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 '639bda69e7'
${ noResults }
infer_clone/infer/tests/codetoanalyze/cpp/shared/attributes/annotate.cpp.dot

359 lines
31 KiB
Raw Normal View History Unescape

Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
/* @generated */
[cfg] rename iCFG to cfg in dotty files Summary: Not sure what an "iCFG" is but the dotty is only about CFGs anyway. Diff obtained by mass-`sed`. Reviewed By: sblackshear Differential Revision: D6324280 fbshipit-source-id: b7603bb
7 years ago
digraph cfg {
Get rid of Declare_locals Reviewed By: jeremydubreil Differential Revision: D9234331 fbshipit-source-id: 70443eabe
6 years ago
"derefFirstArg#11155233742091168443.dc9be6cfdd866625309f594a0c1a3bdc_1" [label="1: Start derefFirstArg\nFormals: a:int* b:int*\nLocals: \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
"derefFirstArg#11155233742091168443.dc9be6cfdd866625309f594a0c1a3bdc_1" -> "derefFirstArg#11155233742091168443.dc9be6cfdd866625309f594a0c1a3bdc_2" ;
"derefFirstArg#11155233742091168443.dc9be6cfdd866625309f594a0c1a3bdc_2" [label="2: Exit derefFirstArg \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
Get rid of Declare_locals Reviewed By: jeremydubreil Differential Revision: D9234331 fbshipit-source-id: 70443eabe
6 years ago
"derefFirstArg2#11387624487828646016.9777f67ce8b8da5c99a0e59eaaf6eb17_1" [label="1: Start derefFirstArg2\nFormals: a:int* b:int*\nLocals: \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
"derefFirstArg2#11387624487828646016.9777f67ce8b8da5c99a0e59eaaf6eb17_1" -> "derefFirstArg2#11387624487828646016.9777f67ce8b8da5c99a0e59eaaf6eb17_3" ;
"derefFirstArg2#11387624487828646016.9777f67ce8b8da5c99a0e59eaaf6eb17_2" [label="2: Exit derefFirstArg2 \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"derefFirstArg2#11387624487828646016.9777f67ce8b8da5c99a0e59eaaf6eb17_3" [label="3: Return Stmt \n n$0=*&b:int* [line 26, column 11]\n n$1=*n$0:int [line 26, column 10]\n *&return:int=n$1 [line 26, column 3]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
"derefFirstArg2#11387624487828646016.9777f67ce8b8da5c99a0e59eaaf6eb17_3" -> "derefFirstArg2#11387624487828646016.9777f67ce8b8da5c99a0e59eaaf6eb17_2" ;
Get rid of Declare_locals Reviewed By: jeremydubreil Differential Revision: D9234331 fbshipit-source-id: 70443eabe
6 years ago
"derefFirstArg2_null_deref#13631548499595216278.23fca23ff6728e4b72a2548ecb3b1ba0_1" [label="1: Start derefFirstArg2_null_deref\nFormals: \nLocals: a:int \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"derefFirstArg2_null_deref#13631548499595216278.23fca23ff6728e4b72a2548ecb3b1ba0_1" -> "derefFirstArg2_null_deref#13631548499595216278.23fca23ff6728e4b72a2548ecb3b1ba0_4" ;
"derefFirstArg2_null_deref#13631548499595216278.23fca23ff6728e4b72a2548ecb3b1ba0_2" [label="2: Exit derefFirstArg2_null_deref \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"derefFirstArg2_null_deref#13631548499595216278.23fca23ff6728e4b72a2548ecb3b1ba0_3" [label="3: Return Stmt \n n$0=*null:int [line 49, column 10]\n *&return:int=n$0 [line 49, column 3]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"derefFirstArg2_null_deref#13631548499595216278.23fca23ff6728e4b72a2548ecb3b1ba0_3" -> "derefFirstArg2_null_deref#13631548499595216278.23fca23ff6728e4b72a2548ecb3b1ba0_2" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"derefFirstArg2_null_deref#13631548499595216278.23fca23ff6728e4b72a2548ecb3b1ba0_4" [label="4: DeclStmt \n VARIABLE_DECLARED(a:int); [line 48, column 3]\n *&a:int=0 [line 48, column 3]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"derefFirstArg2_null_deref#13631548499595216278.23fca23ff6728e4b72a2548ecb3b1ba0_4" -> "derefFirstArg2_null_deref#13631548499595216278.23fca23ff6728e4b72a2548ecb3b1ba0_3" ;
Get rid of Declare_locals Reviewed By: jeremydubreil Differential Revision: D9234331 fbshipit-source-id: 70443eabe
6 years ago
"derefFirstArg2_ok_deref#6873109919028202465.d57ab0b62c0ba18894b8b08d5a8f8e8a_1" [label="1: Start derefFirstArg2_ok_deref\nFormals: \nLocals: a:int \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"derefFirstArg2_ok_deref#6873109919028202465.d57ab0b62c0ba18894b8b08d5a8f8e8a_1" -> "derefFirstArg2_ok_deref#6873109919028202465.d57ab0b62c0ba18894b8b08d5a8f8e8a_4" ;
"derefFirstArg2_ok_deref#6873109919028202465.d57ab0b62c0ba18894b8b08d5a8f8e8a_2" [label="2: Exit derefFirstArg2_ok_deref \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"derefFirstArg2_ok_deref#6873109919028202465.d57ab0b62c0ba18894b8b08d5a8f8e8a_3" [label="3: Return Stmt \n n$0=*&a:int [line 54, column 10]\n *&return:int=n$0 [line 54, column 3]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"derefFirstArg2_ok_deref#6873109919028202465.d57ab0b62c0ba18894b8b08d5a8f8e8a_3" -> "derefFirstArg2_ok_deref#6873109919028202465.d57ab0b62c0ba18894b8b08d5a8f8e8a_2" ;
[models] get rid of include-based C++ models Summary: These have proved to be too fragile to maintain as they would often break compilation of user code. They have been off by default for more than a year now (D7350715). Removing the include models shows a more accurate picture of what infer results look like in production. As such, lots of tests have changed, mostly biabduction but also in inferbo. SIOF was using include-based models too but now libc++ is better and iostreams are implemented in a way that SIOF understands (instead of being magical creatures) so nothing changed there. Reviewed By: skcho Differential Revision: D16602171 fbshipit-source-id: ce38f045b
5 years ago
"derefFirstArg2_ok_deref#6873109919028202465.d57ab0b62c0ba18894b8b08d5a8f8e8a_4" [label="4: DeclStmt \n VARIABLE_DECLARED(a:int); [line 53, column 3]\n *&a:int=0 [line 53, column 3]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"derefFirstArg2_ok_deref#6873109919028202465.d57ab0b62c0ba18894b8b08d5a8f8e8a_4" -> "derefFirstArg2_ok_deref#6873109919028202465.d57ab0b62c0ba18894b8b08d5a8f8e8a_3" ;
Get rid of Declare_locals Reviewed By: jeremydubreil Differential Revision: D9234331 fbshipit-source-id: 70443eabe
6 years ago
"derefFirstArg3#3150650678378709003.fb38cf6b9238ba2f8f6e25136f8beb95_1" [label="1: Start derefFirstArg3\nFormals: a:int* b:int*\nLocals: \n " color=yellow style=filled]
