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 'ef556e048e'
${ noResults }
infer_clone/infer/tests/codetoanalyze/objc/shared/block/dispatch.m.dot

198 lines
21 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
"sharedInstance#DispatchA#class.8992c6086d1ce5c225093940f62386ac_1" [label="1: Start DispatchA_sharedInstance\nFormals: \nLocals: \n DECLARE_LOCALS(&return); [line 27, column 1]\n " color=yellow style=filled]
Fix StmtExpr Summary: The extra dereference in stmtexpr was wrong. When a dereference is needed, we have a cast. This was causing one dereference too many, and creating wrong results. Reviewed By: akotulski Differential Revision: D3393294 fbshipit-source-id: 7a1ec8e
9 years ago
[clang] Change procname file naming scheme Summary: Procnames files are now reversed qualifier lists with `#` as separator (instead of `::` which needs to be escaped in bash). Because of the mechanism that is used to obtain qualifiers, it also affects naming for ObjC classes. Examples: ``` std::unique_ptr<int>::get -> get#unique_ptr<int>#std#__MANGLED,...__ // C++ method folly::split -> split#folly#__MANGLED,..._ // function within namespace NSNumber numberWithBool: -> numberWithBool:#NSNumber#class // ObjC method ``` Reviewed By: jvillard Differential Revision: D4689701 fbshipit-source-id: c3acfc6
8 years ago
"sharedInstance#DispatchA#class.8992c6086d1ce5c225093940f62386ac_1" -> "sharedInstance#DispatchA#class.8992c6086d1ce5c225093940f62386ac_4" ;
"sharedInstance#DispatchA#class.8992c6086d1ce5c225093940f62386ac_2" [label="2: Exit DispatchA_sharedInstance \n " color=yellow style=filled]
Fix StmtExpr Summary: The extra dereference in stmtexpr was wrong. When a dereference is needed, we have a cast. This was causing one dereference too many, and creating wrong results. Reviewed By: akotulski Differential Revision: D3393294 fbshipit-source-id: 7a1ec8e
9 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
"sharedInstance#DispatchA#class.8992c6086d1ce5c225093940f62386ac_3" [label="3: Return Stmt \n n$1=*&#GB<codetoanalyze/objc/shared/block/dispatch.m>$DispatchA_sharedInstance_sharedInstance:objc_object* [line 33, column 10]\n *&return:objc_object*=n$1 [line 33, column 3]\n " shape="box"]
Fix StmtExpr Summary: The extra dereference in stmtexpr was wrong. When a dereference is needed, we have a cast. This was causing one dereference too many, and creating wrong results. Reviewed By: akotulski Differential Revision: D3393294 fbshipit-source-id: 7a1ec8e
9 years ago
[clang] Change procname file naming scheme Summary: Procnames files are now reversed qualifier lists with `#` as separator (instead of `::` which needs to be escaped in bash). Because of the mechanism that is used to obtain qualifiers, it also affects naming for ObjC classes. Examples: ``` std::unique_ptr<int>::get -> get#unique_ptr<int>#std#__MANGLED,...__ // C++ method folly::split -> split#folly#__MANGLED,..._ // function within namespace NSNumber numberWithBool: -> numberWithBool:#NSNumber#class // ObjC method ``` Reviewed By: jvillard Differential Revision: D4689701 fbshipit-source-id: c3acfc6
8 years ago
"sharedInstance#DispatchA#class.8992c6086d1ce5c225093940f62386ac_3" -> "sharedInstance#DispatchA#class.8992c6086d1ce5c225093940f62386ac_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
"sharedInstance#DispatchA#class.8992c6086d1ce5c225093940f62386ac_4" [label="4: Call (_fun___objc_anonymous_block_DispatchA_sharedInstance______1) \n DECLARE_LOCALS(&__objc_anonymous_block_DispatchA_sharedInstance______1); [line 30, column 24]\n n$4=_fun___objc_alloc_no_fail(sizeof(t=__objc_anonymous_block_DispatchA_sharedInstance______1):unsigned long) [line 30, column 24]\n *&__objc_anonymous_block_DispatchA_sharedInstance______1:__objc_anonymous_block_DispatchA_sharedInstance______1=n$4 [line 30, column 24]\n n$5=*&#GB<codetoanalyze/objc/shared/block/dispatch.m>$DispatchA_sharedInstance_sharedInstance:objc_object* [line 30, column 24]\n *n$4.DispatchA_sharedInstance_sharedInstance:objc_object*=n$5 [line 30, column 24]\n n$6=(_fun___objc_anonymous_block_DispatchA_sharedInstance______1)() [line 30, column 3]\n " shape="box"]
Fix StmtExpr Summary: The extra dereference in stmtexpr was wrong. When a dereference is needed, we have a cast. This was causing one dereference too many, and creating wrong results. Reviewed By: akotulski Differential Revision: D3393294 fbshipit-source-id: 7a1ec8e
9 years ago
[clang] Change procname file naming scheme Summary: Procnames files are now reversed qualifier lists with `#` as separator (instead of `::` which needs to be escaped in bash). Because of the mechanism that is used to obtain qualifiers, it also affects naming for ObjC classes. Examples: ``` std::unique_ptr<int>::get -> get#unique_ptr<int>#std#__MANGLED,...__ // C++ method folly::split -> split#folly#__MANGLED,..._ // function within namespace NSNumber numberWithBool: -> numberWithBool:#NSNumber#class // ObjC method ``` Reviewed By: jvillard Differential Revision: D4689701 fbshipit-source-id: c3acfc6
8 years ago
"sharedInstance#DispatchA#class.8992c6086d1ce5c225093940f62386ac_4" -> "sharedInstance#DispatchA#class.8992c6086d1ce5c225093940f62386ac_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
