Hi all You may have noticed that there have been a bunch of updates from <gitbot@bugs.koha-community.org> to bugs in 'Needs Signoff' status. It is running through all bugs in that status, grabbing the patches and applying them, and changing the status if they don't apply clean. If people find this annoying I will turn it off. If people don't I will start on phase 2, which is making it run all the tests after applying the patches, and catch any fails. My idea is that everything in the 'Needs Signoff' should be passing tests and applying cleanly. Chris
Hi, On Wed, Sep 18, 2013 at 4:37 AM, Chris Cormack <chris@bigballofwax.co.nz>wrote:
You may have noticed that there have been a bunch of updates from <gitbot@bugs.koha-community.org> to bugs in 'Needs Signoff' status. It is running through all bugs in that status, grabbing the patches and applying them, and changing the status if they don't apply clean.
I think this is an excellent idea.
If people find this annoying I will turn it off. If people don't I will start on phase 2, which is making it run all the tests after applying the patches, and catch any fails.
I think the annoyance factor would mainly depend on how frequently the bot fails patches incorrectly. I also have a concern about the potential for a new contributor to post a patch in good faith, only to have their first interaction with the community be a bot them that their patch is problematic, but I think we can wait and see how it works out in practice.
My idea is that everything in the 'Needs Signoff' should be passing tests and applying cleanly.
Agreed. One caveat, though -- if it isn't doing so already, the bot should be taking dependent bugs into account when testing whether patches apply. It failing to do so would quickly turn it into an annoyance. Regards, Galen -- Galen Charlton Manager of Implementation Equinox Software, Inc. / The Open Source Experts email: gmc@esilibrary.com direct: +1 770-709-5581 cell: +1 404-984-4366 skype: gmcharlt web: http://www.esilibrary.com/ Supporting Koha and Evergreen: http://koha-community.org & http://evergreen-ils.org
I think it's a great plan, I've already been caught with one minor rebase requirement and was glad to be prompted by the bot instead of wasting a humans time ;) Martin Martin Renvoize Software Engineer, PTFS Europe Ltd Content Management and Library Solutions Skype: Landline: 0203 286 8685 Mobile: 07725985636 http://www.ptfs-europe.com On 18 September 2013 15:31, Galen Charlton <gmc@esilibrary.com> wrote:
Hi,
On Wed, Sep 18, 2013 at 4:37 AM, Chris Cormack <chris@bigballofwax.co.nz>wrote:
You may have noticed that there have been a bunch of updates from <gitbot@bugs.koha-community.org> to bugs in 'Needs Signoff' status. It is running through all bugs in that status, grabbing the patches and applying them, and changing the status if they don't apply clean.
I think this is an excellent idea.
If people find this annoying I will turn it off. If people don't I will start on phase 2, which is making it run all the tests after applying the patches, and catch any fails.
I think the annoyance factor would mainly depend on how frequently the bot fails patches incorrectly. I also have a concern about the potential for a new contributor to post a patch in good faith, only to have their first interaction with the community be a bot them that their patch is problematic, but I think we can wait and see how it works out in practice.
My idea is that everything in the 'Needs Signoff' should be passing tests and applying cleanly.
Agreed. One caveat, though -- if it isn't doing so already, the bot should be taking dependent bugs into account when testing whether patches apply. It failing to do so would quickly turn it into an annoyance.
Regards,
Galen -- Galen Charlton Manager of Implementation Equinox Software, Inc. / The Open Source Experts email: gmc@esilibrary.com direct: +1 770-709-5581 cell: +1 404-984-4366 skype: gmcharlt web: http://www.esilibrary.com/ Supporting Koha and Evergreen: http://koha-community.org & http://evergreen-ils.org
_______________________________________________ Koha-devel mailing list Koha-devel@lists.koha-community.org http://lists.koha-community.org/cgi-bin/mailman/listinfo/koha-devel website : http://www.koha-community.org/ git : http://git.koha-community.org/ bugs : http://bugs.koha-community.org/
+1 for this initiative Just repeating that dependencies could be an issue, and also patches for an older version. It occurs that a patch is only meant for e.g. 3.8.X. If this is in the title of the patch or its report, they could be skipped.
Le 18/09/2013 17:05, Marcel de Rooy a écrit :
+1 for this initiative
Just repeating that dependencies could be an issue, and also patches for an older version. It occurs that a patch is only meant for e.g. 3.8.X. If this is in the title of the patch or its report, they could be skipped. Another possibility: add a field in bugzilla, called "bot: ignore me", that could be checked when one don't want the bot checking a patch.
