pghs975uc
/
infer_clone
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
Branches Tags
${ item.name }
Create tag ${ searchTerm }
Create branch ${ searchTerm }
from 'ec118e62a7'
${ noResults }
infer_clone/infer/tests/codetoanalyze/c/frontend/switchstmt/switch.c.dot

807 lines
46 KiB
Raw Normal View History Unescape

Add @generated comment to icfg.dot files Reviewed By: jvillard Differential Revision: D3358243 fbshipit-source-id: a47cc01
9 years ago
/* @generated */
initial synchronization
10 years ago
digraph iCFG {
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_1" [label="1: Start m1\nFormals: \nLocals: x:int value:int \n DECLARE_LOCALS(&return,&x,&value); [line 12, column 1]\n " color=yellow style=filled]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_1" -> "m1.ae7be26cdaa742ca148068d5ac90eaca_23" ;
"m1.ae7be26cdaa742ca148068d5ac90eaca_2" [label="2: Exit m1 \n " color=yellow style=filled]
initial synchronization
10 years ago
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_3" [label="3: Return Stmt \n *&return:int=0 [line 32, column 3]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_3" -> "m1.ae7be26cdaa742ca148068d5ac90eaca_2" ;
"m1.ae7be26cdaa742ca148068d5ac90eaca_4" [label="4: + \n " ]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_4" -> "m1.ae7be26cdaa742ca148068d5ac90eaca_5" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_5" [label="5: BinaryOperatorStmt: LT \n n$0=*&value:int [line 14, column 10]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_5" -> "m1.ae7be26cdaa742ca148068d5ac90eaca_6" ;
"m1.ae7be26cdaa742ca148068d5ac90eaca_5" -> "m1.ae7be26cdaa742ca148068d5ac90eaca_7" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_6" [label="6: Prune (true branch) \n PRUNE((n$0 < 10), true); [line 14, column 10]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_6" -> "m1.ae7be26cdaa742ca148068d5ac90eaca_9" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_7" [label="7: Prune (false branch) \n PRUNE(!(n$0 < 10), false); [line 14, column 10]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_7" -> "m1.ae7be26cdaa742ca148068d5ac90eaca_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_8" [label="8: Call _fun_printf \n n$1=_fun_printf(\"(after_switch)HELLO WORLD!\":char const *) [line 30, column 5]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_8" -> "m1.ae7be26cdaa742ca148068d5ac90eaca_4" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_9" [label="9: Switch_stmt \n n$2=*&value:int [line 15, column 13]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_9" -> "m1.ae7be26cdaa742ca148068d5ac90eaca_18" ;
"m1.ae7be26cdaa742ca148068d5ac90eaca_9" -> "m1.ae7be26cdaa742ca148068d5ac90eaca_19" ;
"m1.ae7be26cdaa742ca148068d5ac90eaca_10" [label="10: DefaultStmt_placeholder \n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_10" -> "m1.ae7be26cdaa742ca148068d5ac90eaca_11" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_11" [label="11: Call _fun_printf \n n$3=_fun_printf(\"(2/def)HELLO WORLD!\":char const *) [line 27, column 9]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_11" -> "m1.ae7be26cdaa742ca148068d5ac90eaca_4" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_12" [label="12: Prune (true branch) \n PRUNE((n$2 == 2), true); [line 25, column 7]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_12" -> "m1.ae7be26cdaa742ca148068d5ac90eaca_11" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_13" [label="13: Prune (false branch) \n PRUNE(!(n$2 == 2), false); [line 25, column 7]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_13" -> "m1.ae7be26cdaa742ca148068d5ac90eaca_10" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_14" [label="14: Call _fun_printf \n n$4=_fun_printf(\"(1)HELLO WORLD!\":char const *) [line 23, column 9]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_14" -> "m1.ae7be26cdaa742ca148068d5ac90eaca_4" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_15" [label="15: Prune (true branch) \n PRUNE((n$2 == 1), true); [line 22, column 7]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_15" -> "m1.ae7be26cdaa742ca148068d5ac90eaca_14" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_16" [label="16: Prune (false branch) \n PRUNE(!(n$2 == 1), false); [line 22, column 7]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_16" -> "m1.ae7be26cdaa742ca148068d5ac90eaca_12" ;
"m1.ae7be26cdaa742ca148068d5ac90eaca_16" -> "m1.ae7be26cdaa742ca148068d5ac90eaca_13" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_17" [label="17: Call _fun_printf \n n$5=_fun_printf(\"(0)HELLO WORLD!\":char const *) [line 20, column 9]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_17" -> "m1.ae7be26cdaa742ca148068d5ac90eaca_8" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_18" [label="18: Prune (true branch) \n PRUNE((n$2 == 0), true); [line 19, column 7]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_18" -> "m1.ae7be26cdaa742ca148068d5ac90eaca_17" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_19" [label="19: Prune (false branch) \n PRUNE(!(n$2 == 0), false); [line 19, column 7]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_19" -> "m1.ae7be26cdaa742ca148068d5ac90eaca_15" ;
"m1.ae7be26cdaa742ca148068d5ac90eaca_19" -> "m1.ae7be26cdaa742ca148068d5ac90eaca_16" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_20" [label="20: BinaryOperatorStmt: Assign \n n$6=*&value:int [line 18, column 11]\n *&x:int=(n$6 + 1) [line 18, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_20" -> "m1.ae7be26cdaa742ca148068d5ac90eaca_17" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_21" [label="21: Call _fun_printf \n n$7=_fun_printf(\"(out)HELLO WORLD!\":char const *) [line 17, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_21" -> "m1.ae7be26cdaa742ca148068d5ac90eaca_20" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_22" [label="22: DeclStmt \n *&x:int=1 [line 16, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_22" -> "m1.ae7be26cdaa742ca148068d5ac90eaca_21" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_23" [label="23: DeclStmt \n *&value:int=0 [line 13, column 3]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m1.ae7be26cdaa742ca148068d5ac90eaca_23" -> "m1.ae7be26cdaa742ca148068d5ac90eaca_4" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_1" [label="1: Start m2\nFormals: \nLocals: something:int z:int x:int value:int \n DECLARE_LOCALS(&return,&something,&z,&x,&value); [line 35, column 1]\n " color=yellow style=filled]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_1" -> "m2.aaf2f89992379705dac844c0a2a1d45f_22" ;
