The test is now named 'bitrate_controller_unittests'.
This CL also enables it on the bots. The test is excluded on ASAN since
it fails when compiled with projects generated with GYP_DEFINES='asan=1' (see issue 555).
BUG=None
TEST=bitrate_controller_unittests was tested in Debug+Release on Linux, Mac and Windows + TSAN and memcheck.
Review URL: https://webrtc-codereview.appspot.com/612004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@2333 4adac7df-926f-26a2-2b94-8c16560cd09d
Chrome bots are now using the correct scheduler.
Created separate watchlist e-mail for libvpx builds to avoid spamming WebRTC when these builds fail.
BUG=None
TEST=Tested on local master and slaves.
Review URL: https://webrtc-codereview.appspot.com/620004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@2331 4adac7df-926f-26a2-2b94-8c16560cd09d
I think I found a decent way to avoid duplicating the whole configuration for the libvpx waterfall. It's not perfect but it works. I still haven't figured out what the best way to manage the slaves would be however, since they will need to know which master to connect to, or they'll pick the default they're configured for in slaves.cfg, which is the WebRTC waterfall (can be overridden with the TESTING_MASTER property, but that's only to be used for development and testing.
BUG=None
TEST=Tested on local master and slaves.
Review URL: https://webrtc-codereview.appspot.com/595005
git-svn-id: http://webrtc.googlecode.com/svn/trunk@2310 4adac7df-926f-26a2-2b94-8c16560cd09d
Skeleton suppression files for future WebRTC suppressions are added and are included in addition to the ones Chrome are using and maintaining when our wrapper script executes.
Also added tweaked PRESUBMIT checks based on the Chrome code, that verifies
that suppressions are added correctly. I tested that they work by adding an invalid
suppression.
BUG=544
TEST=Tested running tools/valgrind-webrtc/webrtc_tests.sh --tool=tsan -t out/Debug/system_wrappers_unittests and it reports far less errors. Tested adding bad suppression and it was caught by the PRESUBMIT check.
Review URL: https://webrtc-codereview.appspot.com/601004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@2304 4adac7df-926f-26a2-2b94-8c16560cd09d
The old scheduler was missing the branch configuration, that was started being used as a consequence of adapting the Chromium Buildbot scripts.
TBR=phoglund
BUG=None
TEST=Tested on live master+slaves with manual edits.
Review URL: https://webrtc-codereview.appspot.com/576012
git-svn-id: http://webrtc.googlecode.com/svn/trunk@2250 4adac7df-926f-26a2-2b94-8c16560cd09d
Also removed the old Linux Chrome and Chrome Bloat bots from utils.py.
Some of these changes may seem a bit confusing and messy but they must be considered a step on the way to moving away from our inheritance based Factories into the style Chrome users, where most configuration is in the .cfg files, factory setup is in the factories and step details are put into the commands-files.
BUG=None
TEST=Tested with local master and production slaves.
Review URL: https://webrtc-codereview.appspot.com/581006
git-svn-id: http://webrtc.googlecode.com/svn/trunk@2238 4adac7df-926f-26a2-2b94-8c16560cd09d
The change in r2143 changed so we're reusing Chromium scripts for setting up the web status. However Chrome doesn't use the tgrid page it seems (they put the Console page on that URL instead) but we need it for our LKGR parsing.
BUG=None
TEST=Tested on local master.
Review URL: https://webrtc-codereview.appspot.com/541004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@2146 4adac7df-926f-26a2-2b94-8c16560cd09d
The fastbuild flag means debug symbols are not generated. This has no
impact on our tests as long as we don't need to debug with the built
binaries, i.e. is perfect for the buildbots. Compilation is 30% faster
or more.
BUG=None
TEST=Tested on local master and linux, mac and win slaves.
Review URL: https://webrtc-codereview.appspot.com/499001
git-svn-id: http://webrtc.googlecode.com/svn/trunk@2062 4adac7df-926f-26a2-2b94-8c16560cd09d
Added back the gviz_api to the tools deps - will remove it from the main project DEPS in a different patch. Rewired those symlinks too.
Made the build status loader algorithm more scalable. It read in the whole database on each load which is probably unsustainable in the long run. Also, it will now forget bots that have offline for more than ~5 revisions.
Fixed a bug in the coverage tracker.
BUG=
TEST=
Review URL: https://webrtc-codereview.appspot.com/486002
git-svn-id: http://webrtc.googlecode.com/svn/trunk@2039 4adac7df-926f-26a2-2b94-8c16560cd09d
Hijacking the gyp inside_chromium_build finally came back to bite us.
Chrome started using it to control the path to the gold linker, which
broke our build. This change removes use of inside_chromium_build from
WebRTC.
If peerconnection rolls past this point, libjingle.gyp will need to be updated.
BUG=
TEST=build on Linux, Mac, Win
Review URL: https://webrtc-codereview.appspot.com/493007
git-svn-id: http://webrtc.googlecode.com/svn/trunk@2038 4adac7df-926f-26a2-2b94-8c16560cd09d
