New issue
Advanced search Search tips

Issue 693471 link

Starred by 3 users

Issue metadata

Status: Verified
Owner:
Closed: Mar 2017
Components:
EstimatedDays: ----
NextAction: ----
OS: Chrome
Pri: 3
Type: Bug



Sign in to add a comment

authpolicy: Follow up when authpolicy is disabled on x86

Project Member Reported by ljusten@chromium.org, Feb 17 2017

Issue description

When chromeos-chrome-9999.ebuild is marked stable, move disabling-on-x86 logic to package.use.mask and fix TODOs in authpolicy (search for this bug id).
 
Status: Started (was: Assigned)
Project Member

Comment 2 by bugdroid1@chromium.org, Mar 2 2017

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/platform2/+/aa475291bb5c5adc1c0f53586fdc468681a188f2

commit aa475291bb5c5adc1c0f53586fdc468681a188f2
Author: Lutz Justen <ljusten@chromium.org>
Date: Thu Mar 02 02:42:23 2017

authpolicy: Remove workaround code for x86

Removes x86 workaround code in unit tests now that authpolicy
is disabled on x86.

BUG= chromium:693471 
TEST=Compiled, ran tests.

Change-Id: I6c517aa69a5903db1ab329ec50e4892a59cbc8b5
Reviewed-on: https://chromium-review.googlesource.com/446109
Commit-Ready: Lutz Justen <ljusten@chromium.org>
Tested-by: Lutz Justen <ljusten@chromium.org>
Reviewed-by: Roman Sorokin <rsorokin@chromium.org>

[modify] https://crrev.com/aa475291bb5c5adc1c0f53586fdc468681a188f2/authpolicy/authpolicy_unittest.cc
[modify] https://crrev.com/aa475291bb5c5adc1c0f53586fdc468681a188f2/authpolicy/stub_net_main.cc

Project Member

Comment 3 by bugdroid1@chromium.org, Mar 2 2017

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/overlays/chromiumos-overlay/+/d17ed5ff3a168d0d2791e383453ed92b3fe2308a

commit d17ed5ff3a168d0d2791e383453ed92b3fe2308a
Author: Lutz Justen <ljusten@chromium.org>
Date: Thu Mar 02 15:51:53 2017

Disable authpolicy for x86 in package.use.mask

Disables authpolicy for x86 via package.use.mask (aka 'the right
way') and removes similar code from chromeos-chrome-9999.ebuild
(aka 'the wrong way'). package.use.mask references a future version
of Chrome to work around an issue in the build system, see failed
PCQ build in CL:443589.

BUG= chromium:693471 
TEST=build_packages for x86

Change-Id: Ib7550b50f5424668c4d938a96948eb236d0005d8
Reviewed-on: https://chromium-review.googlesource.com/445905
Commit-Ready: Lutz Justen <ljusten@chromium.org>
Tested-by: Lutz Justen <ljusten@chromium.org>
Reviewed-by: Roman Sorokin <rsorokin@chromium.org>

[modify] https://crrev.com/d17ed5ff3a168d0d2791e383453ed92b3fe2308a/profiles/arch/x86/package.use.mask
[modify] https://crrev.com/d17ed5ff3a168d0d2791e383453ed92b3fe2308a/chromeos-base/chromeos-chrome/chromeos-chrome-9999.ebuild

Status: Fixed (was: Started)
there is failure in the waterfall that has been happening for several days:
https://uberchromegw.corp.google.com/i/chromeos/builders/mario-incremental/builds/38861/steps/UnitTest/logs/stdio

authpolicy-0.0.1-r223: [       OK ] AuthPolicyTest.AuthFailsNotJoined (5 ms)
authpolicy-0.0.1-r223: [ RUN      ] AuthPolicyTest.JoinSucceeds
authpolicy-0.0.1-r223: [0305/162103:INFO:authpolicy.cc(121)] Received 'JoinADDomain' request
authpolicy-0.0.1-r223: Error: /var/cache/portage/chromeos-base/authpolicy/out/Default/authpolicy_test: failed with signal SIGIOT|SIGABRT(6)

is this related to your changes? is it expected?

Comment 6 by ljusten@google.com, Mar 5 2017

Yes, this is related, but I didn't expect it. Authpolicy is supposed to be disabled on x86. There was a related change that landed on Mar 02 that probably exposed the error:
https://chromium-review.googlesource.com/c/445905/
The CL changes the way authpolicy is turned off. Modifying Chrome USE flags is a bit tricky because of prebuilts (see build errors in that CL), so I disabled the USE flag for a (at the time) future version of Chrome (58.0.3028.0_rc-r1).
Do you know whether the unit tests use a prebuilt version of Chrome or compile it using the 9999 ebuild? My only explanation is that it's using the 9999 ebuild. In any case, the error should vanish once Chrome gets uprev'ed to 58.0.3028.0_rc-r1. Is this urgent or is it fine if it sits a couple more days?
Tracked in  crbug.com/699663 
Status: Verified (was: Fixed)
bulk Verify of older or not-user-facing Chromad bugs

Sign in to add a comment