"__objc_anonymous_block_DispatchA_sharedInstance______1.4a2e89fcdf390871f5277dca0d16c43b_1" [label="1: Start __objc_anonymous_block_DispatchA_sharedInstance______1\nFormals: \nLocals: \n DECLARE_LOCALS(&return); [line 30, column 24]\n " color=yellow style=filled]
Fix StmtExpr Summary: The extra dereference in stmtexpr was wrong. When a dereference is needed, we have a cast. This was causing one dereference too many, and creating wrong results. Reviewed By: akotulski Differential Revision: D3393294 fbshipit-source-id: 7a1ec8e
9 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
"__objc_anonymous_block_DispatchA_sharedInstance______1.4a2e89fcdf390871f5277dca0d16c43b_1" -> "__objc_anonymous_block_DispatchA_sharedInstance______1.4a2e89fcdf390871f5277dca0d16c43b_3" ;
"__objc_anonymous_block_DispatchA_sharedInstance______1.4a2e89fcdf390871f5277dca0d16c43b_2" [label="2: Exit __objc_anonymous_block_DispatchA_sharedInstance______1 \n " color=yellow style=filled]
Fix StmtExpr Summary: The extra dereference in stmtexpr was wrong. When a dereference is needed, we have a cast. This was causing one dereference too many, and creating wrong results. Reviewed By: akotulski Differential Revision: D3393294 fbshipit-source-id: 7a1ec8e
9 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
"__objc_anonymous_block_DispatchA_sharedInstance______1.4a2e89fcdf390871f5277dca0d16c43b_3" [label="3: BinaryOperatorStmt: Assign \n n$2=_fun___objc_alloc_no_fail(sizeof(t=DispatchA):unsigned long) [line 31, column 23]\n n$3=_fun_DispatchA_init(n$2:DispatchA*) virtual [line 31, column 22]\n *&#GB<codetoanalyze/objc/shared/block/dispatch.m>$DispatchA_sharedInstance_sharedInstance:objc_object*=n$3 [line 31, column 5]\n " shape="box"]
Fix StmtExpr Summary: The extra dereference in stmtexpr was wrong. When a dereference is needed, we have a cast. This was causing one dereference too many, and creating wrong results. Reviewed By: akotulski Differential Revision: D3393294 fbshipit-source-id: 7a1ec8e
9 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
"__objc_anonymous_block_DispatchA_sharedInstance______1.4a2e89fcdf390871f5277dca0d16c43b_3" -> "__objc_anonymous_block_DispatchA_sharedInstance______1.4a2e89fcdf390871f5277dca0d16c43b_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
"trans#DispatchA#class.23f9d908a87deca79c235bc76ca6e941_1" [label="1: Start DispatchA_trans\nFormals: \nLocals: dummy_block:_fn_(*) \n DECLARE_LOCALS(&return,&dummy_block); [line 36, column 1]\n " color=yellow style=filled]
Fix StmtExpr Summary: The extra dereference in stmtexpr was wrong. When a dereference is needed, we have a cast. This was causing one dereference too many, and creating wrong results. Reviewed By: akotulski Differential Revision: D3393294 fbshipit-source-id: 7a1ec8e
9 years ago
[clang] Change procname file naming scheme Summary: Procnames files are now reversed qualifier lists with `#` as separator (instead of `::` which needs to be escaped in bash). Because of the mechanism that is used to obtain qualifiers, it also affects naming for ObjC classes. Examples: ``` std::unique_ptr<int>::get -> get#unique_ptr<int>#std#__MANGLED,...__ // C++ method folly::split -> split#folly#__MANGLED,..._ // function within namespace NSNumber numberWithBool: -> numberWithBool:#NSNumber#class // ObjC method ``` Reviewed By: jvillard Differential Revision: D4689701 fbshipit-source-id: c3acfc6
8 years ago
"trans#DispatchA#class.23f9d908a87deca79c235bc76ca6e941_1" -> "trans#DispatchA#class.23f9d908a87deca79c235bc76ca6e941_5" ;
"trans#DispatchA#class.23f9d908a87deca79c235bc76ca6e941_2" [label="2: Exit DispatchA_trans \n " color=yellow style=filled]
Fix StmtExpr Summary: The extra dereference in stmtexpr was wrong. When a dereference is needed, we have a cast. This was causing one dereference too many, and creating wrong results. Reviewed By: akotulski Differential Revision: D3393294 fbshipit-source-id: 7a1ec8e
9 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
"trans#DispatchA#class.23f9d908a87deca79c235bc76ca6e941_3" [label="3: Return Stmt \n n$7=*&#GB<codetoanalyze/objc/shared/block/dispatch.m>$DispatchA_trans_sharedInstance:objc_object* [line 42, column 10]\n *&return:objc_object*=n$7 [line 42, column 3]\n " shape="box"]
Fix StmtExpr Summary: The extra dereference in stmtexpr was wrong. When a dereference is needed, we have a cast. This was causing one dereference too many, and creating wrong results. Reviewed By: akotulski Differential Revision: D3393294 fbshipit-source-id: 7a1ec8e
9 years ago
[clang] Change procname file naming scheme Summary: Procnames files are now reversed qualifier lists with `#` as separator (instead of `::` which needs to be escaped in bash). Because of the mechanism that is used to obtain qualifiers, it also affects naming for ObjC classes. Examples: ``` std::unique_ptr<int>::get -> get#unique_ptr<int>#std#__MANGLED,...__ // C++ method folly::split -> split#folly#__MANGLED,..._ // function within namespace NSNumber numberWithBool: -> numberWithBool:#NSNumber#class // ObjC method ``` Reviewed By: jvillard Differential Revision: D4689701 fbshipit-source-id: c3acfc6
8 years ago
"trans#DispatchA#class.23f9d908a87deca79c235bc76ca6e941_3" -> "trans#DispatchA#class.23f9d908a87deca79c235bc76ca6e941_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
