gbuild subsequentcheck is clean

classic Classic list List threaded Threaded
5 messages Options
Bjoern Michaelsen Bjoern Michaelsen
Reply | Threaded
Open this post in threaded view
|

gbuild subsequentcheck is clean

Hi all,

subsequentcheck for the gbuild modules is clean. Just type

  make -f GNUmakefile.mk -srkj30 gb_COLOR=t subsequentcheck

in the source root and be amazed by the blinkenlights(*). It should
complete without any errors or hickups.

The dirty secret is of course that I had to disable a few tests for
that. You will find a list of them here:

 https://bugs.freedesktop.org/buglist.cgi?cmdtype=runnamed&namedcmd=subsequenttests

So, if you are looking for a place to hack, these might be good
starting points as they show clearly reproducible bugs (and one of
them leads right into a crash).

The good thing about this is, as soon as there are errors popping up
when running subsequentcheck now, we know them to be regressions and
should take care of them!

Best Regards,

Bjoern

(*) Well, as said: you need to build smoketest before.

--
https://launchpad.net/~bjoern-michaelsen

_______________________________________________
LibreOffice mailing list
[hidden email]
http://lists.freedesktop.org/mailman/listinfo/libreoffice

signature.asc (501 bytes) Download Attachment
Michael Meeks Michael Meeks
Reply | Threaded
Open this post in threaded view
|

Re: gbuild subsequentcheck is clean

Hi Bjoern,

On Fri, 2011-03-25 at 17:37 +0100, Bjoern Michaelsen wrote:
> subsequentcheck for the gbuild modules is clean. Just type

        Ooh - fun :-)

>   make -f GNUmakefile.mk -srkj30 gb_COLOR=t subsequentcheck
>
> in the source root and be amazed by the blinkenlights(*). It should
> complete without any errors or hickups.

        I guess we should add this to the end of the generic all: rule for
compilation, as we should with the smoketest IMHO (at least for Linux
where it runs headless).

        Having said that - it is somewhat annoying all of the graphical thrash
that this introduces: I end up with lots of flickering windows, and some
seem to just hang there ;-) Did I really break -headless somehow with
the new oostart.bin ? [ it works when I try it manually ], and/or do we
need to tweak the tests so they pass that ?

> The dirty secret is of course that I had to disable a few tests for
> that. You will find a list of them here:
>
>  https://bugs.freedesktop.org/buglist.cgi?cmdtype=runnamed&namedcmd=subsequenttests
>
> So, if you are looking for a place to hack, these might be good
> starting points as they show clearly reproducible bugs (and one of
> them leads right into a crash).

        Ooh ! nice - can you add that to the easy hacks ? and we also have a
3.4 blocker bug here: https://bugs.freedesktop.org/show_bug.cgi?id=35673
that might be nice to have those tracked on.

> The good thing about this is, as soon as there are errors popping up
> when running subsequentcheck now, we know them to be regressions and
> should take care of them!

        Right. The only problem is the graphical thrash I guess.

        Anyhow - great to see this in-place.

        Thanks,

                Michael.

--
 [hidden email]  <><, Pseudo Engineer, itinerant idiot

_______________________________________________
LibreOffice mailing list
[hidden email]
http://lists.freedesktop.org/mailman/listinfo/libreoffice
Bjoern Michaelsen Bjoern Michaelsen
Reply | Threaded
Open this post in threaded view
|

Re: gbuild subsequentcheck is clean

Hi Michael,

On Fri, 25 Mar 2011 17:22:12 +0000
Michael Meeks <[hidden email]>
wrote:

> Having said that - it is somewhat annoying all of the
> graphical thrash that this introduces: I end up with lots of
> flickering windows, and some seem to just hang there ;-) Did I really
> break -headless somehow with the new oostart.bin ? [ it works when I
> try it manually ], and/or do we need to tweak the tests so they pass
> that ?

Well, I would prefer to have the tests not run headless as there is
quite a bit of vcl/X11 code coverage that might be cut off doing that. I
usually run those tests with DISPLAY set to a local vnc session (which
I lurk into from time to time in viewonly mode).

> Ooh ! nice - can you add that to the easy hacks ?

Done.

> and we also have a 3.4 blocker bug here:
> https://bugs.freedesktop.org/show_bug.cgi?id=35673 that might be nice
> to have those tracked on.

Done.

Subsequenttests should soon work on the old build system too. However,
I havent blacklisted/disabled the broken tests there yet. I will add
another Easy Task for that.

Best Regards,

Bjoern

--
https://launchpad.net/~bjoern-michaelsen

_______________________________________________
LibreOffice mailing list
[hidden email]
http://lists.freedesktop.org/mailman/listinfo/libreoffice

signature.asc (501 bytes) Download Attachment
Bjoern Michaelsen Bjoern Michaelsen
Reply | Threaded
Open this post in threaded view
|

Re: gbuild subsequentcheck is clean

Hi Michael,

On Fri, 25 Mar 2011 19:49:29 +0100
Bjoern Michaelsen
<[hidden email]> wrote:

> > and we also have a 3.4 blocker bug here:
> > https://bugs.freedesktop.org/show_bug.cgi?id=35673 that might be
> > nice to have those tracked on.
>
> Done.

I just reverted that because of fierce resistance. ;)

Instead I created an own metabug:

 https://bugs.freedesktop.org/show_bug.cgi?id=35690

That one is added as a "see also:" url to 35673.

> Subsequenttests should soon work on the old build system too. However,
> I havent blacklisted/disabled the broken tests there yet. I will add
> another Easy Task for that.

Running the "subsequenttests" command now works. The easy task for
disabling failing tests is here:

 http://wiki.documentfoundation.org/Development/Easy_Hacks#Disable_failing_subsequent_tests_and_file_bugs_for_them


BR,

Bjoern

--
https://launchpad.net/~bjoern-michaelsen

_______________________________________________
LibreOffice mailing list
[hidden email]
http://lists.freedesktop.org/mailman/listinfo/libreoffice

signature.asc (501 bytes) Download Attachment
Bjoern Michaelsen Bjoern Michaelsen
Reply | Threaded
Open this post in threaded view
|

subsequenttests clean (was: gbuild subsequentcheck is clean)

Hi all,

after disabling another bunch of tests (most of them for 35693),
running the "subsequenttest" command should now complete without errors
on master (with the tests in the old and in the new build system).
Anything popping up there now is a regression.

Best Regards,

Bjoern


(*) https://bugs.freedesktop.org/show_bug.cgi?id=35693
--
https://launchpad.net/~bjoern-michaelsen

_______________________________________________
LibreOffice mailing list
[hidden email]
http://lists.freedesktop.org/mailman/listinfo/libreoffice

signature.asc (501 bytes) Download Attachment