ubuntu-phone team mailing list archive
-
ubuntu-phone team
-
Mailing list archive
-
Message #08798
Re: Landing team 26.06.14
-
To:
ubuntu-phone@xxxxxxxxxxxxxxxxxxx
-
From:
Jamie Strandboge <jamie@xxxxxxxxxxxxx>
-
Date:
Fri, 27 Jun 2014 13:17:57 -0500
-
In-reply-to:
<CA+RtrqncuRK2CYr=4rt566PQ1UzRY=erueGBuaVoSFQyhEMaNA@mail.gmail.com>
-
User-agent:
Mozilla/5.0 (X11; Linux x86_64; rv:24.0) Gecko/20100101 Thunderbird/24.6.0
On 06/27/2014 11:10 AM, Leo Arias wrote:
> Hello,
>
> On Fri, Jun 27, 2014 at 8:21 AM, Jussi Pakkanen <jussi.pakkanen@xxxxxxxxxxxxx
> <mailto:jussi.pakkanen@xxxxxxxxxxxxx>> wrote:
>
> Indeed. Here's an MR for that (it just disables the test):
>
> https://code.launchpad.net/~__jpakkane/mediascanner2/__sqliteworkaround/+merge/224845
> <https://code.launchpad.net/~jpakkane/mediascanner2/sqliteworkaround/+merge/224845>
>
> What this means in practice is that queries with only a few letters won't
> return results. Longer queries keep on working just like before. The proper
> fix is to dive into SQLite and see if the old behaviour can be re-enabled
> but that's a lot more work.
>
>
> We don't want to have an artificial dashboard that's always green. If something
> is broken, we want to know about it.
>
I don't think Jussi was suggesting we disable the test and forget about it.
Right now the ftbfs is blocking the apparmor fix to allow the mediascanner to
scan in certain situations, which is potentially blocking image promotion. If it
weren't for this no-code-change rebuild ftbfs, no one here would have seen this
issue, because there is no image test for it, yet it would be present on the
image. Considering that and since there are currently no bugs reported against
sqlite3 for this, there would be no known regression against the image and
therefore no promotion blocker and therefore the image cannot be artificially
green. Furthermore, this issue probably would not have been discussed until the
next time mediascanner2 or some other package was built-- but that presumably
wouldn't be an emergency update.
In other words, while I agree that disabling a test is not the first choice and
almost never the correct permanent choice, a carefully thought out decision may
show it to be the right temporary choice. Disabling this test allows us to push
a package that suffers from the same sqlite3 regression as what is currently in
the archive, yet it fixes an issue that is potentially blocking image promotion.
To make sure this is not lost, I have filed a bug:
https://bugs.launchpad.net/ubuntu/+source/sqlite3/+bug/1335281
I created two tasks: one for sqlite3 (and assigned to doko per Steve's guidance
in this thread) and one for mediascanner2 to reenable the test (and assigned to
Jussi). I added the 'rtm14' tag.
Thanks
PS- I was coordinating this landing and landed it in proposed before I read the
comments in this thread.
--
Jamie Strandboge http://www.ubuntu.com/
Attachment:
signature.asc
Description: OpenPGP digital signature
Follow ups
References