"trans#DispatchA#class.23f9d908a87deca79c235bc76ca6e941_4" [label="4: Call n$8 \n n$8=*&dummy_block:_fn_(*) [line 41, column 3]\n n$8() [line 41, column 3]\n " shape="box"]
Fix StmtExpr Summary: The extra dereference in stmtexpr was wrong. When a dereference is needed, we have a cast. This was causing one dereference too many, and creating wrong results. Reviewed By: akotulski Differential Revision: D3393294 fbshipit-source-id: 7a1ec8e
9 years ago
[clang] Change procname file naming scheme Summary: Procnames files are now reversed qualifier lists with `#` as separator (instead of `::` which needs to be escaped in bash). Because of the mechanism that is used to obtain qualifiers, it also affects naming for ObjC classes. Examples: ``` std::unique_ptr<int>::get -> get#unique_ptr<int>#std#__MANGLED,...__ // C++ method folly::split -> split#folly#__MANGLED,..._ // function within namespace NSNumber numberWithBool: -> numberWithBool:#NSNumber#class // ObjC method ``` Reviewed By: jvillard Differential Revision: D4689701 fbshipit-source-id: c3acfc6
8 years ago
"trans#DispatchA#class.23f9d908a87deca79c235bc76ca6e941_4" -> "trans#DispatchA#class.23f9d908a87deca79c235bc76ca6e941_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
"trans#DispatchA#class.23f9d908a87deca79c235bc76ca6e941_5" [label="5: DeclStmt \n DECLARE_LOCALS(&__objc_anonymous_block_DispatchA_trans______2); [line 38, column 27]\n n$11=_fun___objc_alloc_no_fail(sizeof(t=__objc_anonymous_block_DispatchA_trans______2):unsigned long) [line 38, column 27]\n *&__objc_anonymous_block_DispatchA_trans______2:__objc_anonymous_block_DispatchA_trans______2=n$11 [line 38, column 27]\n n$12=*&#GB<codetoanalyze/objc/shared/block/dispatch.m>$DispatchA_trans_sharedInstance:objc_object* [line 38, column 27]\n *n$11.DispatchA_trans_sharedInstance:objc_object*=n$12 [line 38, column 27]\n *&dummy_block:_fn_(*)=(_fun___objc_anonymous_block_DispatchA_trans______2) [line 38, column 3]\n " shape="box"]
initial synchronization
10 years ago
[clang] Change procname file naming scheme Summary: Procnames files are now reversed qualifier lists with `#` as separator (instead of `::` which needs to be escaped in bash). Because of the mechanism that is used to obtain qualifiers, it also affects naming for ObjC classes. Examples: ``` std::unique_ptr<int>::get -> get#unique_ptr<int>#std#__MANGLED,...__ // C++ method folly::split -> split#folly#__MANGLED,..._ // function within namespace NSNumber numberWithBool: -> numberWithBool:#NSNumber#class // ObjC method ``` Reviewed By: jvillard Differential Revision: D4689701 fbshipit-source-id: c3acfc6
8 years ago
"trans#DispatchA#class.23f9d908a87deca79c235bc76ca6e941_5" -> "trans#DispatchA#class.23f9d908a87deca79c235bc76ca6e941_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
"__objc_anonymous_block_DispatchA_trans______2.8ca180fe9a17b86cb599eced71242770_1" [label="1: Start __objc_anonymous_block_DispatchA_trans______2\nFormals: \nLocals: \n DECLARE_LOCALS(&return); [line 38, column 27]\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
"__objc_anonymous_block_DispatchA_trans______2.8ca180fe9a17b86cb599eced71242770_1" -> "__objc_anonymous_block_DispatchA_trans______2.8ca180fe9a17b86cb599eced71242770_3" ;
"__objc_anonymous_block_DispatchA_trans______2.8ca180fe9a17b86cb599eced71242770_2" [label="2: Exit __objc_anonymous_block_DispatchA_trans______2 \n " color=yellow style=filled]
Fix control-flow graph shape when conditional is last instruction. Summary:public When a conditional is the last instruction, there will be a join node leading directly to the exit node. Some instructions, such as nullification of dead variables, and abstraction, are added to the control flow graph automatically. But, join nodes cannot contain instructions. So when a procedure ends with a conditional, there might be no place to store these instructions. This diff adds one extra node between the join and the exit node in that situation. Reviewed By: jvillard Differential Revision: D3179056 fb-gh-sync-id: 2b9cd7e fbshipit-source-id: 2b9cd7e
9 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
"__objc_anonymous_block_DispatchA_trans______2.8ca180fe9a17b86cb599eced71242770_3" [label="3: BinaryOperatorStmt: Assign \n n$9=_fun___objc_alloc_no_fail(sizeof(t=DispatchA):unsigned long) [line 39, column 23]\n n$10=_fun_DispatchA_init(n$9:DispatchA*) virtual [line 39, column 22]\n *&#GB<codetoanalyze/objc/shared/block/dispatch.m>$DispatchA_trans_sharedInstance:objc_object*=n$10 [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
"__objc_anonymous_block_DispatchA_trans______2.8ca180fe9a17b86cb599eced71242770_3" -> "__objc_anonymous_block_DispatchA_trans______2.8ca180fe9a17b86cb599eced71242770_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
"dispatch_a_block_variable#DispatchA#class.3cc12dd22127281b8293b7c046d21bb2_1" [label="1: Start DispatchA_dispatch_a_block_variable\nFormals: \nLocals: initialization_block__:_fn_(*) \n DECLARE_LOCALS(&return,&initialization_block__); [line 45, column 1]\n " color=yellow style=filled]
initial synchronization
10 years ago
