m59 pfq failing MasterSlaveSync |
|||||||||||||
Issue descriptionSeems to be a problem with the builder: https://uberchromegw.corp.google.com/i/chromeos_release/builders/samus-pre-flight-branch%20release-R59-9460.B/builds/247
,
Jun 8 2017
,
Jun 8 2017
,
Jun 8 2017
looks like a failure in cros_chromite_internal.py?
,
Jun 8 2017
,
Jun 8 2017
,
Jun 8 2017
I'm trying syncing to Chrome 59.0.3071.94, and seeing if I get the same errors with a "gen gen ..." (with the args listed in the build step that failed). Hopefully it does fail, and I can quickly bisect what changed between that and .93 I'll also try looking through the patches added since .93, but the bisect might be faster for me. I didn't find any mention of a dependency bug elsewhere, so I expect this a ChromeOS specific breakage caused by some cherrypick.
,
Jun 8 2017
I did reproduce the build error. The list of changes to base Chrome is actually smaller than I expected, and nothing jumped out of me: https://chromium.googlesource.com/chromium/src/+log/59.0.3071.93..59.0.3071.94?pretty=fuller&n=10000 The error message points at the webrtc build files needing something not defined in the base files. The above CL's included a DEPS roll for a large range of webrtc patches. But looking back at the DEPS roll for .93 ... .92, showed the same starting point but a different (earlier) endpoint. The list of patches different for webrtc between .92 and .93 has length one, and does seem to be the culprit: (diff list) https://chromium.googlesource.com/external/webrtc/trunk/webrtc/+log/d45f54b8540f1953793026a20a828cf50d50120b..352fabe2d063ebf3d2302ac726911b3403704e8f (CL review) https://codereview.webrtc.org/2897523002
,
Jun 8 2017
Reverting back https://codereview.webrtc.org/2897523002 locally makes the "gn gen ..." succeed. I don't know the details of how webrtc changes end up in the release branch, but it looks like it depends on some other change that did not get included in base Chrome. Sending to thomasanderson@ as he might be able to suggest a way to fix it.
,
Jun 8 2017
Ah, a bit more searching shows there was already a revert that landed yesterday https://codereview.webrtc.org/2927763003/ associated with crbug.com/730934 We may need to do nothing other than wait for 59.0.3071.95
,
Jun 8 2017
,
Jun 8 2017
Does that seem reasonable to you Grace?
,
Jun 8 2017
,
Jun 9 2017
I think I'm off the hook here. The GN breaking change was quickly reverted as you already discovered.
,
Jul 13 2017
,
May 14 2018
|
|||||||||||||
►
Sign in to add a comment |
|||||||||||||
Comment 1 by dgarr...@chromium.org
, Jun 8 2017