[Ffmpeg-devel] make test-server failure

Luca Abeni lucabe72
Mon Jan 23 10:06:03 CET 2006


Hi all,

On Fri, 2006-01-20 at 17:34 +0100, Luca Abeni wrote:
> On Fri, 2006-01-20 at 16:57 +0100, Luca Abeni wrote:
> > Hi all,
> > 
> > with today's CVS, configured without any special option (just
> > run ./configure), "make test-server" (in the "tests" directory) fails:
> Update:
> CVS from 2005-11-13 seems to be ok ("make test-server" is succesful most
> of the times - it is not deterministic)
> CVS from 2005-11-14 is ko ("make test-server" always fails - but this
> time it is deterministic :)
I have some more updates, but first, can anyone confirm the problem can
be reproduced? Or am I the only one seeing the failure in the tests?
Maybe it depends on my setup... I am using debian unstable.

> The problem does not seem to depend on the gcc version.
I confirm this. I tried all the gcc versions I have, and changing the
compiler does not change the results of the regression tests.

> I'll investigate more on monday, trying to find the commit that broke
> it. Does anyone know if the cvs commit list is archived somewhere?
Ok.... CVSPS is my friend, and it helped me to find the two commits that
broke test-server on my setup. They are
---------------------
PatchSet 4704
Date: 2005/11/13 04:32:12
Author: alex
Branch: HEAD
Tag: (none)
Log:
Support for streaming: dont write indexes and dont signal HAS_INDEX in header. Also set filesize to max in this case.

Members: 
        libavformat/avienc.c:1.113->1.114

and

---------------------
PatchSet 4705
Date: 2005/11/13 04:33:26
Author: alex
Branch: HEAD
Tag: (none)
Log:
updated regression reference

Members:
        tests/ffserver.regression.ref:1.13->1.14


If I revert these two commits, the regression test for ffserver can
succeed. I say "can succeed" because "make test-server" is not
deterministic: with the two commits reverted, "make test-server"
succeeds most of the times, but sometimes fails with
host3:/tmp/A/ffmpeg/tests# make test-server
./videogen 'vsynth1/'
Launch commandline: ../ffmpeg -loop -flags +bitexact -dct fastint -idct simple -y -f pgmyuv -i vsynth1/%02d.pgm http://127.0.0.1:9999/feed1.ffm 
ffserver started.
New connection: GET /feed1.ffm
127.0.0.1 - - [Mon Jan 23 18:51:08 2006] "GET /feed1.ffm HTTP/1.0" 404 149
Waiting for feeds to startup...
New connection: GET /test.jpg?date=19700101T000000Z
New connection: GET /test.mjpg?date=19700101T000000Z
192.168.25.4 - - [Mon Jan 23 18:51:10 2006] "GET /test.jpg?date=19700101T000000Z HTTP/1.0" 200 66
192.168.25.4 - - [Mon Jan 23 18:51:10 2006] "GET /test.mjpg?date=19700101T000000Z HTTP/1.0" 200 107
New connection: GET /test.swf?date=19700101T000000Z
192.168.25.4 - - [Mon Jan 23 18:51:10 2006] "GET /test.swf?date=19700101T000000Z HTTP/1.0" 200 122
New connection: GET /test_h.asf?date=19700101T000000Z
192.168.25.4 - - [Mon Jan 23 18:51:10 2006] "GET /test_h.asf?date=19700101T000000Z HTTP/1.0" 200 702
New connection: GET /test_h.avi?date=19700101T000000Z
192.168.25.4 - - [Mon Jan 23 18:51:10 2006] "GET /test_h.avi?date=19700101T000000Z HTTP/1.0" 200 300
New connection: GET /test_h.mpg?date=19700101T000000Z
192.168.25.4 - - [Mon Jan 23 18:51:10 2006] "GET /test_h.mpg?date=19700101T000000Z HTTP/1.0" 404 150
New connection: GET /test_h.rm?date=19700101T000000Z
192.168.25.4 - - [Mon Jan 23 18:51:10 2006] "GET /test_h.rm?date=19700101T000000Z HTTP/1.0" 200 521
New connection: GET /test_l.asf?date=19700101T000000Z
192.168.25.4 - - [Mon Jan 23 18:51:10 2006] "GET /test_l.asf?date=19700101T000000Z HTTP/1.0" 200 702
New connection: GET /test_l.avi?date=19700101T000000Z
192.168.25.4 - - [Mon Jan 23 18:51:10 2006] "GET /test_l.avi?date=19700101T000000Z HTTP/1.0" 200 300
New connection: GET /test_l.mpg?date=19700101T000000Z
192.168.25.4 - - [Mon Jan 23 18:51:10 2006] "GET /test_l.mpg?date=19700101T000000Z HTTP/1.0" 404 150
New connection: GET /test_l.rm?date=19700101T000000Z
192.168.25.4 - - [Mon Jan 23 18:51:10 2006] "GET /test_l.rm?date=19700101T000000Z HTTP/1.0" 200 521
New connection: GET /test_small.jpg?date=19700101T000000Z
192.168.25.4 - - [Mon Jan 23 18:51:10 2006] "GET /test_small.jpg?date=19700101T000000Z HTTP/1.0" 200 66
New connection: GET /teststat.html?date=19700101T000000Z
192.168.25.4 - - [Mon Jan 23 18:51:10 2006] "GET /teststat.html?date=19700101T000000Z HTTP/1.0" 200 4070
--- data/ffserver.regression    2006-01-23 18:51:10.705883088 +0100
+++ /tmp/A/ffmpeg/tests/ffserver.regression.ref 2006-01-23 18:37:58.278350384 +0100
@@ -1,6 +1,6 @@
 d41d8cd98f00b204e9800998ecf8427e  ff-test.jpg
 6ce0af834ee0abf4753df419f67648c5  ff-test.mjpg
-0c4803b8a7e8e7beba1b0fb3979cc5bd  ff-test.swf
+0f1795f0496f37aeea6272d272140efa  ff-test.swf
 f842864a654593ea9ee9d7a1bd024627  ff-test_h.asf
 4357b110362c3d4d60ad14ffc5d09550  ff-test_h.avi
 d41d8cd98f00b204e9800998ecf8427e  ff-test_h.mpg

Server regression test: Error.
make: *** [test-server] Error 1


I recall that without reverting the two commits cited above
"make test-server" always fails. I have not been able to find any
version for which "make test-server" never fails.


That's all. I do not know ffserver, nor I use it (I am reporting the
failure and try to track down the problem just to help the project)...
So, I do not really know how to fix the regression tests for ffserver. I
hope the information I provided can help, and I am more than willing to
provide more information or do more tests on request.


				Luca
-- 
_____________________________________________________________________________
Copy this in your signature, if you think it is important:
                               N O    W A R ! ! !

 
 
 --
 Email.it, the professional e-mail, gratis per te: http://www.email.it/f
 
 Sponsor:
 Web Marketing: impara tutti i segreti del Posizionamento sui Motori di
* Ricerca, Pubblicita' on-line, Email Marketing ?- clicca qui
 Clicca qui: http://adv.email.it/cgi-bin/foclick.cgi?mid=3213&d=23-1





More information about the ffmpeg-devel mailing list