[clang] Change procname file naming scheme Summary: Procnames files are now reversed qualifier lists with `#` as separator (instead of `::` which needs to be escaped in bash). Because of the mechanism that is used to obtain qualifiers, it also affects naming for ObjC classes. Examples: ``` std::unique_ptr<int>::get -> get#unique_ptr<int>#std#__MANGLED,...__ // C++ method folly::split -> split#folly#__MANGLED,..._ // function within namespace NSNumber numberWithBool: -> numberWithBool:#NSNumber#class // ObjC method ``` Reviewed By: jvillard Differential Revision: D4689701 fbshipit-source-id: c3acfc6
8 years ago
"dispatch_a_block_variable#DispatchA#class.3cc12dd22127281b8293b7c046d21bb2_1" -> "dispatch_a_block_variable#DispatchA#class.3cc12dd22127281b8293b7c046d21bb2_5" ;
"dispatch_a_block_variable#DispatchA#class.3cc12dd22127281b8293b7c046d21bb2_2" [label="2: Exit DispatchA_dispatch_a_block_variable \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
"dispatch_a_block_variable#DispatchA#class.3cc12dd22127281b8293b7c046d21bb2_3" [label="3: Return Stmt \n n$13=*&#GB<codetoanalyze/objc/shared/block/dispatch.m>$DispatchA_dispatch_a_block_variable_static_storage__:objc_object* [line 52, column 10]\n *&return:objc_object*=n$13 [line 52, column 3]\n " shape="box"]
Adding endtoend test that shows that dispatch once works well Reviewed By: ddino Differential Revision: D2570191 fb-gh-sync-id: 01be61c
9 years ago
[clang] Change procname file naming scheme Summary: Procnames files are now reversed qualifier lists with `#` as separator (instead of `::` which needs to be escaped in bash). Because of the mechanism that is used to obtain qualifiers, it also affects naming for ObjC classes. Examples: ``` std::unique_ptr<int>::get -> get#unique_ptr<int>#std#__MANGLED,...__ // C++ method folly::split -> split#folly#__MANGLED,..._ // function within namespace NSNumber numberWithBool: -> numberWithBool:#NSNumber#class // ObjC method ``` Reviewed By: jvillard Differential Revision: D4689701 fbshipit-source-id: c3acfc6
8 years ago
"dispatch_a_block_variable#DispatchA#class.3cc12dd22127281b8293b7c046d21bb2_3" -> "dispatch_a_block_variable#DispatchA#class.3cc12dd22127281b8293b7c046d21bb2_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
"dispatch_a_block_variable#DispatchA#class.3cc12dd22127281b8293b7c046d21bb2_4" [label="4: Call n$14 \n n$14=*&initialization_block__:_fn_(*) [line 51, column 33]\n n$15=n$14() [line 51, column 3]\n " shape="box"]
Adding endtoend test that shows that dispatch once works well Reviewed By: ddino Differential Revision: D2570191 fb-gh-sync-id: 01be61c
9 years ago
[clang] Change procname file naming scheme Summary: Procnames files are now reversed qualifier lists with `#` as separator (instead of `::` which needs to be escaped in bash). Because of the mechanism that is used to obtain qualifiers, it also affects naming for ObjC classes. Examples: ``` std::unique_ptr<int>::get -> get#unique_ptr<int>#std#__MANGLED,...__ // C++ method folly::split -> split#folly#__MANGLED,..._ // function within namespace NSNumber numberWithBool: -> numberWithBool:#NSNumber#class // ObjC method ``` Reviewed By: jvillard Differential Revision: D4689701 fbshipit-source-id: c3acfc6
8 years ago
"dispatch_a_block_variable#DispatchA#class.3cc12dd22127281b8293b7c046d21bb2_4" -> "dispatch_a_block_variable#DispatchA#class.3cc12dd22127281b8293b7c046d21bb2_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
"dispatch_a_block_variable#DispatchA#class.3cc12dd22127281b8293b7c046d21bb2_5" [label="5: DeclStmt \n DECLARE_LOCALS(&__objc_anonymous_block_DispatchA_dispatch_a_block_variable______3); [line 47, column 38]\n n$18=_fun___objc_alloc_no_fail(sizeof(t=__objc_anonymous_block_DispatchA_dispatch_a_block_variable______3):unsigned long) [line 47, column 38]\n *&__objc_anonymous_block_DispatchA_dispatch_a_block_variable______3:__objc_anonymous_block_DispatchA_dispatch_a_block_variable______3=n$18 [line 47, column 38]\n n$19=*&#GB<codetoanalyze/objc/shared/block/dispatch.m>$DispatchA_dispatch_a_block_variable_static_storage__:objc_object* [line 47, column 38]\n *n$18.DispatchA_dispatch_a_block_variable_static_storage__:objc_object*=n$19 [line 47, column 38]\n *&initialization_block__:_fn_(*)=(_fun___objc_anonymous_block_DispatchA_dispatch_a_block_variable______3) [line 47, column 3]\n " shape="box"]
initial synchronization
10 years ago
[clang] Change procname file naming scheme Summary: Procnames files are now reversed qualifier lists with `#` as separator (instead of `::` which needs to be escaped in bash). Because of the mechanism that is used to obtain qualifiers, it also affects naming for ObjC classes. Examples: ``` std::unique_ptr<int>::get -> get#unique_ptr<int>#std#__MANGLED,...__ // C++ method folly::split -> split#folly#__MANGLED,..._ // function within namespace NSNumber numberWithBool: -> numberWithBool:#NSNumber#class // ObjC method ``` Reviewed By: jvillard Differential Revision: D4689701 fbshipit-source-id: c3acfc6
8 years ago
"dispatch_a_block_variable#DispatchA#class.3cc12dd22127281b8293b7c046d21bb2_5" -> "dispatch_a_block_variable#DispatchA#class.3cc12dd22127281b8293b7c046d21bb2_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
"__objc_anonymous_block_DispatchA_dispatch_a_block_variable______3.9c4c8eed871dc8fb1938edcd3d194533_1" [label="1: Start __objc_anonymous_block_DispatchA_dispatch_a_block_variable______3\nFormals: \nLocals: \n DECLARE_LOCALS(&return); [line 47, column 38]\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