"m2.aaf2f89992379705dac844c0a2a1d45f_2" [label="2: Exit m2 \n " color=yellow style=filled]
initial synchronization
10 years ago
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_3" [label="3: Return Stmt \n *&return:int=0 [line 57, column 3]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_3" -> "m2.aaf2f89992379705dac844c0a2a1d45f_2" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_4" [label="4: Switch_stmt \n n$0=*&value:int [line 37, column 11]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_4" -> "m2.aaf2f89992379705dac844c0a2a1d45f_17" ;
"m2.aaf2f89992379705dac844c0a2a1d45f_4" -> "m2.aaf2f89992379705dac844c0a2a1d45f_18" ;
"m2.aaf2f89992379705dac844c0a2a1d45f_5" [label="5: DefaultStmt_placeholder \n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_5" -> "m2.aaf2f89992379705dac844c0a2a1d45f_12" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_6" [label="6: Prune (true branch) \n PRUNE((n$0 == 3), true); [line 54, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_6" -> "m2.aaf2f89992379705dac844c0a2a1d45f_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_7" [label="7: Prune (false branch) \n PRUNE(!(n$0 == 3), false); [line 54, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_7" -> "m2.aaf2f89992379705dac844c0a2a1d45f_5" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_8" [label="8: Prune (true branch) \n PRUNE((n$0 == 2), true); [line 53, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_8" -> "m2.aaf2f89992379705dac844c0a2a1d45f_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_9" [label="9: Prune (false branch) \n PRUNE(!(n$0 == 2), false); [line 53, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_9" -> "m2.aaf2f89992379705dac844c0a2a1d45f_6" ;
"m2.aaf2f89992379705dac844c0a2a1d45f_9" -> "m2.aaf2f89992379705dac844c0a2a1d45f_7" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_10" [label="10: BinaryOperatorStmt: Assign \n *&z:int=42 [line 51, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_10" -> "m2.aaf2f89992379705dac844c0a2a1d45f_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_11" [label="11: UnaryOperator \n n$1=*&something:int [line 49, column 7]\n *&something:int=(n$1 + 1) [line 49, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_11" -> "m2.aaf2f89992379705dac844c0a2a1d45f_10" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_12" [label="12: DeclStmt \n *&something:int=1 [line 48, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_12" -> "m2.aaf2f89992379705dac844c0a2a1d45f_11" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_13" [label="13: Prune (true branch) \n PRUNE((n$0 == 1), true); [line 47, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_13" -> "m2.aaf2f89992379705dac844c0a2a1d45f_12" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_14" [label="14: Prune (false branch) \n PRUNE(!(n$0 == 1), false); [line 47, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_14" -> "m2.aaf2f89992379705dac844c0a2a1d45f_8" ;
"m2.aaf2f89992379705dac844c0a2a1d45f_14" -> "m2.aaf2f89992379705dac844c0a2a1d45f_9" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_15" [label="15: DeclStmt \n *&z:int=9 [line 44, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_15" -> "m2.aaf2f89992379705dac844c0a2a1d45f_12" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_16" [label="16: Call _fun_printf \n n$2=_fun_printf(\"(0)HELLO WORLD!\":char const *) [line 42, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_16" -> "m2.aaf2f89992379705dac844c0a2a1d45f_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_17" [label="17: Prune (true branch) \n PRUNE((n$0 == 0), true); [line 41, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_17" -> "m2.aaf2f89992379705dac844c0a2a1d45f_16" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_18" [label="18: Prune (false branch) \n PRUNE(!(n$0 == 0), false); [line 41, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_18" -> "m2.aaf2f89992379705dac844c0a2a1d45f_13" ;
"m2.aaf2f89992379705dac844c0a2a1d45f_18" -> "m2.aaf2f89992379705dac844c0a2a1d45f_14" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_19" [label="19: BinaryOperatorStmt: Assign \n n$3=*&value:int [line 40, column 9]\n *&x:int=(n$3 + 1) [line 40, column 5]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_19" -> "m2.aaf2f89992379705dac844c0a2a1d45f_16" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_20" [label="20: Call _fun_printf \n n$4=_fun_printf(\"(out)HELLO WORLD!\":char const *) [line 39, column 5]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_20" -> "m2.aaf2f89992379705dac844c0a2a1d45f_19" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_21" [label="21: DeclStmt \n *&x:int=1 [line 38, column 5]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_21" -> "m2.aaf2f89992379705dac844c0a2a1d45f_20" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_22" [label="22: DeclStmt \n *&value:int=0 [line 36, column 3]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m2.aaf2f89992379705dac844c0a2a1d45f_22" -> "m2.aaf2f89992379705dac844c0a2a1d45f_4" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_1" [label="1: Start m3\nFormals: \nLocals: z:int something:int value:int \n DECLARE_LOCALS(&return,&z,&something,&value); [line 60, column 1]\n " color=yellow style=filled]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_1" -> "m3.9678f7a7939f457fa0d9353761e189c7_17" ;
"m3.9678f7a7939f457fa0d9353761e189c7_2" [label="2: Exit m3 \n " color=yellow style=filled]
initial synchronization
10 years ago
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_3" [label="3: Return Stmt \n *&return:int=0 [line 75, column 3]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_3" -> "m3.9678f7a7939f457fa0d9353761e189c7_2" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_4" [label="4: Switch_stmt \n n$0=*&value:int [line 62, column 11]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_4" -> "m3.9678f7a7939f457fa0d9353761e189c7_15" ;