That would be much easier and secure than trying to find all the cases that the bot should ignore (patch has dependencies -sometimes they are not problematic-, patch applies only on a maintenance version, ...) -- Paul POULAIN - BibLibre http://www.biblibre.com Free & Open Source Softwares for libraries Koha, Drupal, Piwik, Jasper
Hi, On Wed, Sep 18, 2013 at 8:46 AM, Paul Poulain <paul.poulain@biblibre.com>wrote:
Another possibility: add a field in bugzilla, called "bot: ignore me", that could be checked when one don't want the bot checking a patch.
I think this is a good idea.
That would be much easier and secure than trying to find all the cases that the bot should ignore (patch has dependencies -sometimes they are not problematic-, patch applies only on a maintenance version, ...)
Well, ultimately it would be good for the bot to check dependencies and maintenance patches as well, but adding a custom "ignore me" field would reduce any pressure to have the bot become smarter immediately. Regards, Galen -- Galen Charlton Manager of Implementation Equinox Software, Inc. / The Open Source Experts email: gmc@esilibrary.com direct: +1 770-709-5581 cell: +1 404-984-4366 skype: gmcharlt web: http://www.esilibrary.com/ Supporting Koha and Evergreen: http://koha-community.org & http://evergreen-ils.org
On 19 September 2013 04:21, Galen Charlton <gmc@esilibrary.com> wrote:
Hi,
On Wed, Sep 18, 2013 at 8:46 AM, Paul Poulain <paul.poulain@biblibre.com> wrote:
Another possibility: add a field in bugzilla, called "bot: ignore me", that could be checked when one don't want the bot checking a patch.
There is now a field in bugzilla, called Bot Control. If you set it to no status change, the bot will still check and add a comment if the patch is in needs signoff and doesn't apply. But it won't change the status. That way people can at least see straight away if there is a conflict. If you set it to Ignore this bug, then it won't even bother checking. So if you switch a bug back to needs signoff, you can set this flag. Chris
Le 19/09/2013 09:34, Chris Cormack a écrit : >>> Another possibility: add a field in bugzilla, called "bot: ignore me", >>> that could be checked when one don't want the bot checking a patch. > There is now a field in bugzilla, called Bot Control. If you set it to > no status change, the bot will still check and add a comment if the > patch is in needs signoff and doesn't apply. > But it won't change the status. > > That way people can at least see straight away if there is a conflict. > > If you set it to Ignore this bug, then it won't even bother checking. > > So if you switch a bug back to needs signoff, you can set this flag. Isn't this use case problematic: * bug is set to "no status change" * bot running every week (that's the idea isn't it ?) * every week, there will be a new comment saying "the patch does not apply" If no-one reacts, comments & mails will pile quickly isn't it ? -- Paul POULAIN - Associé-gérant Tel : (33) 4 91 81 35 08 http://www.biblibre.com Logiciels Libres pour les bibliothèques et les centres de documentation
On 19 September 2013 20:08, Paul Poulain <paul.poulain@biblibre.com> wrote: > Le 19/09/2013 09:34, Chris Cormack a écrit : >>>> Another possibility: add a field in bugzilla, called "bot: ignore me", >>>> that could be checked when one don't want the bot checking a patch. >> There is now a field in bugzilla, called Bot Control. If you set it to >> no status change, the bot will still check and add a comment if the >> patch is in needs signoff and doesn't apply. >> But it won't change the status. >> >> That way people can at least see straight away if there is a conflict. >> >> If you set it to Ignore this bug, then it won't even bother checking. >> >> So if you switch a bug back to needs signoff, you can set this flag. > Isn't this use case problematic: > * bug is set to "no status change" > * bot running every week (that's the idea isn't it ?) > * every week, there will be a new comment saying "the patch does not apply" > > If no-one reacts, comments & mails will pile quickly isn't it ? > If people aren't going to react then they should set it to Ignore this bug then. Chris
Hi All The bot now will apply any dependencies first, if they apply clean, it will then try to apply the main patches. If it can't apply the dependencies it, will not this on the bug, but not change status, and move on. It also respects 2 settings in the Bot Control flag. If it is Ignore this bug, it will do nothing if it is No status change, it will note that the patches failed to apply but leave it in Needs Signoff. If you have a bug that you are sure is going to conflict a lot, and don't want to rebase it, set it to Ignore this bug. One question, would you like the bot to comment if it applied a patch cleanly? Or just move on to the next one, (as it is doing now). The next thing I plan to add, is having it run the test suite, or at least the 00- tests. Would people like it to mark things failed qa if the tests fail, or just leave a comment? Chris
Hi, On Fri, Sep 20, 2013 at 8:48 PM, Chris Cormack <chris@bigballofwax.co.nz>wrote:
One question, would you like the bot to comment if it applied a patch cleanly? Or just move on to the next one, (as it is doing now). The next thing I plan to add, is having it run the test suite, or at least the 00- tests. Would people like it to mark things failed qa if the tests fail, or just leave a comment?
