Re: An updated version of proposed-migration is available to review

2020-07-16 Thread Steve Langasek
On Thu, Jul 16, 2020 at 06:24:11PM +0200, Matthias Klose wrote:

> the new update_excuses hides the information for successful autopkg tests
> which makes the page more compact.  I dislike that when searching for
> failed autopkg tests which failed for some triggers, but succeeded with
> others.  With the old output you were able to just grep update_excuses to
> look for those tests, which currently doesn't work anymore.  Please
> restore the old behavior, or if that's considered to be too big, then show
> successful autopkg tests when there are also ones which are failing with
> certain triggers.

I agree, the hiding of "irrelevant" information to make the page smaller has
gone too far.  When there is an autopkgtest regression on one architecture,
it is useful to be able to quickly see the corresponding history for the
tests on other architectures, without having to walk autopkgtest.ubuntu.com
to find the information.

Could you file a bug on https://bugs.launchpad.net/britney to track this?

-- 
Steve Langasek   Give me a lever long enough and a Free OS
Debian Developer   to set it on, and I can move the world.
Ubuntu Developer   https://www.debian.org/
slanga...@ubuntu.com vor...@debian.org


signature.asc
Description: PGP signature
-- 
Ubuntu-release mailing list
Ubuntu-release@lists.ubuntu.com
Modify settings or unsubscribe at: 
https://lists.ubuntu.com/mailman/listinfo/ubuntu-release


Re: An updated version of proposed-migration is available to review

2020-07-16 Thread Matthias Klose
On 6/16/20 7:07 PM, Iain Lane wrote:
> Over the last few weeks, I've been working on rebasing our extensive 
> delta to proposed-migration. It's now at a state where it's ready for 
> others to take a look at. Please check out the output from a dry-run 
> (being re-run hourly from cron)
> 
>   
> https://people.canonical.com/~ubuntu-archive/laney/proposed-migration/update_excuses.html
>   
> https://people.canonical.com/~ubuntu-archive/laney/proposed-migration/update_output.txt
> 
> There are some changes worth highlighting, all of which are courtesy of 
> the upstream developers - thanks to them.
> 
>   - Installability of Build-Depends is now considered for migrating.
>   - Skipped / flaky tests are now reported as "No test results" instead 
> of "pass". Treating these as passes created buggy situations before, 
> where skip -> fail was treated as a regression. That won't happen 
> any more.
>   - proposed-migration will now try to work out if a test requests needs 
> more triggers (packages from proposed) for the test-deps to be 
> installable, and add them if so. This should make more tests, 
> particularly for transitions, work first time and not require 
> retrying.
>   - More uninstallability is detected at 'excuses' time, rather than 
> 'output' time. That makes it easier to interpret.
> 
> If you can see anything that's *wrong* in the output linked above, 
> please let me know. If you run any scripts which parse the yaml, please 
> try them against
> 
>   
> https://people.canonical.com/~ubuntu-archive/laney/proposed-migration/update_excuses.yaml
> 
> and ideally adapt them as necessary.
> 
> Providing no issues crop up, I'll be looking to deploy this next week.

the new update_excuses hides the information for successful autopkg tests which
makes the page more compact.  I dislike that when searching for failed autopkg
tests which failed for some triggers, but succeeded with others.  With the old
output you were able to just grep update_excuses to look for those tests, which
currently doesn't work anymore.  Please restore the old behavior, or if that's
considered to be too big, then show successful autopkg tests when there are also
ones which are failing with certain triggers.

Matthias

-- 
Ubuntu-release mailing list
Ubuntu-release@lists.ubuntu.com
Modify settings or unsubscribe at: 
https://lists.ubuntu.com/mailman/listinfo/ubuntu-release


[DEPLOYED] Re: An updated version of proposed-migration is available to review

2020-07-10 Thread Iain Lane
On Thu, Jul 02, 2020 at 04:53:48PM +0100, Iain Lane wrote:
> Just looking into that, but it will delay the rollout into next week.

All done, this is now live. Do let me know if anything looks wrong.