"__objc_anonymous_block_DispatchA_dispatch_a_block_variable______3.9c4c8eed871dc8fb1938edcd3d194533_1" -> "__objc_anonymous_block_DispatchA_dispatch_a_block_variable______3.9c4c8eed871dc8fb1938edcd3d194533_3" ;
"__objc_anonymous_block_DispatchA_dispatch_a_block_variable______3.9c4c8eed871dc8fb1938edcd3d194533_2" [label="2: Exit __objc_anonymous_block_DispatchA_dispatch_a_block_variable______3 \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
"__objc_anonymous_block_DispatchA_dispatch_a_block_variable______3.9c4c8eed871dc8fb1938edcd3d194533_3" [label="3: BinaryOperatorStmt: Assign \n n$16=_fun___objc_alloc_no_fail(sizeof(t=DispatchA):unsigned long) [line 48, column 25]\n n$17=_fun_DispatchA_init(n$16:DispatchA*) virtual [line 48, column 25]\n *&#GB<codetoanalyze/objc/shared/block/dispatch.m>$DispatchA_dispatch_a_block_variable_static_storage__:objc_object*=n$17 [line 48, 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
"__objc_anonymous_block_DispatchA_dispatch_a_block_variable______3.9c4c8eed871dc8fb1938edcd3d194533_3" -> "__objc_anonymous_block_DispatchA_dispatch_a_block_variable______3.9c4c8eed871dc8fb1938edcd3d194533_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
"dispatch_a_block_variable_from_macro#DispatchA#class.92567a38d5ab3cf637f72030b1097441_1" [label="1: Start DispatchA_dispatch_a_block_variable_from_macro\nFormals: \nLocals: initialization_block__:_fn_(*) \n DECLARE_LOCALS(&return,&initialization_block__); [line 55, column 1]\n " color=yellow style=filled]
[tests] Prepare objective-c tests for conversion to the new direct testing format Summary: Rename symbols in test files so they are not duplicated and files can be analyzed together without affecting analysis results. Fix some compilation errors, where files could be analyzed but would fail direct compilation. Add Makefile mimicking the same analysis parameters used for the existing tests. Reviewed By: dulmarod Differential Revision: D3869993 fbshipit-source-id: 6db1baf
8 years ago
[clang] Fix expression statement with conditional inside Summary: CFG nodes were not connected and some instructions ended up in wrong place. Fix those issues Reviewed By: dulmarod Differential Revision: D5406720 fbshipit-source-id: 2a70e1a
7 years ago
"dispatch_a_block_variable_from_macro#DispatchA#class.92567a38d5ab3cf637f72030b1097441_1" -> "dispatch_a_block_variable_from_macro#DispatchA#class.92567a38d5ab3cf637f72030b1097441_5" ;
[clang] Change procname file naming scheme Summary: Procnames files are now reversed qualifier lists with `#` as separator (instead of `::` which needs to be escaped in bash). Because of the mechanism that is used to obtain qualifiers, it also affects naming for ObjC classes. Examples: ``` std::unique_ptr<int>::get -> get#unique_ptr<int>#std#__MANGLED,...__ // C++ method folly::split -> split#folly#__MANGLED,..._ // function within namespace NSNumber numberWithBool: -> numberWithBool:#NSNumber#class // ObjC method ``` Reviewed By: jvillard Differential Revision: D4689701 fbshipit-source-id: c3acfc6
8 years ago
"dispatch_a_block_variable_from_macro#DispatchA#class.92567a38d5ab3cf637f72030b1097441_2" [label="2: Exit DispatchA_dispatch_a_block_variable_from_macro \n " color=yellow style=filled]
[tests] Prepare objective-c tests for conversion to the new direct testing format Summary: Rename symbols in test files so they are not duplicated and files can be analyzed together without affecting analysis results. Fix some compilation errors, where files could be analyzed but would fail direct compilation. Add Makefile mimicking the same analysis parameters used for the existing tests. Reviewed By: dulmarod Differential Revision: D3869993 fbshipit-source-id: 6db1baf
8 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
"dispatch_a_block_variable_from_macro#DispatchA#class.92567a38d5ab3cf637f72030b1097441_3" [label="3: Fallback node \n n$20=*&#GB<codetoanalyze/objc/shared/block/dispatch.m>$DispatchA_dispatch_a_block_variable_from_macro_static_storage__:objc_object* [line 63, column 5]\n " shape="box"]
[tests] Prepare objective-c tests for conversion to the new direct testing format Summary: Rename symbols in test files so they are not duplicated and files can be analyzed together without affecting analysis results. Fix some compilation errors, where files could be analyzed but would fail direct compilation. Add Makefile mimicking the same analysis parameters used for the existing tests. Reviewed By: dulmarod Differential Revision: D3869993 fbshipit-source-id: 6db1baf
8 years ago
[clang] Fix expression statement with conditional inside Summary: CFG nodes were not connected and some instructions ended up in wrong place. Fix those issues Reviewed By: dulmarod Differential Revision: D5406720 fbshipit-source-id: 2a70e1a
7 years ago
"dispatch_a_block_variable_from_macro#DispatchA#class.92567a38d5ab3cf637f72030b1097441_3" -> "dispatch_a_block_variable_from_macro#DispatchA#class.92567a38d5ab3cf637f72030b1097441_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
"dispatch_a_block_variable_from_macro#DispatchA#class.92567a38d5ab3cf637f72030b1097441_4" [label="4: Call n$21 \n n$21=*&initialization_block__:_fn_(*) [line 62, column 35]\n n$22=n$21() [line 62, column 5]\n " shape="box"]
[clang] Fix expression statement with conditional inside Summary: CFG nodes were not connected and some instructions ended up in wrong place. Fix those issues Reviewed By: dulmarod Differential Revision: D5406720 fbshipit-source-id: 2a70e1a