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
"derefFirstArg3#3150650678378709003.fb38cf6b9238ba2f8f6e25136f8beb95_1" -> "derefFirstArg3#3150650678378709003.fb38cf6b9238ba2f8f6e25136f8beb95_3" ;
"derefFirstArg3#3150650678378709003.fb38cf6b9238ba2f8f6e25136f8beb95_2" [label="2: Exit derefFirstArg3 \n " color=yellow style=filled]
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"derefFirstArg3#3150650678378709003.fb38cf6b9238ba2f8f6e25136f8beb95_3" [label="3: Return Stmt \n n$0=*&b:int* [line 33, column 11]\n n$1=*n$0:int [line 33, column 10]\n *&return:int=n$1 [line 33, column 3]\n " shape="box"]
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
"derefFirstArg3#3150650678378709003.fb38cf6b9238ba2f8f6e25136f8beb95_3" -> "derefFirstArg3#3150650678378709003.fb38cf6b9238ba2f8f6e25136f8beb95_2" ;
Get rid of Declare_locals Reviewed By: jeremydubreil Differential Revision: D9234331 fbshipit-source-id: 70443eabe
6 years ago
"derefFirstArg3_null_deref#3036141491555788229.605788dbf5e3c5625520098d1b5d320e_1" [label="1: Start derefFirstArg3_null_deref\nFormals: \nLocals: a:int \n " color=yellow style=filled]
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
"derefFirstArg3_null_deref#3036141491555788229.605788dbf5e3c5625520098d1b5d320e_1" -> "derefFirstArg3_null_deref#3036141491555788229.605788dbf5e3c5625520098d1b5d320e_4" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"derefFirstArg3_null_deref#3036141491555788229.605788dbf5e3c5625520098d1b5d320e_2" [label="2: Exit derefFirstArg3_null_deref \n " color=yellow style=filled]
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"derefFirstArg3_null_deref#3036141491555788229.605788dbf5e3c5625520098d1b5d320e_3" [label="3: Return Stmt \n n$0=_fun_derefFirstArg3(&a:int*,null:int*) [line 64, column 10]\n *&return:int=n$0 [line 64, column 3]\n " shape="box"]
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
"derefFirstArg3_null_deref#3036141491555788229.605788dbf5e3c5625520098d1b5d320e_3" -> "derefFirstArg3_null_deref#3036141491555788229.605788dbf5e3c5625520098d1b5d320e_2" ;
[models] get rid of include-based C++ models Summary: These have proved to be too fragile to maintain as they would often break compilation of user code. They have been off by default for more than a year now (D7350715). Removing the include models shows a more accurate picture of what infer results look like in production. As such, lots of tests have changed, mostly biabduction but also in inferbo. SIOF was using include-based models too but now libc++ is better and iostreams are implemented in a way that SIOF understands (instead of being magical creatures) so nothing changed there. Reviewed By: skcho Differential Revision: D16602171 fbshipit-source-id: ce38f045b
5 years ago
"derefFirstArg3_null_deref#3036141491555788229.605788dbf5e3c5625520098d1b5d320e_4" [label="4: DeclStmt \n VARIABLE_DECLARED(a:int); [line 63, column 3]\n *&a:int=0 [line 63, column 3]\n " shape="box"]
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
"derefFirstArg3_null_deref#3036141491555788229.605788dbf5e3c5625520098d1b5d320e_4" -> "derefFirstArg3_null_deref#3036141491555788229.605788dbf5e3c5625520098d1b5d320e_3" ;
Get rid of Declare_locals Reviewed By: jeremydubreil Differential Revision: D9234331 fbshipit-source-id: 70443eabe
6 years ago
"derefFirstArg3_ok_deref#12266654054137171150.c58c85ea4ba2ebfd89d0336e51301e7a_1" [label="1: Start derefFirstArg3_ok_deref\nFormals: \nLocals: a:int \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"derefFirstArg3_ok_deref#12266654054137171150.c58c85ea4ba2ebfd89d0336e51301e7a_1" -> "derefFirstArg3_ok_deref#12266654054137171150.c58c85ea4ba2ebfd89d0336e51301e7a_4" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"derefFirstArg3_ok_deref#12266654054137171150.c58c85ea4ba2ebfd89d0336e51301e7a_2" [label="2: Exit derefFirstArg3_ok_deref \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"derefFirstArg3_ok_deref#12266654054137171150.c58c85ea4ba2ebfd89d0336e51301e7a_3" [label="3: Return Stmt \n n$0=_fun_derefFirstArg3(null:int*,&a:int*) [line 59, column 10]\n *&return:int=n$0 [line 59, column 3]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"derefFirstArg3_ok_deref#12266654054137171150.c58c85ea4ba2ebfd89d0336e51301e7a_3" -> "derefFirstArg3_ok_deref#12266654054137171150.c58c85ea4ba2ebfd89d0336e51301e7a_2" ;
[models] get rid of include-based C++ models Summary: These have proved to be too fragile to maintain as they would often break compilation of user code. They have been off by default for more than a year now (D7350715). Removing the include models shows a more accurate picture of what infer results look like in production. As such, lots of tests have changed, mostly biabduction but also in inferbo. SIOF was using include-based models too but now libc++ is better and iostreams are implemented in a way that SIOF understands (instead of being magical creatures) so nothing changed there. Reviewed By: skcho Differential Revision: D16602171 fbshipit-source-id: ce38f045b
5 years ago
"derefFirstArg3_ok_deref#12266654054137171150.c58c85ea4ba2ebfd89d0336e51301e7a_4" [label="4: DeclStmt \n VARIABLE_DECLARED(a:int); [line 58, column 3]\n *&a:int=0 [line 58, column 3]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"derefFirstArg3_ok_deref#12266654054137171150.c58c85ea4ba2ebfd89d0336e51301e7a_4" -> "derefFirstArg3_ok_deref#12266654054137171150.c58c85ea4ba2ebfd89d0336e51301e7a_3" ;
Get rid of Declare_locals Reviewed By: jeremydubreil Differential Revision: D9234331 fbshipit-source-id: 70443eabe
6 years ago
"derefFirstArg_null_deref#14830687999166111591.325df3347d8f75d0292cfd33a485d28a_1" [label="1: Start derefFirstArg_null_deref\nFormals: \nLocals: a:int \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
"derefFirstArg_null_deref#14830687999166111591.325df3347d8f75d0292cfd33a485d28a_1" -> "derefFirstArg_null_deref#14830687999166111591.325df3347d8f75d0292cfd33a485d28a_4" ;
