This is a second try on http://reviews.llvm.org/D5724 which was approved and commited but later reverted after causing bugs in upstream. This patch is mostly identical (except for fixing the issue for OSX and rebasing). It relies on BlockinMutex patch in http://reviews.llvm.org/D7171
Details
Details
Diff Detail
Diff Detail
- Repository
- rL LLVM
Event Timeline
lib/asan/asan_report.cc | ||
---|---|---|
56 ↗ | (On Diff #18744) | Feel free to commit this part separately in a cleanup CL. |
lib/sanitizer_common/sanitizer_common.cc | ||
224 ↗ | (On Diff #18744) | I forgot, why do you need this? |
lib/sanitizer_common/sanitizer_linux.cc | ||
754 ↗ | (On Diff #18744) | Use default_module_name here. |
lib/sanitizer_common/sanitizer_printf.cc | ||
258 ↗ | (On Diff #18744) | Hm, do we really want this? I believe some users may scrape logs by searching for ==%d== prefix. |
test/asan/TestCases/verbose-log-path_test.cc | ||
1 ↗ | (On Diff #18744) | Why do you need this? |
12 ↗ | (On Diff #18744) | This test will not be run on Windows anyway - verbose-log-path is not set there. |
lib/sanitizer_common/sanitizer_common.cc | ||
---|---|---|
224 ↗ | (On Diff #18744) | It's for Windows. |
lib/sanitizer_common/sanitizer_printf.cc | ||
258 ↗ | (On Diff #18744) | Knowing pname may be useful when output from multiple processes gets printed into stderr. Perhaps do ==%d==%s==? |
test/asan/TestCases/verbose-log-path_test.cc | ||
1 ↗ | (On Diff #18744) | I simply copy-pasted this from log-path_test.cc and thought that it's better to avoid changes. I could remove useless stuff if necessary. |
12 ↗ | (On Diff #18744) | Ditto. |
Comment Actions
Yikes, looks like I accidentally closed this by a partial cleanup commit. Should I create a new review?
lib/asan/asan_report.cc | ||
---|---|---|
56 ↗ | (On Diff #18744) | Landed in 227559. |