"m3.9678f7a7939f457fa0d9353761e189c7_4" -> "m3.9678f7a7939f457fa0d9353761e189c7_16" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_5" [label="5: Prune (true branch) \n PRUNE((n$0 == 3), true); [line 72, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_5" -> "m3.9678f7a7939f457fa0d9353761e189c7_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_6" [label="6: Prune (false branch) \n PRUNE(!(n$0 == 3), false); [line 72, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_6" -> "m3.9678f7a7939f457fa0d9353761e189c7_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_7" [label="7: Prune (true branch) \n PRUNE((n$0 == 2), true); [line 71, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_7" -> "m3.9678f7a7939f457fa0d9353761e189c7_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_8" [label="8: Prune (false branch) \n PRUNE(!(n$0 == 2), false); [line 71, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_8" -> "m3.9678f7a7939f457fa0d9353761e189c7_5" ;
"m3.9678f7a7939f457fa0d9353761e189c7_8" -> "m3.9678f7a7939f457fa0d9353761e189c7_6" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_9" [label="9: DeclStmt \n *&z:int=9 [line 70, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_9" -> "m3.9678f7a7939f457fa0d9353761e189c7_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_10" [label="10: UnaryOperator \n n$1=*&something:int [line 68, column 7]\n *&something:int=(n$1 + 1) [line 68, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_10" -> "m3.9678f7a7939f457fa0d9353761e189c7_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_11" [label="11: DeclStmt \n *&something:int=1 [line 67, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_11" -> "m3.9678f7a7939f457fa0d9353761e189c7_10" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_12" [label="12: Prune (true branch) \n PRUNE((n$0 == 1), true); [line 66, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_12" -> "m3.9678f7a7939f457fa0d9353761e189c7_11" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_13" [label="13: Prune (false branch) \n PRUNE(!(n$0 == 1), false); [line 66, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_13" -> "m3.9678f7a7939f457fa0d9353761e189c7_7" ;
"m3.9678f7a7939f457fa0d9353761e189c7_13" -> "m3.9678f7a7939f457fa0d9353761e189c7_8" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_14" [label="14: Call _fun_printf \n n$2=_fun_printf(\"(0)HELLO WORLD!\":char const *) [line 64, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_14" -> "m3.9678f7a7939f457fa0d9353761e189c7_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_15" [label="15: Prune (true branch) \n PRUNE((n$0 == 0), true); [line 63, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_15" -> "m3.9678f7a7939f457fa0d9353761e189c7_14" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_16" [label="16: Prune (false branch) \n PRUNE(!(n$0 == 0), false); [line 63, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_16" -> "m3.9678f7a7939f457fa0d9353761e189c7_12" ;
"m3.9678f7a7939f457fa0d9353761e189c7_16" -> "m3.9678f7a7939f457fa0d9353761e189c7_13" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_17" [label="17: DeclStmt \n *&value:int=0 [line 61, column 3]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m3.9678f7a7939f457fa0d9353761e189c7_17" -> "m3.9678f7a7939f457fa0d9353761e189c7_4" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_1" [label="1: Start m4\nFormals: \nLocals: something:int z:int x:int value:int \n DECLARE_LOCALS(&return,&something,&z,&x,&value); [line 78, column 1]\n " color=yellow style=filled]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_1" -> "m4.fd6b6fc9220b72d21683ae8e4f50a210_22" ;
"m4.fd6b6fc9220b72d21683ae8e4f50a210_2" [label="2: Exit m4 \n " color=yellow style=filled]
initial synchronization
10 years ago
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_3" [label="3: Return Stmt \n *&return:int=0 [line 100, column 3]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_3" -> "m4.fd6b6fc9220b72d21683ae8e4f50a210_2" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_4" [label="4: Switch_stmt \n n$0=*&value:int [line 80, column 11]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_4" -> "m4.fd6b6fc9220b72d21683ae8e4f50a210_17" ;
"m4.fd6b6fc9220b72d21683ae8e4f50a210_4" -> "m4.fd6b6fc9220b72d21683ae8e4f50a210_18" ;
"m4.fd6b6fc9220b72d21683ae8e4f50a210_5" [label="5: DefaultStmt_placeholder \n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_5" -> "m4.fd6b6fc9220b72d21683ae8e4f50a210_12" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_6" [label="6: Prune (true branch) \n PRUNE((n$0 == 3), true); [line 97, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_6" -> "m4.fd6b6fc9220b72d21683ae8e4f50a210_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_7" [label="7: Prune (false branch) \n PRUNE(!(n$0 == 3), false); [line 97, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_7" -> "m4.fd6b6fc9220b72d21683ae8e4f50a210_5" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_8" [label="8: Prune (true branch) \n PRUNE((n$0 == 2), true); [line 96, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_8" -> "m4.fd6b6fc9220b72d21683ae8e4f50a210_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_9" [label="9: Prune (false branch) \n PRUNE(!(n$0 == 2), false); [line 96, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_9" -> "m4.fd6b6fc9220b72d21683ae8e4f50a210_6" ;
"m4.fd6b6fc9220b72d21683ae8e4f50a210_9" -> "m4.fd6b6fc9220b72d21683ae8e4f50a210_7" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_10" [label="10: BinaryOperatorStmt: Assign \n *&z:int=42 [line 94, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_10" -> "m4.fd6b6fc9220b72d21683ae8e4f50a210_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_11" [label="11: UnaryOperator \n n$1=*&something:int [line 92, column 7]\n *&something:int=(n$1 + 1) [line 92, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_11" -> "m4.fd6b6fc9220b72d21683ae8e4f50a210_10" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_12" [label="12: DeclStmt \n *&something:int=1 [line 91, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_12" -> "m4.fd6b6fc9220b72d21683ae8e4f50a210_11" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_13" [label="13: Prune (true branch) \n PRUNE((n$0 == 1), true); [line 90, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_13" -> "m4.fd6b6fc9220b72d21683ae8e4f50a210_12" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_14" [label="14: Prune (false branch) \n PRUNE(!(n$0 == 1), false); [line 90, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_14" -> "m4.fd6b6fc9220b72d21683ae8e4f50a210_8" ;