"derefFirstArg_null_deref#14830687999166111591.325df3347d8f75d0292cfd33a485d28a_2" [label="2: Exit derefFirstArg_null_deref \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"derefFirstArg_null_deref#14830687999166111591.325df3347d8f75d0292cfd33a485d28a_3" [label="3: Return Stmt \n n$0=*null:int [line 39, column 10]\n *&return:int=n$0 [line 39, column 3]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
"derefFirstArg_null_deref#14830687999166111591.325df3347d8f75d0292cfd33a485d28a_3" -> "derefFirstArg_null_deref#14830687999166111591.325df3347d8f75d0292cfd33a485d28a_2" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"derefFirstArg_null_deref#14830687999166111591.325df3347d8f75d0292cfd33a485d28a_4" [label="4: DeclStmt \n VARIABLE_DECLARED(a:int); [line 38, column 3]\n *&a:int=0 [line 38, column 3]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
"derefFirstArg_null_deref#14830687999166111591.325df3347d8f75d0292cfd33a485d28a_4" -> "derefFirstArg_null_deref#14830687999166111591.325df3347d8f75d0292cfd33a485d28a_3" ;
Get rid of Declare_locals Reviewed By: jeremydubreil Differential Revision: D9234331 fbshipit-source-id: 70443eabe
6 years ago
"derefFirstArg_ok_deref#70986049112502156.78efafe2cdade07d4257a7cd671e75f5_1" [label="1: Start derefFirstArg_ok_deref\nFormals: \nLocals: a:int \n " color=yellow style=filled]
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
"derefFirstArg_ok_deref#70986049112502156.78efafe2cdade07d4257a7cd671e75f5_1" -> "derefFirstArg_ok_deref#70986049112502156.78efafe2cdade07d4257a7cd671e75f5_4" ;
"derefFirstArg_ok_deref#70986049112502156.78efafe2cdade07d4257a7cd671e75f5_2" [label="2: Exit derefFirstArg_ok_deref \n " color=yellow style=filled]
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"derefFirstArg_ok_deref#70986049112502156.78efafe2cdade07d4257a7cd671e75f5_3" [label="3: Return Stmt \n n$0=*&a:int [line 44, column 10]\n *&return:int=n$0 [line 44, column 3]\n " shape="box"]
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
"derefFirstArg_ok_deref#70986049112502156.78efafe2cdade07d4257a7cd671e75f5_3" -> "derefFirstArg_ok_deref#70986049112502156.78efafe2cdade07d4257a7cd671e75f5_2" ;
[models] get rid of include-based C++ models Summary: These have proved to be too fragile to maintain as they would often break compilation of user code. They have been off by default for more than a year now (D7350715). Removing the include models shows a more accurate picture of what infer results look like in production. As such, lots of tests have changed, mostly biabduction but also in inferbo. SIOF was using include-based models too but now libc++ is better and iostreams are implemented in a way that SIOF understands (instead of being magical creatures) so nothing changed there. Reviewed By: skcho Differential Revision: D16602171 fbshipit-source-id: ce38f045b
5 years ago
"derefFirstArg_ok_deref#70986049112502156.78efafe2cdade07d4257a7cd671e75f5_4" [label="4: DeclStmt \n VARIABLE_DECLARED(a:int); [line 43, column 3]\n *&a:int=0 [line 43, column 3]\n " shape="box"]
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
"derefFirstArg_ok_deref#70986049112502156.78efafe2cdade07d4257a7cd671e75f5_4" -> "derefFirstArg_ok_deref#70986049112502156.78efafe2cdade07d4257a7cd671e75f5_3" ;
Get rid of Declare_locals Reviewed By: jeremydubreil Differential Revision: D9234331 fbshipit-source-id: 70443eabe
6 years ago
"getPtr_null_deref1#10685326586135592861.d05a7735c36f759fec001951cdc51035_1" [label="1: Start getPtr_null_deref1\nFormals: \nLocals: t:int* \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"getPtr_null_deref1#10685326586135592861.d05a7735c36f759fec001951cdc51035_1" -> "getPtr_null_deref1#10685326586135592861.d05a7735c36f759fec001951cdc51035_5" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"getPtr_null_deref1#10685326586135592861.d05a7735c36f759fec001951cdc51035_2" [label="2: Exit getPtr_null_deref1 \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"getPtr_null_deref1#10685326586135592861.d05a7735c36f759fec001951cdc51035_3" [label="3: Return Stmt \n _=*&t:int* [line 92, column 11]\n n$1=*&t:int* [line 92, column 11]\n n$2=*n$1:int [line 92, column 10]\n *&return:int=n$2 [line 92, column 3]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"getPtr_null_deref1#10685326586135592861.d05a7735c36f759fec001951cdc51035_3" -> "getPtr_null_deref1#10685326586135592861.d05a7735c36f759fec001951cdc51035_2" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"getPtr_null_deref1#10685326586135592861.d05a7735c36f759fec001951cdc51035_4" [label="4: Call _fun_TranslateAsPtr<int>::setPtr \n _=*&t:int* [line 91, column 3]\n n$4=_fun_TranslateAsPtr<int>::setPtr(&t:int*&,null:int*) [line 91, column 3]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"getPtr_null_deref1#10685326586135592861.d05a7735c36f759fec001951cdc51035_4" -> "getPtr_null_deref1#10685326586135592861.d05a7735c36f759fec001951cdc51035_3" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"getPtr_null_deref1#10685326586135592861.d05a7735c36f759fec001951cdc51035_5" [label="5: DeclStmt \n VARIABLE_DECLARED(t:int*); [line 90, column 3]\n n$5=_fun_TranslateAsPtr<int>::TranslateAsPtr(&t:int**,null:int*) [line 90, column 23]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"getPtr_null_deref1#10685326586135592861.d05a7735c36f759fec001951cdc51035_5" -> "getPtr_null_deref1#10685326586135592861.d05a7735c36f759fec001951cdc51035_4" ;
Get rid of Declare_locals Reviewed By: jeremydubreil Differential Revision: D9234331 fbshipit-source-id: 70443eabe
6 years ago
