tests/amstream.c doesn't test anything

Hi,

please take a look at test.winehq.org
http://test.winehq.org/data/tests/amstream:amstream.html

It appears that almost all native machines perform 0 tests on amstream.
As no skip is flagged, results appear all green.
The only exception are Wylda's XP machines. They execute 16 tests
but fail one of them.

Can we conclude that the amstream tests are in a bad shape, despite green colour?
Or that some amstream components are optional and should be installed on more test bots?
(Note that the amstream.dll is present, otherwise the message would be different)

Most Wine machines also execute 0 tests, except Nicolas Le Cam's (nlc)
amstream.c:163: Test marked todo: IDirectDrawMediaStream_CreateSample returned: 80004001
What's different about these machines?

Should I take this to bugzilla? I know next to nothing about
amstream. It is about streaming to audio or a DirectDraw surface.
http://msdn.microsoft.com/en-us/library/dd390909%28v=VS.85%29.aspx

Regards,
 Jörg Höhle

[Nicolas Le Cam]

2011/1/12
> Hi,
>
> please take a look at test.winehq.org
> http://test.winehq.org/data/tests/amstream:amstream.html
>
> It appears that almost all native machines perform 0 tests on amstream.
> As no skip is flagged, results appear all green.
> The only exception are Wylda's XP machines. They execute 16 tests
> but fail one of them.
>
> Can we conclude that the amstream tests are in a bad shape, despite green
> colour?
> Or that some amstream components are optional and should be installed on
> more test bots?
> (Note that the amstream.dll is present, otherwise the message would be
> different)
>
> Most Wine machines also execute 0 tests, except Nicolas Le Cam's (nlc)
> amstream.c:163: Test marked todo: IDirectDrawMediaStream_CreateSample
> returned: 80004001
> What's different about these machines?
>
> Should I take this to bugzilla? I know next to nothing about
> amstream. It is about streaming to audio or a DirectDraw surface.
> http://msdn.microsoft.com/en-us/library/dd390909%28v=VS.85%29.aspx
>
> Regards,
>  Jörg Höhle
>
> Hi Jörg,

The difference is that I'm creating a test.avi symlink to an avi file in the
working directory before launching winetest.
This avi file is also used by tests/avisplitter.c (skip) and
tests/filtergraph.c (no skip) and I'm creating another msrle.avi symlink to
the same file to avoid skipping in tests/mmio.c.

The file used is ED_1024.avi from the Elephant Dream short film. As I'm
using winetest_latest.exe, I also need to use  the -d switch to be able to
skip the temp dir creation and put symlink in it.

BTW, tests/filtergraph.c also tries to use a test.mpg file (no skip), but I
didn't found one under CC when I added this to my automated test script.
I'll add one this evening now that I've just found
http://learn.creativecommons.org/.

[André Hentschel]

Am 12.01.2011 13:16, schrieb [e-mail]:
> Hi,
>
> please take a look at test.winehq.org
> http://test.winehq.org/data/tests/amstream:amstream.html
>
> It appears that almost all native machines perform 0 tests on amstream.
> As no skip is flagged, results appear all green.
> The only exception are Wylda's XP machines. They execute 16 tests
> but fail one of them.
>
> Can we conclude that the amstream tests are in a bad shape, despite green colour?
> Or that some amstream components are optional and should be installed on more test bots?
> (Note that the amstream.dll is present, otherwise the message would be different)
>
> Most Wine machines also execute 0 tests, except Nicolas Le Cam's (nlc)
> amstream.c:163: Test marked todo: IDirectDrawMediaStream_CreateSample returned: 80004001
> What's different about these machines?
>
> Should I take this to bugzilla? I know next to nothing about
> amstream. It is about streaming to audio or a DirectDraw surface.
> http://msdn.microsoft.com/en-us/library/dd390909%28v=VS.85%29.aspx
>
> Regards,
>  Jörg Höhle

I think we at least should print out a "skipped" somehow...
And what about the needed videos? can't we create them???

> > As no skip is flagged, results appear all green.
> > The only exception are Wylda's XP machines. They
> > execute 16 tests but fail one of them.

Hi, i looked at that right now. WyldBOT downloads from FTP current
testsuite + msrle.avi + test.avi and puts those video files into wct
folder.

test.avi is actually only a renamed c:\windows\clock.avi video file,
because i read somewhere, that when you put a test.avi into wct
directory, more tests will be performed. And from following line i
guess, that amstream.c uses test.avi aka clock.avi and fails
there...

static const char fileA[FILE_LEN] = "test.avi";

W.