Re: Issue 520586 in chromium: CQ published uprev despite build failure; bluez-5.33-r31 failing in BuildPackages

[chro...@googlecode.com]

Updates:
Cc: i...@chromium.org mar...@chromium.org

Comment #4 on issue 520586 by hs...@chromium.org: CQ published uprev despite
build failure; bluez-5.33-r31 failing in BuildPackages
https://code.google.com/p/chromium/issues/detail?id=520586

ihf@ was looking at the bluez failure as late as 11:30pm last night
he has a CL up https://chromium-review.googlesource.com/#/c/293400/

--
You received this message because this project is configured to send all
issue notifications to this address.
You may adjust your notification preferences at:
https://code.google.com/hosting/settings

[chro...@googlecode.com]

Updates:
Status: Available
Labels: -Pri-1 Pri-2
Mergedinto:

Comment #6 on issue 520586 by akes...@chromium.org: CQ published uprev

despite build failure; bluez-5.33-r31 failing in BuildPackages
https://code.google.com/p/chromium/issues/detail?id=520586

un-duping. Looks like the bluez breakage is fixed, but I still want to
understand how this uprev and breakage made it through the cq.

[chro...@googlecode.com]

Comment #9 on issue 520586 by bugd...@chromium.org: CQ published uprev

despite build failure; bluez-5.33-r31 failing in BuildPackages

https://code.google.com/p/chromium/issues/detail?id=520586#c9

The following revision refers to this bug:

https://chromium.googlesource.com/chromiumos/overlays/chromiumos-overlay/+/20ed6b2fc53d42052deb4847cd608f4d0eec62c1

commit 20ed6b2fc53d42052deb4847cd608f4d0eec62c1
Author: Aviv Keshet <ake...@chromium.org>
Date: Thu Aug 13 21:55:26 2015

bluez: blacklist bluez from automatic uprevving

BUG=chromium:520586
TEST=None

Change-Id: I7ecc6fe6a9c0fa42af68297cf20023194bb2057e
Reviewed-on: https://chromium-review.googlesource.com/293503
Reviewed-by: Aviv Keshet <ake...@chromium.org>
Tested-by: Aviv Keshet <ake...@chromium.org>
Commit-Queue: Aviv Keshet <ake...@chromium.org>

[modify]
http://crrev.com/20ed6b2fc53d42052deb4847cd608f4d0eec62c1/net-wireless/bluez/bluez-9999.ebuild
[rename]
http://crrev.com/20ed6b2fc53d42052deb4847cd608f4d0eec62c1/net-wireless/bluez/bluez-5.33-r34.ebuild

[chro...@googlecode.com]

Updates:
Status: Assigned
Owner: so...@chromium.org
Cc: jpawlow...@chromium.org armans...@chromium.org
Labels: -Pri-2 -Hotlist-Recharge Pri-1 Cr-OS-Systems-Bluetooth
ReleaseBlock-Stable

Comment #11 on issue 520586 by msa...@chromium.org: CQ published uprev

despite build failure; bluez-5.33-r31 failing in BuildPackages
https://code.google.com/p/chromium/issues/detail?id=520586

We have been working on Smart Lock over Bluetooth Low Energy on Chrome OS
that was supposed to be in fishfood in Chrome M47. This had a large
dependency on the bluez. Over the past 2 months Jakub put a lot of effort
into fixing a lot of bugs in bluez and we were expecting to have them all
in Chrome OS M47.

Due to this bug, all these fixes missed the branch and this is really
hurting our product. I think this bug needs to be addressed to avoid such
product issues in the future.

I am raising its priority and marking it as release block beta as I think
we should not branch M48 without this bug being fixes.

[chro...@googlecode.com]

Updates:
Labels: M-48

Comment #12 on issue 520586 by msa...@chromium.org: CQ published uprev

despite build failure; bluez-5.33-r31 failing in BuildPackages
https://code.google.com/p/chromium/issues/detail?id=520586

(No comment was entered for this change.)

[chro...@googlecode.com]

Updates:
Status: Fixed

Comment #14 on issue 520586 by david...@chromium.org: CQ published uprev

despite build failure; bluez-5.33-r31 failing in BuildPackages
https://code.google.com/p/chromium/issues/detail?id=520586

This was fixed back in Aug 17. msarda@ did you have a different fix in mind?

[chro...@googlecode.com]

Updates:
Status: Assigned

Comment #16 on issue 520586 by msa...@chromium.org: CQ published uprev

despite build failure; bluez-5.33-r31 failing in BuildPackages
https://code.google.com/p/chromium/issues/detail?id=520586

davidjames@:

I'm reopening the bug as I do not understand what the policy regarding
repositories is. Let me explain.