"getPtr_null_deref2#10682492045158632578.de31216813faa493761802feb6f997f2_1" [label="1: Start getPtr_null_deref2\nFormals: \nLocals: t:int* \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"getPtr_null_deref2#10682492045158632578.de31216813faa493761802feb6f997f2_1" -> "getPtr_null_deref2#10682492045158632578.de31216813faa493761802feb6f997f2_5" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"getPtr_null_deref2#10682492045158632578.de31216813faa493761802feb6f997f2_2" [label="2: Exit getPtr_null_deref2 \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"getPtr_null_deref2#10682492045158632578.de31216813faa493761802feb6f997f2_3" [label="3: Return Stmt \n _=*&t:int* [line 98, column 11]\n n$1=*&t:int* [line 98, column 11]\n n$2=*n$1:int [line 98, column 10]\n *&return:int=n$2 [line 98, column 3]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"getPtr_null_deref2#10682492045158632578.de31216813faa493761802feb6f997f2_3" -> "getPtr_null_deref2#10682492045158632578.de31216813faa493761802feb6f997f2_2" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"getPtr_null_deref2#10682492045158632578.de31216813faa493761802feb6f997f2_4" [label="4: Call _fun_TranslateAsPtr<int>::setPtr \n _=*&t:int* [line 97, column 3]\n n$4=_fun_TranslateAsPtr<int>::setPtr(&t:int*&,null:int*) [line 97, column 3]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"getPtr_null_deref2#10682492045158632578.de31216813faa493761802feb6f997f2_4" -> "getPtr_null_deref2#10682492045158632578.de31216813faa493761802feb6f997f2_3" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"getPtr_null_deref2#10682492045158632578.de31216813faa493761802feb6f997f2_5" [label="5: DeclStmt \n VARIABLE_DECLARED(t:int*); [line 96, column 3]\n n$5=_fun_TranslateAsPtr<int>::TranslateAsPtr(&t:int**,null:int*) [line 96, column 23]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"getPtr_null_deref2#10682492045158632578.de31216813faa493761802feb6f997f2_5" -> "getPtr_null_deref2#10682492045158632578.de31216813faa493761802feb6f997f2_4" ;
Get rid of Declare_locals Reviewed By: jeremydubreil Differential Revision: D9234331 fbshipit-source-id: 70443eabe
6 years ago
"getPtr_ok_deref#15608473391071478730.49e56fac5bd82269c2093a9c1e438200_1" [label="1: Start getPtr_ok_deref\nFormals: \nLocals: t:int* a:int \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"getPtr_ok_deref#15608473391071478730.49e56fac5bd82269c2093a9c1e438200_1" -> "getPtr_ok_deref#15608473391071478730.49e56fac5bd82269c2093a9c1e438200_6" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"getPtr_ok_deref#15608473391071478730.49e56fac5bd82269c2093a9c1e438200_2" [label="2: Exit getPtr_ok_deref \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"getPtr_ok_deref#15608473391071478730.49e56fac5bd82269c2093a9c1e438200_3" [label="3: Return Stmt \n _=*&t:int* [line 105, column 11]\n n$1=*&t:int* [line 105, column 11]\n n$2=*n$1:int [line 105, column 10]\n *&return:int=n$2 [line 105, column 3]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"getPtr_ok_deref#15608473391071478730.49e56fac5bd82269c2093a9c1e438200_3" -> "getPtr_ok_deref#15608473391071478730.49e56fac5bd82269c2093a9c1e438200_2" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"getPtr_ok_deref#15608473391071478730.49e56fac5bd82269c2093a9c1e438200_4" [label="4: Call _fun_TranslateAsPtr<int>::setPtr \n _=*&t:int* [line 104, column 3]\n n$4=_fun_TranslateAsPtr<int>::setPtr(&t:int*&,&a:int*) [line 104, column 3]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"getPtr_ok_deref#15608473391071478730.49e56fac5bd82269c2093a9c1e438200_4" -> "getPtr_ok_deref#15608473391071478730.49e56fac5bd82269c2093a9c1e438200_3" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"getPtr_ok_deref#15608473391071478730.49e56fac5bd82269c2093a9c1e438200_5" [label="5: DeclStmt \n VARIABLE_DECLARED(t:int*); [line 103, column 3]\n n$5=_fun_TranslateAsPtr<int>::TranslateAsPtr(&t:int**,null:int*) [line 103, column 23]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"getPtr_ok_deref#15608473391071478730.49e56fac5bd82269c2093a9c1e438200_5" -> "getPtr_ok_deref#15608473391071478730.49e56fac5bd82269c2093a9c1e438200_4" ;
[models] get rid of include-based C++ models Summary: These have proved to be too fragile to maintain as they would often break compilation of user code. They have been off by default for more than a year now (D7350715). Removing the include models shows a more accurate picture of what infer results look like in production. As such, lots of tests have changed, mostly biabduction but also in inferbo. SIOF was using include-based models too but now libc++ is better and iostreams are implemented in a way that SIOF understands (instead of being magical creatures) so nothing changed there. Reviewed By: skcho Differential Revision: D16602171 fbshipit-source-id: ce38f045b
5 years ago
"getPtr_ok_deref#15608473391071478730.49e56fac5bd82269c2093a9c1e438200_6" [label="6: DeclStmt \n VARIABLE_DECLARED(a:int); [line 102, column 3]\n *&a:int=0 [line 102, column 3]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"getPtr_ok_deref#15608473391071478730.49e56fac5bd82269c2093a9c1e438200_6" -> "getPtr_ok_deref#15608473391071478730.49e56fac5bd82269c2093a9c1e438200_5" ;
Get rid of Declare_locals Reviewed By: jeremydubreil Differential Revision: D9234331 fbshipit-source-id: 70443eabe
6 years ago
"getRef_null_deref1#4264296374417396044.654d24b6c4af017d90a5ceff83c121c2_1" [label="1: Start getRef_null_deref1\nFormals: \nLocals: t:int* \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"getRef_null_deref1#4264296374417396044.654d24b6c4af017d90a5ceff83c121c2_1" -> "getRef_null_deref1#4264296374417396044.654d24b6c4af017d90a5ceff83c121c2_5" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"getRef_null_deref1#4264296374417396044.654d24b6c4af017d90a5ceff83c121c2_2" [label="2: Exit getRef_null_deref1 \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"getRef_null_deref1#4264296374417396044.654d24b6c4af017d90a5ceff83c121c2_3" [label="3: Return Stmt \n _=*&t:int* [line 130, column 10]\n n$1=*&t:int* [line 130, column 10]\n n$2=*n$1:int [line 130, column 10]\n *&return:int=n$2 [line 130, column 3]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"getRef_null_deref1#4264296374417396044.654d24b6c4af017d90a5ceff83c121c2_3" -> "getRef_null_deref1#4264296374417396044.654d24b6c4af017d90a5ceff83c121c2_2" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"getRef_null_deref1#4264296374417396044.654d24b6c4af017d90a5ceff83c121c2_4" [label="4: Call _fun_TranslateAsPtr<int>::setPtr \n _=*&t:int* [line 129, column 3]\n n$4=_fun_TranslateAsPtr<int>::setPtr(&t:int*&,null:int*) [line 129, column 3]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"getRef_null_deref1#4264296374417396044.654d24b6c4af017d90a5ceff83c121c2_4" -> "getRef_null_deref1#4264296374417396044.654d24b6c4af017d90a5ceff83c121c2_3" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"getRef_null_deref1#4264296374417396044.654d24b6c4af017d90a5ceff83c121c2_5" [label="5: DeclStmt \n VARIABLE_DECLARED(t:int*); [line 128, column 3]\n n$5=_fun_TranslateAsPtr<int>::TranslateAsPtr(&t:int**,null:int*) [line 128, column 23]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"getRef_null_deref1#4264296374417396044.654d24b6c4af017d90a5ceff83c121c2_5" -> "getRef_null_deref1#4264296374417396044.654d24b6c4af017d90a5ceff83c121c2_4" ;