-- 
Iain Lane  [ i...@orangesquash.org.uk ]
Debian Developer   [ la...@debian.org ]
Ubuntu Developer   [ la...@ubuntu.com ]


signature.asc
Description: PGP signature
-- 
Ubuntu-release mailing list
Ubuntu-release@lists.ubuntu.com
Modify settings or unsubscribe at: 
https://lists.ubuntu.com/mailman/listinfo/ubuntu-release


Re: An updated version of proposed-migration is available to review

2020-07-02 Thread Iain Lane
On Tue, Jun 30, 2020 at 10:20:18AM +0100, Iain Lane wrote:
> I'm thinking that I'll make the cut over on Thursday UK time, so 
> please have a look at the output before then and check your 
> .yaml-parsing scripts against the new output (location changed since 
> my initial post; it's now .xz compressed)
> 
>   
> https://people.canonical.com/~ubuntu-archive/laney/proposed-migration/update_excuses.yaml.xz

Quick update on this - I enabled some test runs for stable series (focal 
& eoan) just to be sure that it worked there, and - would you believe 
it! - the thing crashes.

Just looking into that, but it will delay the rollout into next week.

Cheers,

-- 
Iain Lane  [ i...@orangesquash.org.uk ]
Debian Developer   [ la...@debian.org ]
Ubuntu Developer   [ la...@ubuntu.com ]


signature.asc
Description: PGP signature
-- 
Ubuntu-release mailing list
Ubuntu-release@lists.ubuntu.com
Modify settings or unsubscribe at: 
https://lists.ubuntu.com/mailman/listinfo/ubuntu-release


Re: An updated version of proposed-migration is available to review

2020-06-30 Thread Iain Lane
On Tue, Jun 16, 2020 at 06:07:09PM +0100, Iain Lane wrote:
> Over the last few weeks, I've been working on rebasing our extensive 
> delta to proposed-migration. It's now at a state where it's ready for 
> others to take a look at. Please check out the output from a dry-run 
> (being re-run hourly from cron)
> 
>   
> https://people.canonical.com/~ubuntu-archive/laney/proposed-migration/update_excuses.html
>   
> https://people.canonical.com/~ubuntu-archive/laney/proposed-migration/update_output.txt

Update on this: Testing found a couple of bugs. I think I've fixed them.

They were mainly around arch:all handling. In Debian, these packages are 
usually built on separate buildds, and britney had made a couple of 
assumptions that relied on this.

I'm thinking that I'll make the cut over on Thursday UK time, so please 
have a look at the output before then and check your .yaml-parsing 
scripts against the new output (location changed since my initial post; 
it's now .xz compressed)

  
https://people.canonical.com/~ubuntu-archive/laney/proposed-migration/update_excuses.yaml.xz

Cheers,

-- 
Iain Lane  [ i...@orangesquash.org.uk ]
Debian Developer   [ la...@debian.org ]
Ubuntu Developer   [ la...@ubuntu.com ]


signature.asc
Description: PGP signature
-- 
Ubuntu-release mailing list
Ubuntu-release@lists.ubuntu.com
Modify settings or unsubscribe at: 
https://lists.ubuntu.com/mailman/listinfo/ubuntu-release


Re: An updated version of proposed-migration is available to review

2020-06-17 Thread Sebastien Bacher
Hey Iain, thanks for the work!

Le 16/06/2020 à 19:07, Iain Lane a écrit :
> If you can see anything that's *wrong* in the output linked above, 
> please let me know. If you run any scripts which parse the yaml, please 
> try them against
>
>   
> https://people.canonical.com/~ubuntu-archive/laney/proposed-migration/update_excuses.yaml
>
> and ideally adapt them as necessary.

(That's a .xz compressed version which is a bit confusing)

The by-team-report is unhappy with the new file, it doesn't like the
policy_info/autopkgtest section having no package listed, e.g from gcc-9

  policy_info:
    autopkgtest:
  verdict: REJECTED_TEMPORARILY

Skipping those cases as a test workaround gives a report where bug
references are buggy (listed as #0).