It seems crazy to me that we consider blacklisting the bluez repository as
the final fix for this bug. In this particular case, there were a lot of
fixes on the upstream bluez repository that were never mirrored in Chrome
OS due to this blacklisting. It seems that the bluez owners in Chrome OS
were not aware of this (armansito@ in CC), as Chrome OS M47 will ship with
bluez from early August.

How will this work from now on? Do you expect the bluez repo to be manually
rolled? If so, was the Chrome OS Bluetooth owners informed of this decision?

[chro...@googlecode.com]

Comment #21 on issue 520586 by i...@chromium.org: CQ published uprev

despite build failure; bluez-5.33-r31 failing in BuildPackages
https://code.google.com/p/chromium/issues/detail?id=520586

As an aside, #3 sounds exciting (like chromeos-chrome I guess) but it would
require automatic notifications and gardening of failures to prevent the
same situation as here. Is there a doc/bug for this to subscribe to?

[chro...@googlecode.com]

Comment #22 on issue 520586 by david...@chromium.org: CQ published uprev

despite build failure; bluez-5.33-r31 failing in BuildPackages
https://code.google.com/p/chromium/issues/detail?id=520586

For #3, Don hasn't designed the feature yet, but it will be sent to
chromeos-infra-discuss and I can ask him to CC you.

[chro...@googlecode.com]

Comment #23 on issue 520586 by bugd...@chromium.org: CQ published uprev

despite build failure; bluez-5.33-r31 failing in BuildPackages

https://code.google.com/p/chromium/issues/detail?id=520586#c23

The following revision refers to this bug:

https://chromium.googlesource.com/chromiumos/overlays/chromiumos-overlay/+/c006bfd463b8f4f2a6acd8278ffa5783ef504ca1

commit c006bfd463b8f4f2a6acd8278ffa5783ef504ca1
Author: David James <david...@google.com>
Date: Tue Oct 20 19:29:25 2015

Unblacklist bluez ebuild.

The bluetooth team is now using the commit queue so there's no need
to blacklist.

BUG=chromium:520586
TEST=none

Change-Id: I116570aa675df4e246f9cb03ca6bb5618e13d9f4
Reviewed-on: https://chromium-review.googlesource.com/307458
Commit-Ready: David James <david...@chromium.org>
Tested-by: David James <david...@chromium.org>
Reviewed-by: Arman Uguray <arma...@chromium.org>

[modify]
http://crrev.com/c006bfd463b8f4f2a6acd8278ffa5783ef504ca1/net-wireless/bluez/bluez-9999.ebuild
[delete]
http://crrev.com/ad1a6e5f1a053fcee88a7bc18e280cce11611cae/net-wireless/bluez/bluez-5.28.ebuild
[modify]
http://crrev.com/c006bfd463b8f4f2a6acd8278ffa5783ef504ca1/net-wireless/bluez/bluez-5.35-r37.ebuild

[chro...@googlecode.com]

Comment #24 on issue 520586 by fbeauf...@chromium.org: CQ published uprev

despite build failure; bluez-5.33-r31 failing in BuildPackages
https://code.google.com/p/chromium/issues/detail?id=520586

It looks like we're on the right path from now on. Thank you!

Quick question: These 3 bluez commits below
(https://chromium.googlesource.com/chromiumos/third_party/bluez/) have
still not been picked up yet by the CQ.

a2f815a UPSTREAM: src/profile: Fix segmentation fault
9091ceb UPSTREAM: core/device: Fix crash when disconnecting
b741078 UPSTREAM: deviceinfo: Enable external flag

Should we manually update the ebuild file for those?

[chro...@googlecode.com]

Comment #25 on issue 520586 by msa...@chromium.org: CQ published uprev

despite build failure; bluez-5.33-r31 failing in BuildPackages
https://code.google.com/p/chromium/issues/detail?id=520586

David: Arman: Thank you for fixing this bug.

[chro...@googlecode.com]

Updates:
Status: Verified

Comment #26 on issue 520586 by david...@chromium.org: CQ published uprev

despite build failure; bluez-5.33-r31 failing in BuildPackages
https://code.google.com/p/chromium/issues/detail?id=520586

Confirmed that ebuild uprev went through on Wed Oct 21 @ 01:24:41 PST
(presumably just after Francois checked). We've had some problems with
ebuild uprevving lately which phobbs is working on that may have delayed
the uprev, but the uprev did go through.

[chro...@googlecode.com]

Comment #27 on issue 520586 by fbeauf...@chromium.org: CQ published uprev

despite build failure; bluez-5.33-r31 failing in BuildPackages
https://code.google.com/p/chromium/issues/detail?id=520586

I confirm. Thanks David!