Get rid of Declare_locals Reviewed By: jeremydubreil Differential Revision: D9234331 fbshipit-source-id: 70443eabe
6 years ago
"getRef_null_deref2#4263471740696427019.45bed1239309132cabf29f4cdd81f3cc_1" [label="1: Start getRef_null_deref2\nFormals: \nLocals: t:int* \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"getRef_null_deref2#4263471740696427019.45bed1239309132cabf29f4cdd81f3cc_1" -> "getRef_null_deref2#4263471740696427019.45bed1239309132cabf29f4cdd81f3cc_5" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"getRef_null_deref2#4263471740696427019.45bed1239309132cabf29f4cdd81f3cc_2" [label="2: Exit getRef_null_deref2 \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"getRef_null_deref2#4263471740696427019.45bed1239309132cabf29f4cdd81f3cc_3" [label="3: Return Stmt \n _=*&t:int* [line 136, column 10]\n n$1=*&t:int* [line 136, column 10]\n n$2=*n$1:int [line 136, column 10]\n *&return:int=n$2 [line 136, column 3]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"getRef_null_deref2#4263471740696427019.45bed1239309132cabf29f4cdd81f3cc_3" -> "getRef_null_deref2#4263471740696427019.45bed1239309132cabf29f4cdd81f3cc_2" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"getRef_null_deref2#4263471740696427019.45bed1239309132cabf29f4cdd81f3cc_4" [label="4: Call _fun_TranslateAsPtr<int>::setPtr \n _=*&t:int* [line 135, column 3]\n n$4=_fun_TranslateAsPtr<int>::setPtr(&t:int*&,null:int*) [line 135, column 3]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"getRef_null_deref2#4263471740696427019.45bed1239309132cabf29f4cdd81f3cc_4" -> "getRef_null_deref2#4263471740696427019.45bed1239309132cabf29f4cdd81f3cc_3" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"getRef_null_deref2#4263471740696427019.45bed1239309132cabf29f4cdd81f3cc_5" [label="5: DeclStmt \n VARIABLE_DECLARED(t:int*); [line 134, column 3]\n n$5=_fun_TranslateAsPtr<int>::TranslateAsPtr(&t:int**,null:int*) [line 134, column 23]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"getRef_null_deref2#4263471740696427019.45bed1239309132cabf29f4cdd81f3cc_5" -> "getRef_null_deref2#4263471740696427019.45bed1239309132cabf29f4cdd81f3cc_4" ;
Get rid of Declare_locals Reviewed By: jeremydubreil Differential Revision: D9234331 fbshipit-source-id: 70443eabe
6 years ago
"getRef_ok_deref#10111201054364386601.e514c65ac6978a31376e6032d81b3d16_1" [label="1: Start getRef_ok_deref\nFormals: \nLocals: t:int* a:int \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"getRef_ok_deref#10111201054364386601.e514c65ac6978a31376e6032d81b3d16_1" -> "getRef_ok_deref#10111201054364386601.e514c65ac6978a31376e6032d81b3d16_6" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"getRef_ok_deref#10111201054364386601.e514c65ac6978a31376e6032d81b3d16_2" [label="2: Exit getRef_ok_deref \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"getRef_ok_deref#10111201054364386601.e514c65ac6978a31376e6032d81b3d16_3" [label="3: Return Stmt \n _=*&t:int* [line 143, column 10]\n n$1=*&t:int* [line 143, column 10]\n n$2=*n$1:int [line 143, column 10]\n *&return:int=n$2 [line 143, column 3]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"getRef_ok_deref#10111201054364386601.e514c65ac6978a31376e6032d81b3d16_3" -> "getRef_ok_deref#10111201054364386601.e514c65ac6978a31376e6032d81b3d16_2" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"getRef_ok_deref#10111201054364386601.e514c65ac6978a31376e6032d81b3d16_4" [label="4: Call _fun_TranslateAsPtr<int>::setPtr \n _=*&t:int* [line 142, column 3]\n n$4=_fun_TranslateAsPtr<int>::setPtr(&t:int*&,&a:int*) [line 142, column 3]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"getRef_ok_deref#10111201054364386601.e514c65ac6978a31376e6032d81b3d16_4" -> "getRef_ok_deref#10111201054364386601.e514c65ac6978a31376e6032d81b3d16_3" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"getRef_ok_deref#10111201054364386601.e514c65ac6978a31376e6032d81b3d16_5" [label="5: DeclStmt \n VARIABLE_DECLARED(t:int*); [line 141, column 3]\n n$5=_fun_TranslateAsPtr<int>::TranslateAsPtr(&t:int**,null:int*) [line 141, column 23]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"getRef_ok_deref#10111201054364386601.e514c65ac6978a31376e6032d81b3d16_5" -> "getRef_ok_deref#10111201054364386601.e514c65ac6978a31376e6032d81b3d16_4" ;
[models] get rid of include-based C++ models Summary: These have proved to be too fragile to maintain as they would often break compilation of user code. They have been off by default for more than a year now (D7350715). Removing the include models shows a more accurate picture of what infer results look like in production. As such, lots of tests have changed, mostly biabduction but also in inferbo. SIOF was using include-based models too but now libc++ is better and iostreams are implemented in a way that SIOF understands (instead of being magical creatures) so nothing changed there. Reviewed By: skcho Differential Revision: D16602171 fbshipit-source-id: ce38f045b
5 years ago
"getRef_ok_deref#10111201054364386601.e514c65ac6978a31376e6032d81b3d16_6" [label="6: DeclStmt \n VARIABLE_DECLARED(a:int); [line 140, column 3]\n *&a:int=0 [line 140, column 3]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"getRef_ok_deref#10111201054364386601.e514c65ac6978a31376e6032d81b3d16_6" -> "getRef_ok_deref#10111201054364386601.e514c65ac6978a31376e6032d81b3d16_5" ;
Get rid of Declare_locals Reviewed By: jeremydubreil Differential Revision: D9234331 fbshipit-source-id: 70443eabe
6 years ago
"operator_star_null_deref1#14187169119337849630.74372e24230903d2d4cacecae74f498d_1" [label="1: Start operator_star_null_deref1\nFormals: \nLocals: t:int* \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