"m4.fd6b6fc9220b72d21683ae8e4f50a210_14" -> "m4.fd6b6fc9220b72d21683ae8e4f50a210_9" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_15" [label="15: DeclStmt \n *&z:int=9 [line 87, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_15" -> "m4.fd6b6fc9220b72d21683ae8e4f50a210_12" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_16" [label="16: Call _fun_printf \n n$2=_fun_printf(\"(0)HELLO WORLD!\":char const *) [line 85, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_16" -> "m4.fd6b6fc9220b72d21683ae8e4f50a210_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_17" [label="17: Prune (true branch) \n PRUNE((n$0 == 0), true); [line 84, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_17" -> "m4.fd6b6fc9220b72d21683ae8e4f50a210_16" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_18" [label="18: Prune (false branch) \n PRUNE(!(n$0 == 0), false); [line 84, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_18" -> "m4.fd6b6fc9220b72d21683ae8e4f50a210_13" ;
"m4.fd6b6fc9220b72d21683ae8e4f50a210_18" -> "m4.fd6b6fc9220b72d21683ae8e4f50a210_14" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_19" [label="19: BinaryOperatorStmt: Assign \n n$3=*&value:int [line 83, column 9]\n *&x:int=(n$3 + 1) [line 83, column 5]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_19" -> "m4.fd6b6fc9220b72d21683ae8e4f50a210_16" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_20" [label="20: Call _fun_printf \n n$4=_fun_printf(\"(out)HELLO WORLD!\":char const *) [line 82, column 5]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_20" -> "m4.fd6b6fc9220b72d21683ae8e4f50a210_19" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_21" [label="21: DeclStmt \n *&x:int=1 [line 81, column 5]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_21" -> "m4.fd6b6fc9220b72d21683ae8e4f50a210_20" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_22" [label="22: DeclStmt \n *&value:int=0 [line 79, column 3]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m4.fd6b6fc9220b72d21683ae8e4f50a210_22" -> "m4.fd6b6fc9220b72d21683ae8e4f50a210_4" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_1" [label="1: Start m5\nFormals: \nLocals: x:int value:int \n DECLARE_LOCALS(&return,&x,&value); [line 103, column 1]\n " color=yellow style=filled]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_1" -> "m5.7b1f6dff14d8c2dfeb7da9487be0612d_15" ;
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_2" [label="2: Exit m5 \n " color=yellow style=filled]
initial synchronization
10 years ago
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_3" [label="3: Return Stmt \n *&return:int=0 [line 116, column 3]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_3" -> "m5.7b1f6dff14d8c2dfeb7da9487be0612d_2" ;
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_4" [label="4: + \n " ]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_4" -> "m5.7b1f6dff14d8c2dfeb7da9487be0612d_5" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_5" [label="5: BinaryOperatorStmt: LT \n n$0=*&value:int [line 105, column 10]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_5" -> "m5.7b1f6dff14d8c2dfeb7da9487be0612d_6" ;
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_5" -> "m5.7b1f6dff14d8c2dfeb7da9487be0612d_7" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_6" [label="6: Prune (true branch) \n PRUNE((n$0 < 10), true); [line 105, column 10]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_6" -> "m5.7b1f6dff14d8c2dfeb7da9487be0612d_8" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_7" [label="7: Prune (false branch) \n PRUNE(!(n$0 < 10), false); [line 105, column 10]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_7" -> "m5.7b1f6dff14d8c2dfeb7da9487be0612d_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_8" [label="8: Switch_stmt \n n$1=*&value:int [line 106, column 13]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_8" -> "m5.7b1f6dff14d8c2dfeb7da9487be0612d_10" ;
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_8" -> "m5.7b1f6dff14d8c2dfeb7da9487be0612d_11" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_9" [label="9: Call _fun_printf \n n$2=_fun_printf(\"(0)HELLO WORLD!\":char const *) [line 112, column 9]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_9" -> "m5.7b1f6dff14d8c2dfeb7da9487be0612d_4" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_10" [label="10: Prune (true branch) \n PRUNE((n$1 == 0), true); [line 111, column 7]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_10" -> "m5.7b1f6dff14d8c2dfeb7da9487be0612d_9" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_11" [label="11: Prune (false branch) \n PRUNE(!(n$1 == 0), false); [line 111, column 7]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_11" -> "m5.7b1f6dff14d8c2dfeb7da9487be0612d_4" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_12" [label="12: BinaryOperatorStmt: Assign \n n$3=*&value:int [line 109, column 11]\n *&x:int=(n$3 + 1) [line 109, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_12" -> "m5.7b1f6dff14d8c2dfeb7da9487be0612d_4" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_13" [label="13: Call _fun_printf \n n$4=_fun_printf(\"(out)HELLO WORLD!\":char const *) [line 108, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_13" -> "m5.7b1f6dff14d8c2dfeb7da9487be0612d_12" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_14" [label="14: DeclStmt \n *&x:int=1 [line 107, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_14" -> "m5.7b1f6dff14d8c2dfeb7da9487be0612d_13" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_15" [label="15: DeclStmt \n *&value:int=0 [line 104, column 3]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m5.7b1f6dff14d8c2dfeb7da9487be0612d_15" -> "m5.7b1f6dff14d8c2dfeb7da9487be0612d_4" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_1" [label="1: Start m6\nFormals: \nLocals: 0$?%__sil_tmpSIL_temp_conditional___n$0:int z:int something:int value:int \n DECLARE_LOCALS(&return,&0$?%__sil_tmpSIL_temp_conditional___n$0,&z,&something,&value); [line 119, column 1]\n " color=yellow style=filled]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_1" -> "m6.36604411a85db2bd9e97e22bfb5b692d_23" ;
"m6.36604411a85db2bd9e97e22bfb5b692d_2" [label="2: Exit m6 \n " color=yellow style=filled]
initial synchronization
10 years ago
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_3" [label="3: Return Stmt \n *&return:int=0 [line 134, column 3]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_3" -> "m6.36604411a85db2bd9e97e22bfb5b692d_2" ;
