diff options
author | newt@chromium.org <newt@chromium.org@0039d316-1c4b-4281-b951-d872f2087c98> | 2013-12-06 23:24:37 +0000 |
---|---|---|
committer | newt@chromium.org <newt@chromium.org@0039d316-1c4b-4281-b951-d872f2087c98> | 2013-12-06 23:24:37 +0000 |
commit | ad0dfb76499ae73a0f52a06e905680d5cd72bf0b (patch) | |
tree | 45b6cac1913feebdcd0746888f7287feeca8c82d /chrome/browser/defaults.h | |
parent | 7f693e5cfc0018313a78d69e72f2f915ea262361 (diff) | |
download | chromium_src-ad0dfb76499ae73a0f52a06e905680d5cd72bf0b.zip chromium_src-ad0dfb76499ae73a0f52a06e905680d5cd72bf0b.tar.gz chromium_src-ad0dfb76499ae73a0f52a06e905680d5cd72bf0b.tar.bz2 |
Don't use sys.exit() in build_utils.CheckCallDie().
Calling sys.exit() in a thread spawned by
multiprocessing.pool.ThreadPool causes an infinite hang that even SIGINT
can't fix. This is bad. Instead of calling sys.exit(), raise an
exception, crafted so that the important information (the failed
command's stdout and stderr) comes after the stacktrace and the copyable
command.
This also renames CheckCallDie() to CheckOutput() and changes it to
print stderr but not stdout by default. Suppressing stderr has hid bugs
in the past (e.g. aapt crunch failures), and hiding stdout is what we
usually want anyway.
R=cjhopman@chromium.org
Review URL: https://codereview.chromium.org/106923002
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@239276 0039d316-1c4b-4281-b951-d872f2087c98
Diffstat (limited to 'chrome/browser/defaults.h')
0 files changed, 0 insertions, 0 deletions