I think it should make a comment if it successfully applies the patch (or maybe set a custom last-looked-at-by-bot field?). I also think that it should set the bug to failed QA if whatever tests are run don't pass, but I imagine some tweaking will be required so that it doesn't do things like auto-fail a patch that adds a new dependency -- so maybe it should just add a comment at first. Once we're confident that it isn't generating false negatives, then we could have it change the status. Will the bot process bugs again after a suitable interval has passed? A patch that applies today may not apply two months from now if master has moved on but no human has yet looked at the bug. Regards, Galen -- Galen Charlton Manager of Implementation Equinox Software, Inc. / The Open Source Experts email: gmc@esilibrary.com direct: +1 770-709-5581 cell: +1 404-984-4366 skype: gmcharlt web: http://www.esilibrary.com/ Supporting Koha and Evergreen: http://koha-community.org & http://evergreen-ils.org
On 22 September 2013 05:16, Galen Charlton <gmc@esilibrary.com> wrote:
Hi,
On Fri, Sep 20, 2013 at 8:48 PM, Chris Cormack <chris@bigballofwax.co.nz> wrote:
One question, would you like the bot to comment if it applied a patch cleanly? Or just move on to the next one, (as it is doing now). The next thing I plan to add, is having it run the test suite, or at least the 00- tests. Would people like it to mark things failed qa if the tests fail, or just leave a comment?
I think it should make a comment if it successfully applies the patch (or maybe set a custom last-looked-at-by-bot field?). I also think that it should set the bug to failed QA if whatever tests are run don't pass, but I imagine some tweaking will be required so that it doesn't do things like auto-fail a patch that adds a new dependency -- so maybe it should just add a comment at first. Once we're confident that it isn't generating false negatives, then we could have it change the status.
Will the bot process bugs again after a suitable interval has passed? A patch that applies today may not apply two months from now if master has moved on but no human has yet looked at the bug.
Yeah, it looks at anything in the Needs Signoff queue, that isn't set to be ignored. Chris
Regards,
Galen -- Galen Charlton Manager of Implementation Equinox Software, Inc. / The Open Source Experts email: gmc@esilibrary.com direct: +1 770-709-5581 cell: +1 404-984-4366 skype: gmcharlt web: http://www.esilibrary.com/ Supporting Koha and Evergreen: http://koha-community.org & http://evergreen-ils.org
Hi, On Sat, Sep 21, 2013 at 12:39 PM, Chris Cormack <chris@bigballofwax.co.nz>wrote:
Yeah, it looks at anything in the Needs Signoff queue, that isn't set to be ignored.
In that case, if it adds comments saying that it has successfully applied a patch -- which would be good IMO -- in order to not be annoying it should also skip any Needs Signoff bugs that it has already successfully applied in the past X days, where X is something like two weeks or a month. Regards, Galen -- Galen Charlton Manager of Implementation Equinox Software, Inc. / The Open Source Experts email: gmc@esilibrary.com direct: +1 770-709-5581 cell: +1 404-984-4366 skype: gmcharlt web: http://www.esilibrary.com/ Supporting Koha and Evergreen: http://koha-community.org & http://evergreen-ils.org
On 22 September 2013 08:20, Galen Charlton <gmc@esilibrary.com> wrote:
Hi,
On Sat, Sep 21, 2013 at 12:39 PM, Chris Cormack <chris@bigballofwax.co.nz> wrote:
Yeah, it looks at anything in the Needs Signoff queue, that isn't set to be ignored.
In that case, if it adds comments saying that it has successfully applied a patch -- which would be good IMO -- in order to not be annoying it should also skip any Needs Signoff bugs that it has already successfully applied in the past X days, where X is something like two weeks or a month.
I was thinking 2 weeks or if the bug has changed. IE if someone has updated it since it last checked it will try it again. Sound ok? Chris
Hi, On Sat, Sep 21, 2013 at 1:22 PM, Chris Cormack <chris@bigballofwax.co.nz>wrote:
I was thinking 2 weeks or if the bug has changed. IE if someone has updated it since it last checked it will try it again.
Sounds good. Regards, Galen -- Galen Charlton Manager of Implementation Equinox Software, Inc. / The Open Source Experts email: gmc@esilibrary.com direct: +1 770-709-5581 cell: +1 404-984-4366 skype: gmcharlt web: http://www.esilibrary.com/ Supporting Koha and Evergreen: http://koha-community.org & http://evergreen-ils.org
participants (5)
-
Chris Cormack -
Galen Charlton -
Marcel de Rooy -
Martin Renvoize -
Paul Poulain