----- "James Laska" jlaska@redhat.com wrote:
See attached patch to convey the stand-alone idea noted above.
Ok, I see your point now. Yes, we can do it like that. We will create "stand-alone" script check_upgradepath.py, that will depend on autoqa (therefore be able to use autoqa library and repoinfo.conf), but can be run alone like:
$ ./check_upgradepath.py --envrs blobby-0.9b-1.fc13 --kojitag dist-f13-updates
I'm not sure how large the benefit is over calling it through the hook, but surely the CLI interface is a little simpler than:
$ autoqa post-bodhi-update --title blobby-0.9b-1.fc13 --target-tag dist-f13-updates blobby-0.9b-1.fc13 --local -t upgradepath
Both should achieve the same functionality.
On Fri, 2010-08-20 at 09:38 -0400, Kamil Paral wrote:
----- "James Laska" jlaska@redhat.com wrote:
See attached patch to convey the stand-alone idea noted above.
Ok, I see your point now. Yes, we can do it like that. We will create "stand-alone" script check_upgradepath.py, that will depend on autoqa (therefore be able to use autoqa library and repoinfo.conf), but can be run alone like:
$ ./check_upgradepath.py --envrs blobby-0.9b-1.fc13 --kojitag dist-f13-updates
I'm not sure how large the benefit is over calling it through the hook, but surely the CLI interface is a little simpler than:
$ autoqa post-bodhi-update --title blobby-0.9b-1.fc13 --target-tag dist-f13-updates blobby-0.9b-1.fc13 --local -t upgradepath
You know, I completely forgot about this method of executing tests against --local. Sounds like I need to reacquaint myself with this approach. If running a test is just a matter of 1) yum install autoqa, 2) autoqa $hook --args ... that's not too bad. We already have the argument handling/parsing being done by autoqa and the hook. Apologies, I never thought of this method.
Wwoods, what do you think?
Thanks, James
On Fri, 2010-08-20 at 10:23 -0400, James Laska wrote:
On Fri, 2010-08-20 at 09:38 -0400, Kamil Paral wrote:
----- "James Laska" jlaska@redhat.com wrote:
See attached patch to convey the stand-alone idea noted above.
Ok, I see your point now. Yes, we can do it like that. We will create "stand-alone" script check_upgradepath.py, that will depend on autoqa (therefore be able to use autoqa library and repoinfo.conf), but can be run alone like:
$ ./check_upgradepath.py --envrs blobby-0.9b-1.fc13 --kojitag dist-f13-updates
I'm not sure how large the benefit is over calling it through the hook, but surely the CLI interface is a little simpler than:
$ autoqa post-bodhi-update --title blobby-0.9b-1.fc13 --target-tag dist-f13-updates blobby-0.9b-1.fc13 --local -t upgradepath
You know, I completely forgot about this method of executing tests against --local. Sounds like I need to reacquaint myself with this approach. If running a test is just a matter of 1) yum install autoqa, 2) autoqa $hook --args ... that's not too bad. We already have the argument handling/parsing being done by autoqa and the hook. Apologies, I never thought of this method.
Wwoods, what do you think?
Just a quick update from the fedora-qa meeting today [1]. The group agreed that the above command-line method for running tests works just fine. The group felt that stand-alone tests are easier for autoqa newbies to understand, but that the above method integrates nicely with AutoQA and does require a bit less work. The team also concluded that the method should be documented on the wiki somewhere.
1. Joza noted he would file a ticket to add 'How to run autoqa tests' 2. Jlaska noted he would be interested in drafting some content for review
Thanks, James
On Mon, 2010-08-23 at 15:57 -0400, James Laska wrote:
Just a quick update from the fedora-qa meeting today [1]. The group agreed that the above command-line method for running tests works just fine.
..and since that was really the only objection to the patch, it's been applied to master. Thanks again, Vojta!
-w
----- "Will Woods" wwoods@redhat.com wrote:
..and since that was really the only objection to the patch, it's been applied to master. Thanks again, Vojta!
Thanks Will, I believe the new test case should work just fine. There are just 2 outstanding issues, which I believe vaschenb will patch shortly:
1. Schedule upgradepath for running with post-bodhi-hook (modify control.autoqa) 2. Improve output format a little - I requested a minor improvement in the test output so it's easy to see whether and which checks have failed.
On Tue, Aug 24, 2010 at 04:08:11AM -0400, Kamil Paral wrote:
Thanks Will, I believe the new test case should work just fine. There are just 2 outstanding issues, which I believe vaschenb will patch shortly:
- Schedule upgradepath for running with post-bodhi-hook (modify control.autoqa)
Done.
- Improve output format a little - I requested a minor improvement in the test output so it's easy to see whether and which checks have failed.
Done.
Changes are pushed into master (sorry for not sending patch, but I'm to lazy :-), so please don't kill me for pushing into master)
autoqa-devel@lists.fedorahosted.org