Re: [cfe-commits] [PATCH] WIP: catch CFMakeCollectable(NULL) with static analyzer
Committed in r167537. Thanks, Sean! Jordan On Nov 5, 2012, at 18:32 , Sean McBride s...@rogue-research.com wrote: Jordan, Updated patch attached. Thanks, Sean On Fri, 2 Nov 2012 09:49:37 -0700, Jordan Rose said: Hi, Sean. Thanks for the patch! Most of the tests in retain-release.m are just top-level functions, numbered numerically from lack of imagination. The arguments are fairly arbitrary—'p' is just 'pointer' (I think), and 'x' is being used as a sort of coin toss to arbitrarily create conditions where the analyzer considers both branches equally likely. The actual patch itself looks good, except I'd stick an 'llvm_unreachable' in a last 'else' case for that chained if-statement. Your test output is fine; if something broke it would say Unexpected Failures and probably print output from that test as well. Once you've updated the patch, please send it back to the list! Jordan CFMakeCollectable-with-null2.patch ___ cfe-commits mailing list cfe-commits@cs.uiuc.edu http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
Re: [cfe-commits] [PATCH] WIP: catch CFMakeCollectable(NULL) with static analyzer
Jordan, Updated patch attached. Thanks, Sean On Fri, 2 Nov 2012 09:49:37 -0700, Jordan Rose said: Hi, Sean. Thanks for the patch! Most of the tests in retain-release.m are just top-level functions, numbered numerically from lack of imagination. The arguments are fairly arbitrary—'p' is just 'pointer' (I think), and 'x' is being used as a sort of coin toss to arbitrarily create conditions where the analyzer considers both branches equally likely. The actual patch itself looks good, except I'd stick an 'llvm_unreachable' in a last 'else' case for that chained if-statement. Your test output is fine; if something broke it would say Unexpected Failures and probably print output from that test as well. Once you've updated the patch, please send it back to the list! Jordan CFMakeCollectable-with-null2.patch Description: Binary data ___ cfe-commits mailing list cfe-commits@cs.uiuc.edu http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
Re: [cfe-commits] [PATCH] WIP: catch CFMakeCollectable(NULL) with static analyzer
Hi, Sean. Thanks for the patch! Most of the tests in retain-release.m are just top-level functions, numbered numerically from lack of imagination. The arguments are fairly arbitrary—'p' is just 'pointer' (I think), and 'x' is being used as a sort of coin toss to arbitrarily create conditions where the analyzer considers both branches equally likely. The actual patch itself looks good, except I'd stick an 'llvm_unreachable' in a last 'else' case for that chained if-statement. Your test output is fine; if something broke it would say Unexpected Failures and probably print output from that test as well. Once you've updated the patch, please send it back to the list! Jordan On Nov 1, 2012, at 12:17 , Sean McBride s...@rogue-research.com wrote: Hi all, Attached is a patch that expands the analyzer's check for CFRelease(NULL) to also catch CFMakeCollectable(NULL). It builds, and the resulting clang does now warn on CFMakeCollectable(NULL), but I have not added a test. I'm pretty sure the test belongs in retain-release.m in f16() and f17(), but don't see who calls them, what their parameters mean, or what their names mean. I ran the test suite: $ make check-clang [100%] Running the Clang regression tests lit.py: lit.cfg:171: note: using clang: '/Users/smcbride/Documents/Development/llvm-bin/bin/./clang' Testing Time: 169.28s Expected Passes: 5511 Expected Failures : 24 Unsupported Tests : 5 [100%] Built target check-clang Not sure what output I should see... I guess this indicates nothing got broken? Thanks, -- Sean McBride, B. Eng s...@rogue-research.com Rogue Researchwww.rogue-research.com Mac Software Developer Montréal, Québec, Canada CFMakeCollectable-with-null.patch___ cfe-commits mailing list cfe-commits@cs.uiuc.edu http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits ___ cfe-commits mailing list cfe-commits@cs.uiuc.edu http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
[cfe-commits] [PATCH] WIP: catch CFMakeCollectable(NULL) with static analyzer
Hi all, Attached is a patch that expands the analyzer's check for CFRelease(NULL) to also catch CFMakeCollectable(NULL). It builds, and the resulting clang does now warn on CFMakeCollectable(NULL), but I have not added a test. I'm pretty sure the test belongs in retain-release.m in f16() and f17(), but don't see who calls them, what their parameters mean, or what their names mean. I ran the test suite: $ make check-clang [100%] Running the Clang regression tests lit.py: lit.cfg:171: note: using clang: '/Users/smcbride/Documents/Development/llvm-bin/bin/./clang' Testing Time: 169.28s Expected Passes: 5511 Expected Failures : 24 Unsupported Tests : 5 [100%] Built target check-clang Not sure what output I should see... I guess this indicates nothing got broken? Thanks, -- Sean McBride, B. Eng s...@rogue-research.com Rogue Researchwww.rogue-research.com Mac Software Developer Montréal, Québec, Canada CFMakeCollectable-with-null.patch Description: Binary data ___ cfe-commits mailing list cfe-commits@cs.uiuc.edu http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
