diff options
| author | viettrungluu@chromium.org <viettrungluu@chromium.org@0039d316-1c4b-4281-b951-d872f2087c98> | 2014-05-28 21:55:45 +0000 |
|---|---|---|
| committer | viettrungluu@chromium.org <viettrungluu@chromium.org@0039d316-1c4b-4281-b951-d872f2087c98> | 2014-05-28 21:55:45 +0000 |
| commit | 9646057ea365fd0d0ccdde7d4989517d1e56bdd3 (patch) | |
| tree | 650c949fbfbf6f536b46e0f9436343e0ae4e9465 /content/browser/download/save_package.h | |
| parent | 45e8f89a95998a616d73121b15da567e06472176 (diff) | |
| download | chromium_src-9646057ea365fd0d0ccdde7d4989517d1e56bdd3.zip chromium_src-9646057ea365fd0d0ccdde7d4989517d1e56bdd3.tar.gz chromium_src-9646057ea365fd0d0ccdde7d4989517d1e56bdd3.tar.bz2 | |
Mojo: Refactor the sending of FDs (a.k.a. PlatformHandles) on POSIX.
This should (hopefully) solve some correctness bugs. (The semantics of
sending control messages with FDs on a SOCK_STREAM Unix domain socket
are somewhat tricky. In particular, if you do a sendmsg() with some FDs
and some data, followed by sending more data (using sendmsg() again,
say), then a recvmsg() on the other side may receive the FDs together
with data from multiple sendmsg()s, or possibly also incomplete data
from the first sendmsg().)
Separately: I'll add more careful tests , and also re-enable
MultiprocessMessagePipeTest.PlatformHandlePassing on Mac -- which caught
the bug.
R=yzshen@chromium.org
Review URL: https://codereview.chromium.org/294393014
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273367 0039d316-1c4b-4281-b951-d872f2087c98
Diffstat (limited to 'content/browser/download/save_package.h')
0 files changed, 0 insertions, 0 deletions