"m6.36604411a85db2bd9e97e22bfb5b692d_4" [label="4: + \n " ]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_4" -> "m6.36604411a85db2bd9e97e22bfb5b692d_10" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_5" [label="5: BinaryOperatorStmt: GT \n n$1=*&value:int [line 121, column 11]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_5" -> "m6.36604411a85db2bd9e97e22bfb5b692d_6" ;
"m6.36604411a85db2bd9e97e22bfb5b692d_5" -> "m6.36604411a85db2bd9e97e22bfb5b692d_7" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_6" [label="6: Prune (true branch) \n PRUNE((n$1 > 0), true); [line 121, column 11]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_6" -> "m6.36604411a85db2bd9e97e22bfb5b692d_8" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_7" [label="7: Prune (false branch) \n PRUNE(!(n$1 > 0), false); [line 121, column 11]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_7" -> "m6.36604411a85db2bd9e97e22bfb5b692d_9" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_8" [label="8: ConditinalStmt Branch \n *&0$?%__sil_tmpSIL_temp_conditional___n$0:int=1 [line 121, column 11]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_8" -> "m6.36604411a85db2bd9e97e22bfb5b692d_4" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_9" [label="9: ConditinalStmt Branch \n *&0$?%__sil_tmpSIL_temp_conditional___n$0:int=0 [line 121, column 11]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_9" -> "m6.36604411a85db2bd9e97e22bfb5b692d_4" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_10" [label="10: Switch_stmt \n n$2=*&0$?%__sil_tmpSIL_temp_conditional___n$0:int [line 121, column 11]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_10" -> "m6.36604411a85db2bd9e97e22bfb5b692d_21" ;
"m6.36604411a85db2bd9e97e22bfb5b692d_10" -> "m6.36604411a85db2bd9e97e22bfb5b692d_22" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_11" [label="11: Prune (true branch) \n PRUNE((n$2 == 3), true); [line 131, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_11" -> "m6.36604411a85db2bd9e97e22bfb5b692d_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_12" [label="12: Prune (false branch) \n PRUNE(!(n$2 == 3), false); [line 131, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_12" -> "m6.36604411a85db2bd9e97e22bfb5b692d_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_13" [label="13: Prune (true branch) \n PRUNE((n$2 == 2), true); [line 130, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_13" -> "m6.36604411a85db2bd9e97e22bfb5b692d_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_14" [label="14: Prune (false branch) \n PRUNE(!(n$2 == 2), false); [line 130, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_14" -> "m6.36604411a85db2bd9e97e22bfb5b692d_11" ;
"m6.36604411a85db2bd9e97e22bfb5b692d_14" -> "m6.36604411a85db2bd9e97e22bfb5b692d_12" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_15" [label="15: DeclStmt \n *&z:int=9 [line 129, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_15" -> "m6.36604411a85db2bd9e97e22bfb5b692d_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_16" [label="16: UnaryOperator \n n$3=*&something:int [line 127, column 7]\n *&something:int=(n$3 + 1) [line 127, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_16" -> "m6.36604411a85db2bd9e97e22bfb5b692d_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_17" [label="17: DeclStmt \n *&something:int=1 [line 126, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_17" -> "m6.36604411a85db2bd9e97e22bfb5b692d_16" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_18" [label="18: Prune (true branch) \n PRUNE((n$2 == 1), true); [line 125, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_18" -> "m6.36604411a85db2bd9e97e22bfb5b692d_17" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_19" [label="19: Prune (false branch) \n PRUNE(!(n$2 == 1), false); [line 125, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_19" -> "m6.36604411a85db2bd9e97e22bfb5b692d_13" ;
"m6.36604411a85db2bd9e97e22bfb5b692d_19" -> "m6.36604411a85db2bd9e97e22bfb5b692d_14" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_20" [label="20: Call _fun_printf \n n$4=_fun_printf(\"(0)HELLO WORLD!\":char const *) [line 123, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_20" -> "m6.36604411a85db2bd9e97e22bfb5b692d_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_21" [label="21: Prune (true branch) \n PRUNE((n$2 == 0), true); [line 122, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_21" -> "m6.36604411a85db2bd9e97e22bfb5b692d_20" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_22" [label="22: Prune (false branch) \n PRUNE(!(n$2 == 0), false); [line 122, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_22" -> "m6.36604411a85db2bd9e97e22bfb5b692d_18" ;
"m6.36604411a85db2bd9e97e22bfb5b692d_22" -> "m6.36604411a85db2bd9e97e22bfb5b692d_19" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_23" [label="23: DeclStmt \n *&value:int=0 [line 120, column 3]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m6.36604411a85db2bd9e97e22bfb5b692d_23" -> "m6.36604411a85db2bd9e97e22bfb5b692d_5" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"getValue.faa0c7b1433b0c97fcdc15fa47c8180f_1" [label="1: Start getValue\nFormals: \nLocals: \n DECLARE_LOCALS(&return); [line 137, column 1]\n " color=yellow style=filled]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"getValue.faa0c7b1433b0c97fcdc15fa47c8180f_1" -> "getValue.faa0c7b1433b0c97fcdc15fa47c8180f_3" ;
"getValue.faa0c7b1433b0c97fcdc15fa47c8180f_2" [label="2: Exit getValue \n " color=yellow style=filled]
initial synchronization
10 years ago
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"getValue.faa0c7b1433b0c97fcdc15fa47c8180f_3" [label="3: Return Stmt \n *&return:int=1 [line 137, column 18]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"getValue.faa0c7b1433b0c97fcdc15fa47c8180f_3" -> "getValue.faa0c7b1433b0c97fcdc15fa47c8180f_2" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_1" [label="1: Start m7\nFormals: \nLocals: z:int something:int value:int \n DECLARE_LOCALS(&return,&z,&something,&value); [line 139, column 1]\n " color=yellow style=filled]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_1" -> "m7.0449904fbf32607bf8ce5c26823dbc29_17" ;
"m7.0449904fbf32607bf8ce5c26823dbc29_2" [label="2: Exit m7 \n " color=yellow style=filled]
initial synchronization
10 years ago