"operator_star_null_deref1#14187169119337849630.74372e24230903d2d4cacecae74f498d_1" -> "operator_star_null_deref1#14187169119337849630.74372e24230903d2d4cacecae74f498d_5" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"operator_star_null_deref1#14187169119337849630.74372e24230903d2d4cacecae74f498d_2" [label="2: Exit operator_star_null_deref1 \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"operator_star_null_deref1#14187169119337849630.74372e24230903d2d4cacecae74f498d_3" [label="3: Return Stmt \n n$1=*&t:int* [line 111, column 10]\n n$2=*n$1:int [line 111, column 10]\n *&return:int=n$2 [line 111, column 3]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
"operator_star_null_deref1#14187169119337849630.74372e24230903d2d4cacecae74f498d_3" -> "operator_star_null_deref1#14187169119337849630.74372e24230903d2d4cacecae74f498d_2" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"operator_star_null_deref1#14187169119337849630.74372e24230903d2d4cacecae74f498d_4" [label="4: Call _fun_TranslateAsPtr<int>::setPtr \n _=*&t:int* [line 110, column 3]\n n$4=_fun_TranslateAsPtr<int>::setPtr(&t:int*&,null:int*) [line 110, column 3]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
"operator_star_null_deref1#14187169119337849630.74372e24230903d2d4cacecae74f498d_4" -> "operator_star_null_deref1#14187169119337849630.74372e24230903d2d4cacecae74f498d_3" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"operator_star_null_deref1#14187169119337849630.74372e24230903d2d4cacecae74f498d_5" [label="5: DeclStmt \n VARIABLE_DECLARED(t:int*); [line 109, column 3]\n n$5=_fun_TranslateAsPtr<int>::TranslateAsPtr(&t:int**,null:int*) [line 109, column 23]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
"operator_star_null_deref1#14187169119337849630.74372e24230903d2d4cacecae74f498d_5" -> "operator_star_null_deref1#14187169119337849630.74372e24230903d2d4cacecae74f498d_4" ;
Get rid of Declare_locals Reviewed By: jeremydubreil Differential Revision: D9234331 fbshipit-source-id: 70443eabe
6 years ago
"operator_star_null_deref2#14189968475942707161.6f6b808f2059b0f1bd8edd63f3e0c27b_1" [label="1: Start operator_star_null_deref2\nFormals: \nLocals: t:int* \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
"operator_star_null_deref2#14189968475942707161.6f6b808f2059b0f1bd8edd63f3e0c27b_1" -> "operator_star_null_deref2#14189968475942707161.6f6b808f2059b0f1bd8edd63f3e0c27b_5" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"operator_star_null_deref2#14189968475942707161.6f6b808f2059b0f1bd8edd63f3e0c27b_2" [label="2: Exit operator_star_null_deref2 \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"operator_star_null_deref2#14189968475942707161.6f6b808f2059b0f1bd8edd63f3e0c27b_3" [label="3: Return Stmt \n _=*&t:int* [line 117, column 10]\n n$1=*&t:int* [line 117, column 10]\n n$2=*n$1:int [line 117, column 10]\n *&return:int=n$2 [line 117, column 3]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
"operator_star_null_deref2#14189968475942707161.6f6b808f2059b0f1bd8edd63f3e0c27b_3" -> "operator_star_null_deref2#14189968475942707161.6f6b808f2059b0f1bd8edd63f3e0c27b_2" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"operator_star_null_deref2#14189968475942707161.6f6b808f2059b0f1bd8edd63f3e0c27b_4" [label="4: Call _fun_TranslateAsPtr<int>::setPtr \n _=*&t:int* [line 116, column 3]\n n$4=_fun_TranslateAsPtr<int>::setPtr(&t:int*&,null:int*) [line 116, column 3]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
"operator_star_null_deref2#14189968475942707161.6f6b808f2059b0f1bd8edd63f3e0c27b_4" -> "operator_star_null_deref2#14189968475942707161.6f6b808f2059b0f1bd8edd63f3e0c27b_3" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"operator_star_null_deref2#14189968475942707161.6f6b808f2059b0f1bd8edd63f3e0c27b_5" [label="5: DeclStmt \n VARIABLE_DECLARED(t:int*); [line 115, column 3]\n n$5=_fun_TranslateAsPtr<int>::TranslateAsPtr(&t:int**,null:int*) [line 115, column 23]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
"operator_star_null_deref2#14189968475942707161.6f6b808f2059b0f1bd8edd63f3e0c27b_5" -> "operator_star_null_deref2#14189968475942707161.6f6b808f2059b0f1bd8edd63f3e0c27b_4" ;
Get rid of Declare_locals Reviewed By: jeremydubreil Differential Revision: D9234331 fbshipit-source-id: 70443eabe
6 years ago
"operator_star_ok_deref#11345277927099423171.138b78e88dab5887cd2f20f2590c779f_1" [label="1: Start operator_star_ok_deref\nFormals: \nLocals: t:int* a:int \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
"operator_star_ok_deref#11345277927099423171.138b78e88dab5887cd2f20f2590c779f_1" -> "operator_star_ok_deref#11345277927099423171.138b78e88dab5887cd2f20f2590c779f_5" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"operator_star_ok_deref#11345277927099423171.138b78e88dab5887cd2f20f2590c779f_2" [label="2: Exit operator_star_ok_deref \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"operator_star_ok_deref#11345277927099423171.138b78e88dab5887cd2f20f2590c779f_3" [label="3: Return Stmt \n _=*&t:int* [line 124, column 10]\n n$1=*&t:int* [line 124, column 10]\n n$2=*n$1:int [line 124, column 10]\n *&return:int=n$2 [line 124, column 3]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
"operator_star_ok_deref#11345277927099423171.138b78e88dab5887cd2f20f2590c779f_3" -> "operator_star_ok_deref#11345277927099423171.138b78e88dab5887cd2f20f2590c779f_2" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"operator_star_ok_deref#11345277927099423171.138b78e88dab5887cd2f20f2590c779f_4" [label="4: Call _fun_TranslateAsPtr<int>::setPtr \n _=*&t:int* [line 123, column 3]\n n$4=_fun_TranslateAsPtr<int>::setPtr(&t:int*&,&a:int*) [line 123, column 3]\n " shape="box"]
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
"operator_star_ok_deref#11345277927099423171.138b78e88dab5887cd2f20f2590c779f_4" -> "operator_star_ok_deref#11345277927099423171.138b78e88dab5887cd2f20f2590c779f_3" ;
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"operator_star_ok_deref#11345277927099423171.138b78e88dab5887cd2f20f2590c779f_5" [label="5: DeclStmt \n VARIABLE_DECLARED(t:int*); [line 122, column 3]\n n$5=_fun_TranslateAsPtr<int>::TranslateAsPtr(&t:int**,null:int*) [line 122, column 23]\n " shape="box"]