7 years ago
"dispatch_a_block_variable_from_macro#DispatchA#class.92567a38d5ab3cf637f72030b1097441_4" -> "dispatch_a_block_variable_from_macro#DispatchA#class.92567a38d5ab3cf637f72030b1097441_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
"dispatch_a_block_variable_from_macro#DispatchA#class.92567a38d5ab3cf637f72030b1097441_5" [label="5: DeclStmt \n DECLARE_LOCALS(&__objc_anonymous_block_DispatchA_dispatch_a_block_variable_from_macro______4); [line 58, column 40]\n n$25=_fun___objc_alloc_no_fail(sizeof(t=__objc_anonymous_block_DispatchA_dispatch_a_block_variable_from_macro______4):unsigned long) [line 58, column 40]\n *&__objc_anonymous_block_DispatchA_dispatch_a_block_variable_from_macro______4:__objc_anonymous_block_DispatchA_dispatch_a_block_variable_from_macro______4=n$25 [line 58, column 40]\n n$26=*&#GB<codetoanalyze/objc/shared/block/dispatch.m>$DispatchA_dispatch_a_block_variable_from_macro_static_storage__:objc_object* [line 58, column 40]\n *n$25.DispatchA_dispatch_a_block_variable_from_macro_static_storage__:objc_object*=n$26 [line 58, column 40]\n *&initialization_block__:_fn_(*)=(_fun___objc_anonymous_block_DispatchA_dispatch_a_block_variable_from_macro______4) [line 58, column 5]\n " shape="box"]
[clang] Fix expression statement with conditional inside Summary: CFG nodes were not connected and some instructions ended up in wrong place. Fix those issues Reviewed By: dulmarod Differential Revision: D5406720 fbshipit-source-id: 2a70e1a
7 years ago
"dispatch_a_block_variable_from_macro#DispatchA#class.92567a38d5ab3cf637f72030b1097441_5" -> "dispatch_a_block_variable_from_macro#DispatchA#class.92567a38d5ab3cf637f72030b1097441_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
"dispatch_a_block_variable_from_macro#DispatchA#class.92567a38d5ab3cf637f72030b1097441_6" [label="6: Return Stmt \n *&return:objc_object*=n$20 [line 56, column 3]\n " shape="box"]
[clang] Fix expression statement with conditional inside Summary: CFG nodes were not connected and some instructions ended up in wrong place. Fix those issues Reviewed By: dulmarod Differential Revision: D5406720 fbshipit-source-id: 2a70e1a
7 years ago
"dispatch_a_block_variable_from_macro#DispatchA#class.92567a38d5ab3cf637f72030b1097441_6" -> "dispatch_a_block_variable_from_macro#DispatchA#class.92567a38d5ab3cf637f72030b1097441_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
"__objc_anonymous_block_DispatchA_dispatch_a_block_variable_from_macro______4.82bfd971252ed3dd1cbfd850406db887_1" [label="1: Start __objc_anonymous_block_DispatchA_dispatch_a_block_variable_from_macro______4\nFormals: \nLocals: \n DECLARE_LOCALS(&return); [line 58, column 40]\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
"__objc_anonymous_block_DispatchA_dispatch_a_block_variable_from_macro______4.82bfd971252ed3dd1cbfd850406db887_1" -> "__objc_anonymous_block_DispatchA_dispatch_a_block_variable_from_macro______4.82bfd971252ed3dd1cbfd850406db887_3" ;
"__objc_anonymous_block_DispatchA_dispatch_a_block_variable_from_macro______4.82bfd971252ed3dd1cbfd850406db887_2" [label="2: Exit __objc_anonymous_block_DispatchA_dispatch_a_block_variable_from_macro______4 \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
"__objc_anonymous_block_DispatchA_dispatch_a_block_variable_from_macro______4.82bfd971252ed3dd1cbfd850406db887_3" [label="3: BinaryOperatorStmt: Assign \n n$23=_fun___objc_alloc_no_fail(sizeof(t=DispatchA):unsigned long) [line 59, column 27]\n n$24=_fun_DispatchA_init(n$23:DispatchA*) virtual [line 59, column 27]\n *&#GB<codetoanalyze/objc/shared/block/dispatch.m>$DispatchA_dispatch_a_block_variable_from_macro_static_storage__:objc_object*=n$24 [line 59, 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
"__objc_anonymous_block_DispatchA_dispatch_a_block_variable_from_macro______4.82bfd971252ed3dd1cbfd850406db887_3" -> "__objc_anonymous_block_DispatchA_dispatch_a_block_variable_from_macro______4.82bfd971252ed3dd1cbfd850406db887_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
"dispatch_a_block_variable_from_macro_delivers_initialised_object#DispatchA#class.a58ef5afb5e1e9480b49788e2400c52c_1" [label="1: Start DispatchA_dispatch_a_block_variable_from_macro_delivers_initialised_object\nFormals: \nLocals: a:DispatchA* \n DECLARE_LOCALS(&return,&a); [line 67, column 1]\n " color=yellow style=filled]
initial synchronization
10 years ago
[clang] Change procname file naming scheme Summary: Procnames files are now reversed qualifier lists with `#` as separator (instead of `::` which needs to be escaped in bash). Because of the mechanism that is used to obtain qualifiers, it also affects naming for ObjC classes. Examples: ``` std::unique_ptr<int>::get -> get#unique_ptr<int>#std#__MANGLED,...__ // C++ method folly::split -> split#folly#__MANGLED,..._ // function within namespace NSNumber numberWithBool: -> numberWithBool:#NSNumber#class // ObjC method ``` Reviewed By: jvillard Differential Revision: D4689701 fbshipit-source-id: c3acfc6
8 years ago
"dispatch_a_block_variable_from_macro_delivers_initialised_object#DispatchA#class.a58ef5afb5e1e9480b49788e2400c52c_1" -> "dispatch_a_block_variable_from_macro_delivers_initialised_object#DispatchA#class.a58ef5afb5e1e9480b49788e2400c52c_5" ;
