On Mon, 2011-10-10 at 12:36 +0200, Petr Mladek wrote: > Hi, > > I add Yi Fan into CC who actively works on LO QA. > > Michael Meeks píše v Pá 07. 10. 2011 v 20:49 +0100: > > Hi Drew, > > > > On Fri, 2011-10-07 at 13:28 -0400, drew wrote: > > > I read over some material regarding the opensuse.openQA goings on ... > > > > Wow, encouraging research there. > > > > > http://susestudio.org/a/EvUJ20/lo-sbs-libre-34x > > Looks pretty interesting. I have few questions: > > Is is supposed for testing of the openSUSE LO packages or the plain LO > build?
*chuckling*...good question - it is currently carrying the Suse(ified) version of LibO. Vanilla (plain) LibO would make more sense and my preference. > Is the testing automatic or manual? It doesn't do any testing *yet* (I doubt [know, as the config script isn't written yet] it will even load properly today) - but when I first thought about something like this the plan was to have a set of automatic tests and then have it as a resource for manual testing. > > I wonder how to synchronize it with the other testing effort. Note that > test cases and test runs for the plain build are organized via Litmus, > see http://wiki.documentfoundation.org/QA#Function_test . Yes, I'm aware of that tool - read over the wiki pages, again, this weekend. It seemed to me that there would be a connection between the two, or to say that I would use the litmus system to document the test cases and the VM to execute them. Actually in the first incantation of thought I had planned on having a very pronounced split between a set of back end images and front end images so that one could build a testing system to cover multiple vendor releases of OO.o [1] - I did get to where I had a number of scripts working for some basic test, the scripts where all OOoBasic and exercised a set of OO.o API functions (though OpenSolaris proved bothersome and then I kind of lost my cool on the whole OO.o experience, but that is neither here nor there) - I still think this is a reasonable way to go, though I might change the scripting language from Basic to something else perhaps, in the end the API is the constant. - This API testing is IMO certainly useful but for GUI stuff I did not have a good solution. Sounds like the OpenQA system could step in here. I have some old experience with automated GUI testing tools and I have to say they never left a good taste in my mouth, primarily because of false positives, but that experience is old and hopefully much of that problem has been fixed by folks much smarter then me. > Test cases > for the openSUSE LO packages are in Testopia that is part of > https://bugzilla.novell.com. I am not sure if people outside SUSE have > access to it, though. > > Also I wonder where to report bugs. I would prefer > https://bugs.freedesktop.org/ if the bugs are in the plain build as > well. Where else? (so yes) > > > I wonder how we get the latest LibreOffice versions into there ? > > whether we build them, or install packages or something ? Bernhard > > and/or Petr - what is the easiest way to get say a nightly snapshot into > > an openSUSE virtual machine for automated / openQA testing ? > > One possibility would be to use night builds from the plain build. > Alternatively, I could try to produce nightly builds in BS. I already > have many useful scripts at hands. > so bottom line - I'm willing to put some time and effort into this - and open to suggestions on how best to integrate into what others are doing.. Best wishes, //drew [1] http://drewjensen.typepad.com/blog/2009/12/meet-sybil.html _______________________________________________ List Name: Libreoffice-qa mailing list Mail address: Libreoffice-qa@lists.freedesktop.org Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/ Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette List archive: http://lists.freedesktop.org/archives/libreoffice-qa/