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
"operator_star_ok_deref#11345277927099423171.138b78e88dab5887cd2f20f2590c779f_5" -> "operator_star_ok_deref#11345277927099423171.138b78e88dab5887cd2f20f2590c779f_4" ;
RFC: stop using _ to separate ObjC/C++ class name from method in Typ.Procname.to_string Reviewed By: jvillard Differential Revision: D14736442 fbshipit-source-id: 500df354b
6 years ago
"getPtr#TranslateAsPtr<int>#(11398425810130716395).657b6b87ee3e6dc84e17d734bcfc55b1_1" [label="1: Start TranslateAsPtr<int>::getPtr\nFormals: this:int**\nLocals: \n " color=yellow style=filled]
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
"getPtr#TranslateAsPtr<int>#(11398425810130716395).657b6b87ee3e6dc84e17d734bcfc55b1_1" -> "getPtr#TranslateAsPtr<int>#(11398425810130716395).657b6b87ee3e6dc84e17d734bcfc55b1_2" ;
RFC: stop using _ to separate ObjC/C++ class name from method in Typ.Procname.to_string Reviewed By: jvillard Differential Revision: D14736442 fbshipit-source-id: 500df354b
6 years ago
"getPtr#TranslateAsPtr<int>#(11398425810130716395).657b6b87ee3e6dc84e17d734bcfc55b1_2" [label="2: Exit TranslateAsPtr<int>::getPtr \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
RFC: stop using _ to separate ObjC/C++ class name from method in Typ.Procname.to_string Reviewed By: jvillard Differential Revision: D14736442 fbshipit-source-id: 500df354b
6 years ago
"setPtr#TranslateAsPtr<int>#(11427652750021041520).3f4d983a0a5cf5a43b2e4fd66c30c6a9_1" [label="1: Start TranslateAsPtr<int>::setPtr\nFormals: this:int** v:int*\nLocals: \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] Adding parameters as part of the procname for C++/ObjC methods and ObjC blocks Reviewed By: mbouaziz Differential Revision: D8237112 fbshipit-source-id: c0ec4b4
7 years ago
"setPtr#TranslateAsPtr<int>#(11427652750021041520).3f4d983a0a5cf5a43b2e4fd66c30c6a9_1" -> "setPtr#TranslateAsPtr<int>#(11427652750021041520).3f4d983a0a5cf5a43b2e4fd66c30c6a9_3" ;
RFC: stop using _ to separate ObjC/C++ class name from method in Typ.Procname.to_string Reviewed By: jvillard Differential Revision: D14736442 fbshipit-source-id: 500df354b
6 years ago
"setPtr#TranslateAsPtr<int>#(11427652750021041520).3f4d983a0a5cf5a43b2e4fd66c30c6a9_2" [label="2: Exit TranslateAsPtr<int>::setPtr \n " color=yellow style=filled]
[clang] Adding parameters as part of the procname for C++/ObjC methods and ObjC blocks Reviewed By: mbouaziz Differential Revision: D8237112 fbshipit-source-id: c0ec4b4
7 years ago
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"setPtr#TranslateAsPtr<int>#(11427652750021041520).3f4d983a0a5cf5a43b2e4fd66c30c6a9_3" [label="3: BinaryOperatorStmt: Assign \n n$0=*&this:int** [line 86, column 34]\n n$1=*&v:int* [line 86, column 43]\n *n$0:void*=n$1 [line 86, column 23]\n " shape="box"]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] Adding parameters as part of the procname for C++/ObjC methods and ObjC blocks Reviewed By: mbouaziz Differential Revision: D8237112 fbshipit-source-id: c0ec4b4
7 years ago
"setPtr#TranslateAsPtr<int>#(11427652750021041520).3f4d983a0a5cf5a43b2e4fd66c30c6a9_3" -> "setPtr#TranslateAsPtr<int>#(11427652750021041520).3f4d983a0a5cf5a43b2e4fd66c30c6a9_2" ;
RFC: stop using _ to separate ObjC/C++ class name from method in Typ.Procname.to_string Reviewed By: jvillard Differential Revision: D14736442 fbshipit-source-id: 500df354b
6 years ago
"getRef#TranslateAsPtr<int>#(12157952070639259276).02c97edc35db5f793a7a4d1e6c16b00b_1" [label="1: Start TranslateAsPtr<int>::getRef\nFormals: this:int** a:int b:int\nLocals: \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] update clang plugin to hash mangled names Summary: With this change and the previous facebook-clang-plugins change, infer no longer exhausts the biniou buffer when reading the serialized C++ AST. update-submodule: facebook-clang-plugins Reviewed By: mbouaziz Differential Revision: D5891081 fbshipit-source-id: cf48eac
7 years ago
"getRef#TranslateAsPtr<int>#(12157952070639259276).02c97edc35db5f793a7a4d1e6c16b00b_1" -> "getRef#TranslateAsPtr<int>#(12157952070639259276).02c97edc35db5f793a7a4d1e6c16b00b_2" ;
RFC: stop using _ to separate ObjC/C++ class name from method in Typ.Procname.to_string Reviewed By: jvillard Differential Revision: D14736442 fbshipit-source-id: 500df354b
6 years ago
"getRef#TranslateAsPtr<int>#(12157952070639259276).02c97edc35db5f793a7a4d1e6c16b00b_2" [label="2: Exit TranslateAsPtr<int>::getRef \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
RFC: stop using _ to separate ObjC/C++ class name from method in Typ.Procname.to_string Reviewed By: jvillard Differential Revision: D14736442 fbshipit-source-id: 500df354b
6 years ago
"operator*#TranslateAsPtr<int>#(2957914813032465436).d9d28f5b3fa89d06894336545dfa919e_1" [label="1: Start TranslateAsPtr<int>::operator*\nFormals: this:int**\nLocals: \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] Adding parameters as part of the procname for C++/ObjC methods and ObjC blocks Reviewed By: mbouaziz Differential Revision: D8237112 fbshipit-source-id: c0ec4b4
7 years ago
"operator*#TranslateAsPtr<int>#(2957914813032465436).d9d28f5b3fa89d06894336545dfa919e_1" -> "operator*#TranslateAsPtr<int>#(2957914813032465436).d9d28f5b3fa89d06894336545dfa919e_2" ;
RFC: stop using _ to separate ObjC/C++ class name from method in Typ.Procname.to_string Reviewed By: jvillard Differential Revision: D14736442 fbshipit-source-id: 500df354b
6 years ago
"operator*#TranslateAsPtr<int>#(2957914813032465436).d9d28f5b3fa89d06894336545dfa919e_2" [label="2: Exit TranslateAsPtr<int>::operator* \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
RFC: stop using _ to separate ObjC/C++ class name from method in Typ.Procname.to_string Reviewed By: jvillard Differential Revision: D14736442 fbshipit-source-id: 500df354b
6 years ago
"getPtr#TranslateAsPtr<int>#(5108725798531153105).ddae4f977672452bac54a30a4533059d_1" [label="1: Start TranslateAsPtr<int>::getPtr\nFormals: this:int** a:int b:int\nLocals: \n " color=yellow style=filled]