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_3" [label="3: Return Stmt \n *&return:int=0 [line 154, column 3]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_3" -> "m7.0449904fbf32607bf8ce5c26823dbc29_2" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_4" [label="4: Switch_stmt \n n$0=_fun_getValue() [line 141, column 11]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_4" -> "m7.0449904fbf32607bf8ce5c26823dbc29_15" ;
"m7.0449904fbf32607bf8ce5c26823dbc29_4" -> "m7.0449904fbf32607bf8ce5c26823dbc29_16" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_5" [label="5: Prune (true branch) \n PRUNE((n$0 == 3), true); [line 151, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_5" -> "m7.0449904fbf32607bf8ce5c26823dbc29_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_6" [label="6: Prune (false branch) \n PRUNE(!(n$0 == 3), false); [line 151, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_6" -> "m7.0449904fbf32607bf8ce5c26823dbc29_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_7" [label="7: Prune (true branch) \n PRUNE((n$0 == 2), true); [line 150, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_7" -> "m7.0449904fbf32607bf8ce5c26823dbc29_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_8" [label="8: Prune (false branch) \n PRUNE(!(n$0 == 2), false); [line 150, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_8" -> "m7.0449904fbf32607bf8ce5c26823dbc29_5" ;
"m7.0449904fbf32607bf8ce5c26823dbc29_8" -> "m7.0449904fbf32607bf8ce5c26823dbc29_6" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_9" [label="9: DeclStmt \n *&z:int=9 [line 149, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_9" -> "m7.0449904fbf32607bf8ce5c26823dbc29_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_10" [label="10: UnaryOperator \n n$1=*&something:int [line 147, column 7]\n *&something:int=(n$1 + 1) [line 147, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_10" -> "m7.0449904fbf32607bf8ce5c26823dbc29_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_11" [label="11: DeclStmt \n *&something:int=1 [line 146, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_11" -> "m7.0449904fbf32607bf8ce5c26823dbc29_10" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_12" [label="12: Prune (true branch) \n PRUNE((n$0 == 1), true); [line 145, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_12" -> "m7.0449904fbf32607bf8ce5c26823dbc29_11" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_13" [label="13: Prune (false branch) \n PRUNE(!(n$0 == 1), false); [line 145, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_13" -> "m7.0449904fbf32607bf8ce5c26823dbc29_7" ;
"m7.0449904fbf32607bf8ce5c26823dbc29_13" -> "m7.0449904fbf32607bf8ce5c26823dbc29_8" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_14" [label="14: Call _fun_printf \n n$2=_fun_printf(\"(0)HELLO WORLD!\":char const *) [line 143, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_14" -> "m7.0449904fbf32607bf8ce5c26823dbc29_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_15" [label="15: Prune (true branch) \n PRUNE((n$0 == 0), true); [line 142, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_15" -> "m7.0449904fbf32607bf8ce5c26823dbc29_14" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_16" [label="16: Prune (false branch) \n PRUNE(!(n$0 == 0), false); [line 142, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_16" -> "m7.0449904fbf32607bf8ce5c26823dbc29_12" ;
"m7.0449904fbf32607bf8ce5c26823dbc29_16" -> "m7.0449904fbf32607bf8ce5c26823dbc29_13" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_17" [label="17: DeclStmt \n *&value:int=0 [line 140, column 3]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m7.0449904fbf32607bf8ce5c26823dbc29_17" -> "m7.0449904fbf32607bf8ce5c26823dbc29_4" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_1" [label="1: Start m8\nFormals: \nLocals: a:int 0$?%__sil_tmpSIL_temp_conditional___n$1:int z:int something:int value:int \n DECLARE_LOCALS(&return,&a,&0$?%__sil_tmpSIL_temp_conditional___n$1,&z,&something,&value); [line 157, column 1]\n " color=yellow style=filled]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_1" -> "m8.980b79c2a71b9bcc117e08a990b5b332_29" ;
"m8.980b79c2a71b9bcc117e08a990b5b332_2" [label="2: Exit m8 \n " color=yellow style=filled]
initial synchronization
10 years ago
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_3" [label="3: Return Stmt \n *&return:int=0 [line 176, column 3]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_3" -> "m8.980b79c2a71b9bcc117e08a990b5b332_2" ;
"m8.980b79c2a71b9bcc117e08a990b5b332_4" [label="4: + \n " ]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_4" -> "m8.980b79c2a71b9bcc117e08a990b5b332_5" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_5" [label="5: BinaryOperatorStmt: LT \n n$0=*&value:int [line 159, column 10]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_5" -> "m8.980b79c2a71b9bcc117e08a990b5b332_6" ;
"m8.980b79c2a71b9bcc117e08a990b5b332_5" -> "m8.980b79c2a71b9bcc117e08a990b5b332_7" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_6" [label="6: Prune (true branch) \n PRUNE((n$0 < 10), true); [line 159, column 10]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_6" -> "m8.980b79c2a71b9bcc117e08a990b5b332_10" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_7" [label="7: Prune (false branch) \n PRUNE(!(n$0 < 10), false); [line 159, column 10]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_7" -> "m8.980b79c2a71b9bcc117e08a990b5b332_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_8" [label="8: DeclStmt \n *&a:int=0 [line 174, column 5]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_8" -> "m8.980b79c2a71b9bcc117e08a990b5b332_4" ;
"m8.980b79c2a71b9bcc117e08a990b5b332_9" [label="9: + \n " ]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_9" -> "m8.980b79c2a71b9bcc117e08a990b5b332_15" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_10" [label="10: BinaryOperatorStmt: EQ \n n$2=_fun_getValue() [line 160, column 13]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_10" -> "m8.980b79c2a71b9bcc117e08a990b5b332_11" ;