"dispatch_a_block_variable_from_macro_delivers_initialised_object#DispatchA#class.a58ef5afb5e1e9480b49788e2400c52c_2" [label="2: Exit DispatchA_dispatch_a_block_variable_from_macro_delivers_initialised_object \n " color=yellow style=filled]
Fix dyspatch_async calls Reviewed By: akotulski Differential Revision: D3392473 fbshipit-source-id: 37b5b67
9 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
"dispatch_a_block_variable_from_macro_delivers_initialised_object#DispatchA#class.a58ef5afb5e1e9480b49788e2400c52c_3" [label="3: Return Stmt \n n$27=*&a:DispatchA* [line 70, column 15]\n n$28=*n$27._x:int [line 70, column 15]\n *&return:int=(1 / (n$28 - 5)) [line 70, column 3]\n " shape="box"]
Fix dyspatch_async calls Reviewed By: akotulski Differential Revision: D3392473 fbshipit-source-id: 37b5b67
9 years ago
[clang] Change procname file naming scheme Summary: Procnames files are now reversed qualifier lists with `#` as separator (instead of `::` which needs to be escaped in bash). Because of the mechanism that is used to obtain qualifiers, it also affects naming for ObjC classes. Examples: ``` std::unique_ptr<int>::get -> get#unique_ptr<int>#std#__MANGLED,...__ // C++ method folly::split -> split#folly#__MANGLED,..._ // function within namespace NSNumber numberWithBool: -> numberWithBool:#NSNumber#class // ObjC method ``` Reviewed By: jvillard Differential Revision: D4689701 fbshipit-source-id: c3acfc6
8 years ago
"dispatch_a_block_variable_from_macro_delivers_initialised_object#DispatchA#class.a58ef5afb5e1e9480b49788e2400c52c_3" -> "dispatch_a_block_variable_from_macro_delivers_initialised_object#DispatchA#class.a58ef5afb5e1e9480b49788e2400c52c_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
"dispatch_a_block_variable_from_macro_delivers_initialised_object#DispatchA#class.a58ef5afb5e1e9480b49788e2400c52c_4" [label="4: BinaryOperatorStmt: Assign \n n$29=*&a:DispatchA* [line 69, column 3]\n *n$29._x:int=5 [line 69, column 3]\n " shape="box"]
Fix dyspatch_async calls Reviewed By: akotulski Differential Revision: D3392473 fbshipit-source-id: 37b5b67
9 years ago
[clang] Change procname file naming scheme Summary: Procnames files are now reversed qualifier lists with `#` as separator (instead of `::` which needs to be escaped in bash). Because of the mechanism that is used to obtain qualifiers, it also affects naming for ObjC classes. Examples: ``` std::unique_ptr<int>::get -> get#unique_ptr<int>#std#__MANGLED,...__ // C++ method folly::split -> split#folly#__MANGLED,..._ // function within namespace NSNumber numberWithBool: -> numberWithBool:#NSNumber#class // ObjC method ``` Reviewed By: jvillard Differential Revision: D4689701 fbshipit-source-id: c3acfc6
8 years ago
"dispatch_a_block_variable_from_macro_delivers_initialised_object#DispatchA#class.a58ef5afb5e1e9480b49788e2400c52c_4" -> "dispatch_a_block_variable_from_macro_delivers_initialised_object#DispatchA#class.a58ef5afb5e1e9480b49788e2400c52c_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
"dispatch_a_block_variable_from_macro_delivers_initialised_object#DispatchA#class.a58ef5afb5e1e9480b49788e2400c52c_5" [label="5: DeclStmt \n n$30=_fun_DispatchA_dispatch_a_block_variable_from_macro() [line 68, column 18]\n *&a:DispatchA*=n$30 [line 68, column 3]\n " shape="box"]
Fix dyspatch_async calls Reviewed By: akotulski Differential Revision: D3392473 fbshipit-source-id: 37b5b67
9 years ago
[clang] Change procname file naming scheme Summary: Procnames files are now reversed qualifier lists with `#` as separator (instead of `::` which needs to be escaped in bash). Because of the mechanism that is used to obtain qualifiers, it also affects naming for ObjC classes. Examples: ``` std::unique_ptr<int>::get -> get#unique_ptr<int>#std#__MANGLED,...__ // C++ method folly::split -> split#folly#__MANGLED,..._ // function within namespace NSNumber numberWithBool: -> numberWithBool:#NSNumber#class // ObjC method ``` Reviewed By: jvillard Differential Revision: D4689701 fbshipit-source-id: c3acfc6
8 years ago
"dispatch_a_block_variable_from_macro_delivers_initialised_object#DispatchA#class.a58ef5afb5e1e9480b49788e2400c52c_5" -> "dispatch_a_block_variable_from_macro_delivers_initialised_object#DispatchA#class.a58ef5afb5e1e9480b49788e2400c52c_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
"DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_1" [label="1: Start DispatchMain\nFormals: \nLocals: p:int* b:DispatchA* \n DECLARE_LOCALS(&return,&p,&b); [line 75, column 1]\n " color=yellow style=filled]
Fix dyspatch_async calls Reviewed By: akotulski Differential Revision: D3392473 fbshipit-source-id: 37b5b67
9 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
"DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_1" -> "DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_11" ;
"DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_2" [label="2: Exit DispatchMain \n " color=yellow style=filled]
Fix dyspatch_async calls Reviewed By: akotulski Differential Revision: D3392473 fbshipit-source-id: 37b5b67
9 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
"DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_3" [label="3: + \n " ]
Fix dyspatch_async calls Reviewed By: akotulski Differential Revision: D3392473 fbshipit-source-id: 37b5b67