[debug] print procedures in alphabetical order in cfgs Summary: When looking at large CFGs, at least in `xdot`, it's often difficult to find the procedure you're looking for. Sorting the proc names puts them in alphabetical order, which makes searching one procedure easier. Reviewed By: mbouaziz Differential Revision: D7758521 fbshipit-source-id: 8e9997f
7 years ago
[clang] Adding parameters as part of the procname for C++/ObjC methods and ObjC blocks Reviewed By: mbouaziz Differential Revision: D8237112 fbshipit-source-id: c0ec4b4
7 years ago
"getPtr#TranslateAsPtr<int>#(5108725798531153105).ddae4f977672452bac54a30a4533059d_1" -> "getPtr#TranslateAsPtr<int>#(5108725798531153105).ddae4f977672452bac54a30a4533059d_2" ;
RFC: stop using _ to separate ObjC/C++ class name from method in Typ.Procname.to_string Reviewed By: jvillard Differential Revision: D14736442 fbshipit-source-id: 500df354b
6 years ago
"getPtr#TranslateAsPtr<int>#(5108725798531153105).ddae4f977672452bac54a30a4533059d_2" [label="2: Exit TranslateAsPtr<int>::getPtr \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
RFC: stop using _ to separate ObjC/C++ class name from method in Typ.Procname.to_string Reviewed By: jvillard Differential Revision: D14736442 fbshipit-source-id: 500df354b
6 years ago
"getRef#TranslateAsPtr<int>#(8980454460906194048).c1f61acdfdda98d0f31dfdad70fac6a4_1" [label="1: Start TranslateAsPtr<int>::getRef\nFormals: this:int**\nLocals: \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] Adding parameters as part of the procname for C++/ObjC methods and ObjC blocks Reviewed By: mbouaziz Differential Revision: D8237112 fbshipit-source-id: c0ec4b4
7 years ago
"getRef#TranslateAsPtr<int>#(8980454460906194048).c1f61acdfdda98d0f31dfdad70fac6a4_1" -> "getRef#TranslateAsPtr<int>#(8980454460906194048).c1f61acdfdda98d0f31dfdad70fac6a4_2" ;
RFC: stop using _ to separate ObjC/C++ class name from method in Typ.Procname.to_string Reviewed By: jvillard Differential Revision: D14736442 fbshipit-source-id: 500df354b
6 years ago
"getRef#TranslateAsPtr<int>#(8980454460906194048).c1f61acdfdda98d0f31dfdad70fac6a4_2" [label="2: Exit TranslateAsPtr<int>::getRef \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
RFC: stop using _ to separate ObjC/C++ class name from method in Typ.Procname.to_string Reviewed By: jvillard Differential Revision: D14736442 fbshipit-source-id: 500df354b
6 years ago
"TranslateAsPtr#TranslateAsPtr<int>#{16989717360382977660}.33ce04b76efc158540bbe4b4b3c6897f_1" [label="1: Start TranslateAsPtr<int>::TranslateAsPtr\nFormals: this:int** t:int*\nLocals: \n " color=yellow style=filled]
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
[clang] Adding parameters as part of the procname for C++/ObjC methods and ObjC blocks Reviewed By: mbouaziz Differential Revision: D8237112 fbshipit-source-id: c0ec4b4
7 years ago
"TranslateAsPtr#TranslateAsPtr<int>#{16989717360382977660}.33ce04b76efc158540bbe4b4b3c6897f_1" -> "TranslateAsPtr#TranslateAsPtr<int>#{16989717360382977660}.33ce04b76efc158540bbe4b4b3c6897f_3" ;
RFC: stop using _ to separate ObjC/C++ class name from method in Typ.Procname.to_string Reviewed By: jvillard Differential Revision: D14736442 fbshipit-source-id: 500df354b
6 years ago
"TranslateAsPtr#TranslateAsPtr<int>#{16989717360382977660}.33ce04b76efc158540bbe4b4b3c6897f_2" [label="2: Exit TranslateAsPtr<int>::TranslateAsPtr \n " color=yellow style=filled]
[clang] Adding parameters as part of the procname for C++/ObjC methods and ObjC blocks Reviewed By: mbouaziz Differential Revision: D8237112 fbshipit-source-id: c0ec4b4
7 years ago
[cfg] run pre-analysis lazily in ondemand Summary: This also prints the CFGs *after* pre-analysis for individual procedures in infer-out/captured/<filename>/<proc>.dot. One can also look up the CFGs before pre-analysis in infer-out/captured/proc_cfgs_frontend.dot. Context: I want to add a pre-analysis that needs to look at proc attributes inter-procedurally. For this to make sense it has to happen *after* all of capture, and before analysis. Thus, this diff brings back the lazy running of the pre-analysis like in D15803492, except that we still make sure to run the pre-analyses systematically regardless of the checkers being run by running the pre-analysis from ondemand.ml. Also we don't need to re-introduce the "did_preanalysis" proc attribute for the same reason that the pre-analysis is now run once and for all by ondemand.ml (instead of each individual checker back in the days). This has the benefit of running the pre-analysis only when needed, and the drawback that several concurrent processes analysing the same proc descs will duplicate work. Since pre-analyses are supposed to be very fast I assume that neither is a big deal. If they become more expensive then the benefit gets bigger and the drawback is just the same as with regular analyses. Reviewed By: skcho Differential Revision: D18573920 fbshipit-source-id: de350eaef
5 years ago
"TranslateAsPtr#TranslateAsPtr<int>#{16989717360382977660}.33ce04b76efc158540bbe4b4b3c6897f_3" [label="3: Call _fun_TranslateAsPtr<int>::setPtr \n n$0=*&this:int** [line 76, column 36]\n _=*n$0:int* [line 76, column 36]\n n$2=*&t:int* [line 76, column 43]\n n$3=_fun_TranslateAsPtr<int>::setPtr(n$0:int**,n$2:int*) [line 76, column 36]\n " shape="box"]
[clang] Adding parameters as part of the procname for C++/ObjC methods and ObjC blocks Reviewed By: mbouaziz Differential Revision: D8237112 fbshipit-source-id: c0ec4b4
7 years ago
"TranslateAsPtr#TranslateAsPtr<int>#{16989717360382977660}.33ce04b76efc158540bbe4b4b3c6897f_3" -> "TranslateAsPtr#TranslateAsPtr<int>#{16989717360382977660}.33ce04b76efc158540bbe4b4b3c6897f_2" ;
Skip Sil.Call instruction for some shared_ptr method calls Summary: 1. Add capability to clang frontend to replace some function calls with another SIL code based on `__deprecated__` attribute. 2. Given this capability, use those attributes for shared_ptr getters to generate `Sil.Load` instruction instead of method call 3. Add test that mimics shared_ptr model, but it doesn't have that much scary C++ templated code Reviewed By: jvillard, jberdine Differential Revision: D3729176 fbshipit-source-id: 2a330d5
8 years ago
}