"m8.980b79c2a71b9bcc117e08a990b5b332_10" -> "m8.980b79c2a71b9bcc117e08a990b5b332_12" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_11" [label="11: Prune (true branch) \n PRUNE((n$2 == 0), true); [line 160, column 13]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_11" -> "m8.980b79c2a71b9bcc117e08a990b5b332_13" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_12" [label="12: Prune (false branch) \n PRUNE(!(n$2 == 0), false); [line 160, column 13]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_12" -> "m8.980b79c2a71b9bcc117e08a990b5b332_14" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_13" [label="13: ConditinalStmt Branch \n *&0$?%__sil_tmpSIL_temp_conditional___n$1:int=1 [line 160, column 13]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_13" -> "m8.980b79c2a71b9bcc117e08a990b5b332_9" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_14" [label="14: ConditinalStmt Branch \n *&0$?%__sil_tmpSIL_temp_conditional___n$1:int=2 [line 160, column 13]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_14" -> "m8.980b79c2a71b9bcc117e08a990b5b332_9" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_15" [label="15: Switch_stmt \n n$3=*&0$?%__sil_tmpSIL_temp_conditional___n$1:int [line 160, column 13]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_15" -> "m8.980b79c2a71b9bcc117e08a990b5b332_27" ;
"m8.980b79c2a71b9bcc117e08a990b5b332_15" -> "m8.980b79c2a71b9bcc117e08a990b5b332_28" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_16" [label="16: Prune (true branch) \n PRUNE((n$3 == 3), true); [line 171, column 7]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_16" -> "m8.980b79c2a71b9bcc117e08a990b5b332_8" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_17" [label="17: Prune (false branch) \n PRUNE(!(n$3 == 3), false); [line 171, column 7]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_17" -> "m8.980b79c2a71b9bcc117e08a990b5b332_8" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_18" [label="18: Prune (true branch) \n PRUNE((n$3 == 2), true); [line 170, column 7]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_18" -> "m8.980b79c2a71b9bcc117e08a990b5b332_8" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_19" [label="19: Prune (false branch) \n PRUNE(!(n$3 == 2), false); [line 170, column 7]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_19" -> "m8.980b79c2a71b9bcc117e08a990b5b332_16" ;
"m8.980b79c2a71b9bcc117e08a990b5b332_19" -> "m8.980b79c2a71b9bcc117e08a990b5b332_17" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_20" [label="20: DeclStmt \n *&z:int=9 [line 169, column 9]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_20" -> "m8.980b79c2a71b9bcc117e08a990b5b332_8" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_21" [label="21: UnaryOperator \n n$4=*&something:int [line 166, column 9]\n *&something:int=(n$4 + 1) [line 166, column 9]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_21" -> "m8.980b79c2a71b9bcc117e08a990b5b332_4" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_22" [label="22: DeclStmt \n *&something:int=1 [line 165, column 9]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_22" -> "m8.980b79c2a71b9bcc117e08a990b5b332_21" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_23" [label="23: Prune (true branch) \n PRUNE((n$3 == 1), true); [line 164, column 7]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_23" -> "m8.980b79c2a71b9bcc117e08a990b5b332_22" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_24" [label="24: Prune (false branch) \n PRUNE(!(n$3 == 1), false); [line 164, column 7]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_24" -> "m8.980b79c2a71b9bcc117e08a990b5b332_18" ;
"m8.980b79c2a71b9bcc117e08a990b5b332_24" -> "m8.980b79c2a71b9bcc117e08a990b5b332_19" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_25" [label="25: Return Stmt \n *&return:int=0 [line 163, column 9]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_25" -> "m8.980b79c2a71b9bcc117e08a990b5b332_2" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_26" [label="26: Call _fun_printf \n n$5=_fun_printf(\"(0)HELLO WORLD!\":char const *) [line 162, column 9]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_26" -> "m8.980b79c2a71b9bcc117e08a990b5b332_25" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_27" [label="27: Prune (true branch) \n PRUNE((n$3 == 0), true); [line 161, column 7]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_27" -> "m8.980b79c2a71b9bcc117e08a990b5b332_26" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_28" [label="28: Prune (false branch) \n PRUNE(!(n$3 == 0), false); [line 161, column 7]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_28" -> "m8.980b79c2a71b9bcc117e08a990b5b332_23" ;
"m8.980b79c2a71b9bcc117e08a990b5b332_28" -> "m8.980b79c2a71b9bcc117e08a990b5b332_24" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_29" [label="29: DeclStmt \n *&value:int=0 [line 158, column 3]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m8.980b79c2a71b9bcc117e08a990b5b332_29" -> "m8.980b79c2a71b9bcc117e08a990b5b332_4" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m9.5bbb291cc1e38a051365ee9edb7cbd14_1" [label="1: Start m9\nFormals: \nLocals: value:int \n DECLARE_LOCALS(&return,&value); [line 179, column 1]\n " color=yellow style=filled]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m9.5bbb291cc1e38a051365ee9edb7cbd14_1" -> "m9.5bbb291cc1e38a051365ee9edb7cbd14_5" ;
"m9.5bbb291cc1e38a051365ee9edb7cbd14_2" [label="2: Exit m9 \n " color=yellow style=filled]
initial synchronization
10 years ago
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m9.5bbb291cc1e38a051365ee9edb7cbd14_3" [label="3: Return Stmt \n *&return:int=0 [line 182, column 3]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m9.5bbb291cc1e38a051365ee9edb7cbd14_3" -> "m9.5bbb291cc1e38a051365ee9edb7cbd14_2" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m9.5bbb291cc1e38a051365ee9edb7cbd14_4" [label="4: Switch_stmt \n n$0=*&value:int [line 181, column 11]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m9.5bbb291cc1e38a051365ee9edb7cbd14_4" -> "m9.5bbb291cc1e38a051365ee9edb7cbd14_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m9.5bbb291cc1e38a051365ee9edb7cbd14_5" [label="5: DeclStmt \n *&value:int=0 [line 180, column 3]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m9.5bbb291cc1e38a051365ee9edb7cbd14_5" -> "m9.5bbb291cc1e38a051365ee9edb7cbd14_4" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m10.e66050aa5d0a7e0ecb49429ea4b0a32b_1" [label="1: Start m10\nFormals: \nLocals: value:int \n DECLARE_LOCALS(&return,&value); [line 185, column 1]\n " color=yellow style=filled]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m10.e66050aa5d0a7e0ecb49429ea4b0a32b_1" -> "m10.e66050aa5d0a7e0ecb49429ea4b0a32b_5" ;
"m10.e66050aa5d0a7e0ecb49429ea4b0a32b_2" [label="2: Exit m10 \n " color=yellow style=filled]
initial synchronization
10 years ago