I plan to work on those issues tomorrow


Cheers,
Sebastien Bacher



-- 
Ubuntu-release mailing list
Ubuntu-release@lists.ubuntu.com
Modify settings or unsubscribe at: 
https://lists.ubuntu.com/mailman/listinfo/ubuntu-release


Re: An updated version of proposed-migration is available to review

2020-06-17 Thread Iain Lane
On Tue, Jun 16, 2020 at 07:49:18PM +0100, Iain Lane wrote:
> I could maybe make a diff of subsets of the yaml, if that would be 
> helpful. Perhaps even just a diff of the candidates, and then people 
> can manually go back and inspect why something does or does not 
> migrate when it didn't/did before.

Done:

  
https://people.canonical.com/~ubuntu-archive/laney/proposed-migration/update_excuses_diff.html

The current and new proposed-migration runs are not synced - they run on 
slightly different archive state - please bear this in mind when looking 
at that diff if you see things that aren't there but 'should' be, or 
differences in state like test results.

(rough-n-ready code: 
https://gist.github.com/iainlane/2607b426837fe885a1760e9639871a9c)

One thing this made me notice that I could have mentioned in the OP:

  - Sets of autopkgtest results where there is no action required (all 
pass, all alwaysfailed) are *omitted from the HTML output*, as you 
don't need to think about them. That could be a bit confusing if 
you're used to the current way.

Cheers,

-- 
Iain Lane  [ i...@orangesquash.org.uk ]
Debian Developer   [ la...@debian.org ]
Ubuntu Developer   [ la...@ubuntu.com ]


signature.asc
Description: PGP signature
-- 
Ubuntu-release mailing list
Ubuntu-release@lists.ubuntu.com
Modify settings or unsubscribe at: 
https://lists.ubuntu.com/mailman/listinfo/ubuntu-release


Re: An updated version of proposed-migration is available to review

2020-06-16 Thread Steve Langasek
On Wed, Jun 17, 2020 at 12:15:29AM +0100, Dimitri John Ledkov wrote:
> On Tue, 16 Jun 2020 at 19:23, Steve Langasek  
> wrote:

> > Personally, my desire to manually review proposed-migration output is low :)
> > Is it possible to get diffs between the old and new output, over a few
> > iterations, that we could review in order to identify any behavior changes?

> diff of html2text output

> https://people.canonical.com/~xnox/update_excuses.diff.txt

> 9MiB big your welcome?

Was really looking for a logical diff of the yaml tbh, not of the html.

-- 
Steve Langasek   Give me a lever long enough and a Free OS
Debian Developer   to set it on, and I can move the world.
Ubuntu Developer   https://www.debian.org/
slanga...@ubuntu.com vor...@debian.org


signature.asc
Description: PGP signature
-- 
Ubuntu-release mailing list
Ubuntu-release@lists.ubuntu.com
Modify settings or unsubscribe at: 
https://lists.ubuntu.com/mailman/listinfo/ubuntu-release


Re: An updated version of proposed-migration is available to review

2020-06-16 Thread Iain Lane
On Tue, Jun 16, 2020 at 11:23:12AM -0700, Steve Langasek wrote:
> Hi Iain,
> 
> Personally, my desire to manually review proposed-migration output is low :)
> Is it possible to get diffs between the old and new output, over a few
> iterations, that we could review in order to identify any behavior changes?

I tried, but the full diff has way too many changes to be readable. It's 
worse than if you just page through update_excuses.html and look at the 
items individually.

I could maybe make a diff of subsets of the yaml, if that would be 
helpful. Perhaps even just a diff of the candidates, and then people can 
manually go back and inspect why something does or does not migrate when 
it didn't/did before.

Cheers,

-- 
Iain Lane  [ i...@orangesquash.org.uk ]
Debian Developer   [ la...@debian.org ]
Ubuntu Developer   [ la...@ubuntu.com ]


signature.asc
Description: PGP signature
-- 
Ubuntu-release mailing list
Ubuntu-release@lists.ubuntu.com
Modify settings or unsubscribe at: 
https://lists.ubuntu.com/mailman/listinfo/ubuntu-release