9 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
"DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_3" -> "DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_4" ;
"DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_4" [label="4: between_join_and_exit \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
"DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_4" -> "DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_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
"DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_5" [label="5: BinaryOperatorStmt: EQ \n n$0=*&b:DispatchA* [line 78, 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
"DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_5" -> "DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_6" ;
"DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_5" -> "DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_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
"DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_6" [label="6: Prune (true branch) \n PRUNE((n$0 == null), true); [line 78, 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
"DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_6" -> "DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_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
"DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_7" [label="7: Prune (false branch) \n PRUNE(!(n$0 == null), false); [line 78, 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
"DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_7" -> "DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_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
"DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_8" [label="8: Return Stmt \n n$1=*&p:int* [line 79, column 13]\n n$2=*n$1:int [line 79, column 12]\n *&return:int=n$2 [line 79, 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
"DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_8" -> "DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_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
"DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_9" [label="9: Return Stmt \n *&return:int=0 [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
"DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_9" -> "DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_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
"DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_10" [label="10: DeclStmt \n *&p:int*=null [line 77, 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
"DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_10" -> "DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_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
"DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_11" [label="11: DeclStmt \n n$3=_fun_DispatchA_sharedInstance() [line 76, column 18]\n *&b:DispatchA*=n$3 [line 76, 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
"DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_11" -> "DispatchMain.f6461dbdaeaf9a114cbe40f5f72fbb3f_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
"init#DispatchA#instance.ff6c7b9a5a49bb46493519a4290a6582_1" [label="1: Start DispatchA_init\nFormals: self:DispatchA*\nLocals: \n DECLARE_LOCALS(&return); [line 23, column 1]\n " color=yellow style=filled]
initial synchronization
10 years ago
[clang] Change procname file naming scheme Summary: Procnames files are now reversed qualifier lists with `#` as separator (instead of `::` which needs to be escaped in bash). Because of the mechanism that is used to obtain qualifiers, it also affects naming for ObjC classes. Examples: ``` std::unique_ptr<int>::get -> get#unique_ptr<int>#std#__MANGLED,...__ // C++ method folly::split -> split#folly#__MANGLED,..._ // function within namespace NSNumber numberWithBool: -> numberWithBool:#NSNumber#class // ObjC method ``` Reviewed By: jvillard Differential Revision: D4689701 fbshipit-source-id: c3acfc6
8 years ago
"init#DispatchA#instance.ff6c7b9a5a49bb46493519a4290a6582_1" -> "init#DispatchA#instance.ff6c7b9a5a49bb46493519a4290a6582_3" ;
"init#DispatchA#instance.ff6c7b9a5a49bb46493519a4290a6582_2" [label="2: Exit DispatchA_init \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
"init#DispatchA#instance.ff6c7b9a5a49bb46493519a4290a6582_3" [label="3: Return Stmt \n n$0=*&self:DispatchA* [line 24, column 10]\n *&return:objc_object*=n$0 [line 24, column 3]\n " shape="box"]
initial synchronization
10 years ago
[clang] Change procname file naming scheme Summary: Procnames files are now reversed qualifier lists with `#` as separator (instead of `::` which needs to be escaped in bash). Because of the mechanism that is used to obtain qualifiers, it also affects naming for ObjC classes. Examples: ``` std::unique_ptr<int>::get -> get#unique_ptr<int>#std#__MANGLED,...__ // C++ method folly::split -> split#folly#__MANGLED,..._ // function within namespace NSNumber numberWithBool: -> numberWithBool:#NSNumber#class // ObjC method ``` Reviewed By: jvillard Differential Revision: D4689701 fbshipit-source-id: c3acfc6
8 years ago
"init#DispatchA#instance.ff6c7b9a5a49bb46493519a4290a6582_3" -> "init#DispatchA#instance.ff6c7b9a5a49bb46493519a4290a6582_2" ;
initial synchronization
10 years ago
}