summaryrefslogtreecommitdiffstats
path: root/net/http/http_security_headers_unittest.cc
diff options
context:
space:
mode:
authormpearson@chromium.org <mpearson@chromium.org@0039d316-1c4b-4281-b951-d872f2087c98>2013-02-26 18:39:24 +0000
committermpearson@chromium.org <mpearson@chromium.org@0039d316-1c4b-4281-b951-d872f2087c98>2013-02-26 18:39:24 +0000
commitd22837bfda1972f69bca9a29784ae6f69f5d9e3f (patch)
tree0f5e17116a9efe2ef155a4739cac1c5508044b87 /net/http/http_security_headers_unittest.cc
parent28865f1149f18ad7089572bd8ec86257268d222f (diff)
downloadchromium_src-d22837bfda1972f69bca9a29784ae6f69f5d9e3f.zip
chromium_src-d22837bfda1972f69bca9a29784ae6f69f5d9e3f.tar.gz
chromium_src-d22837bfda1972f69bca9a29784ae6f69f5d9e3f.tar.bz2
Experimental Revert of 184577
ASAN bots started timing out consistently last night at the time this was submitted. It's a possible culprit. It doesn't involve extensions, but it does things with streams and I assume those are connected to the feeds in this test somehow. browser_tests browser_tests failed 4 ( 12 mins, 52 secs ) stdio ParseFeedInvalidFeed1 ParseFeedInvalidFeed2 ParseFeedValidFeed3 ParseFeedValidFeed4 http://build.chromium.org/p/chromium.memory/builders/Linux%20Chromium%20OS%20ASAN%20Tests%20%281%29/builds/3019 http://build.chromium.org/p/chromium.memory/builders/Linux%20Chromium%20OS%20ASAN%20Tests%20%282%29/builds/3186 http://build.chromium.org/p/chromium.memory/builders/Linux%20Chromium%20OS%20ASAN%20Tests%20%283%29/builds/2716 > Fix net::FileStream to handle all errors correctly. > > This CL adds FileStream::Context::IOResult struct that's used to pass > results of IO operations inside FileStream::Context. Following bugs are > fixes: > > 1. On POSIX net::FileStream::Read() and net::FileStream::Write() would > return a positive result in case of an error. This happens because > POSIX errors are positive integers and FileStream was written with > assumption that they are negative. > 2. On Windows Seek() and Flush() errors were not handled correctly. > This is because CheckForIOError() was assuming that all error codes > are negative, but RecordAndMapError() was mapping positive errors. > 3. On Windows results of asynchronous ReadFile() and WriteFile() were > not handled correctly - ERR_IO_PENDING would be returned even when > operation completes synchronously. > 4. On Windows OVERLAPPED struct wasn't set to zeros as necessary. > > Also added unittests to check that error codes are handled correctly > now. > > TBR=mmenke@chromium.org > > Review URL: https://codereview.chromium.org/12326107 TBR=sergeyu@chromium.org Review URL: https://codereview.chromium.org/12330144 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@184672 0039d316-1c4b-4281-b951-d872f2087c98
Diffstat (limited to 'net/http/http_security_headers_unittest.cc')
0 files changed, 0 insertions, 0 deletions