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m10.e66050aa5d0a7e0ecb49429ea4b0a32b_3" [label="3: Return Stmt \n *&return:int=0 [line 188, column 3]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m10.e66050aa5d0a7e0ecb49429ea4b0a32b_3" -> "m10.e66050aa5d0a7e0ecb49429ea4b0a32b_2" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m10.e66050aa5d0a7e0ecb49429ea4b0a32b_4" [label="4: Switch_stmt \n *&value:int=7 [line 187, column 11]\n n$0=*&value:int [line 187, column 11]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m10.e66050aa5d0a7e0ecb49429ea4b0a32b_4" -> "m10.e66050aa5d0a7e0ecb49429ea4b0a32b_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m10.e66050aa5d0a7e0ecb49429ea4b0a32b_5" [label="5: DeclStmt \n *&value:int=0 [line 186, column 3]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m10.e66050aa5d0a7e0ecb49429ea4b0a32b_5" -> "m10.e66050aa5d0a7e0ecb49429ea4b0a32b_4" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m11.c4534fe0ca256b331e9a3f14fe17229d_1" [label="1: Start m11\nFormals: \nLocals: 0$?%__sil_tmpSIL_temp_conditional___n$0:int value:int \n DECLARE_LOCALS(&return,&0$?%__sil_tmpSIL_temp_conditional___n$0,&value); [line 191, column 1]\n " color=yellow style=filled]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m11.c4534fe0ca256b331e9a3f14fe17229d_1" -> "m11.c4534fe0ca256b331e9a3f14fe17229d_14" ;
"m11.c4534fe0ca256b331e9a3f14fe17229d_2" [label="2: Exit m11 \n " color=yellow style=filled]
initial synchronization
10 years ago
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m11.c4534fe0ca256b331e9a3f14fe17229d_3" [label="3: Return Stmt \n *&return:int=0 [line 197, column 3]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m11.c4534fe0ca256b331e9a3f14fe17229d_3" -> "m11.c4534fe0ca256b331e9a3f14fe17229d_2" ;
"m11.c4534fe0ca256b331e9a3f14fe17229d_4" [label="4: + \n " ]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m11.c4534fe0ca256b331e9a3f14fe17229d_4" -> "m11.c4534fe0ca256b331e9a3f14fe17229d_10" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m11.c4534fe0ca256b331e9a3f14fe17229d_5" [label="5: BinaryOperatorStmt: EQ \n n$1=*&value:int [line 193, column 20]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m11.c4534fe0ca256b331e9a3f14fe17229d_5" -> "m11.c4534fe0ca256b331e9a3f14fe17229d_6" ;
"m11.c4534fe0ca256b331e9a3f14fe17229d_5" -> "m11.c4534fe0ca256b331e9a3f14fe17229d_7" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m11.c4534fe0ca256b331e9a3f14fe17229d_6" [label="6: Prune (true branch) \n PRUNE((n$1 == 0), true); [line 193, column 20]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m11.c4534fe0ca256b331e9a3f14fe17229d_6" -> "m11.c4534fe0ca256b331e9a3f14fe17229d_8" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m11.c4534fe0ca256b331e9a3f14fe17229d_7" [label="7: Prune (false branch) \n PRUNE(!(n$1 == 0), false); [line 193, column 20]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m11.c4534fe0ca256b331e9a3f14fe17229d_7" -> "m11.c4534fe0ca256b331e9a3f14fe17229d_9" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m11.c4534fe0ca256b331e9a3f14fe17229d_8" [label="8: ConditinalStmt Branch \n *&0$?%__sil_tmpSIL_temp_conditional___n$0:int=7 [line 193, column 20]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m11.c4534fe0ca256b331e9a3f14fe17229d_8" -> "m11.c4534fe0ca256b331e9a3f14fe17229d_4" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m11.c4534fe0ca256b331e9a3f14fe17229d_9" [label="9: ConditinalStmt Branch \n *&0$?%__sil_tmpSIL_temp_conditional___n$0:int=9 [line 193, column 20]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m11.c4534fe0ca256b331e9a3f14fe17229d_9" -> "m11.c4534fe0ca256b331e9a3f14fe17229d_4" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m11.c4534fe0ca256b331e9a3f14fe17229d_10" [label="10: Switch_stmt \n n$2=*&0$?%__sil_tmpSIL_temp_conditional___n$0:int [line 193, column 20]\n *&value:int=n$2 [line 193, column 11]\n n$3=*&value:int [line 193, column 11]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m11.c4534fe0ca256b331e9a3f14fe17229d_10" -> "m11.c4534fe0ca256b331e9a3f14fe17229d_12" ;
"m11.c4534fe0ca256b331e9a3f14fe17229d_10" -> "m11.c4534fe0ca256b331e9a3f14fe17229d_13" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m11.c4534fe0ca256b331e9a3f14fe17229d_11" [label="11: Call _fun_printf \n n$4=_fun_printf(\"(0)HELLO WORLD!\":char const *) [line 195, column 7]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m11.c4534fe0ca256b331e9a3f14fe17229d_11" -> "m11.c4534fe0ca256b331e9a3f14fe17229d_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m11.c4534fe0ca256b331e9a3f14fe17229d_12" [label="12: Prune (true branch) \n PRUNE((n$3 == 0), true); [line 194, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m11.c4534fe0ca256b331e9a3f14fe17229d_12" -> "m11.c4534fe0ca256b331e9a3f14fe17229d_11" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m11.c4534fe0ca256b331e9a3f14fe17229d_13" [label="13: Prune (false branch) \n PRUNE(!(n$3 == 0), false); [line 194, column 5]\n " shape="invhouse"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m11.c4534fe0ca256b331e9a3f14fe17229d_13" -> "m11.c4534fe0ca256b331e9a3f14fe17229d_3" ;
[siof] one access per sink, better report deduplication Summary: The previous domain for SIOF was duplicating some work with the generic Trace domain, and basically was a bit confused and confusing. A sink was a set of global accesses, and a state contains a set of sinks. Then the checker has to needlessly jump through hoops to normalize this set of sets of accesses into a set of accesses. The new domain has one sink = one access, as suggested by sblackshear. This simplifies a few things, and makes the dedup logic much easier: just grab the first report of the list of reports for a function. We only report on the fake procedures generated to initialise a global, and the filtering means that we keep only one report per global. Reviewed By: sblackshear Differential Revision: D5932138 fbshipit-source-id: acb7285
7 years ago
"m11.c4534fe0ca256b331e9a3f14fe17229d_14" [label="14: DeclStmt \n *&value:int=0 [line 192, column 3]\n " shape="box"]
initial synchronization
10 years ago
Sort nodes when printing cfg to dot file Summary: Currently cfg nodes are written into dot files in whatever order they appear in a hash table. This seems unnecessarily sensitive, so this diff sorts the nodes. Reviewed By: dulmarod Differential Revision: D4232377 fbshipit-source-id: a907cc6
8 years ago
"m11.c4534fe0ca256b331e9a3f14fe17229d_14" -> "m11.c4534fe0ca256b331e9a3f14fe17229d_5" ;
initial synchronization
10 years ago
}