#55 make regress fails at [45011]

latest_mercurial
open
spadix
None
5
2011-05-14
2011-05-13
No

While trying to test my patch I discovered pure hg clone (r 311) does not pass the regress-decoder test.

[45009] SEED=613829541
[45010] SEED=-441921086
[45011] SEED=-1770279485
MISSING CODE-39:QGJXY48
SEED=-1770279485
make: *** [regress-decoder] Aborted

My configuration is:
autoreconf --install && ./configure --without-gtk --without-qt --without-python && make regress-decoder

Discussion

  • spadix

    spadix - 2011-05-14

    > hg clone (r 311)

    fyi, you should always refer to a revision hash (probably 3e11458a41a0 in this case); the sequence number is only useful for a specific clone...

    > MISSING CODE-39:QGJXY48
    > SEED=-1770279485

    Unfortunately I'm not able to reproduce the error; apparently the random sequence on this platform does not match what you're getting. I'll have to try another box later...

    You can avoid the problem, which should be unrelated to your testing, by starting with a different seed, eg:

    $ test/test_decode -s 2 -n 100000

    and running the image regression separately

    $ make regress-images

     
  • spadix

    spadix - 2011-05-14
    • milestone: --> latest_mercurial
    • assigned_to: nobody --> spadix
     
  • Jarek Czekalski

    Jarek Czekalski - 2011-05-14

    hg id: 3e11458a41a0 tip

    With your command line I get:

    [11549] SEED=-114709036
    [11550] SEED=1131458861
    [11551] SEED=-203299454
    decode EAN-2:93
    UNEXPECTED!
    SEED=-203299454
    Aborted

    I wanted to include pcard testing in test_decode, but at the moment it seems to be not possible. We should use some deterministic random number generator to avoid such problems. There are definitely random number generators available for c language, as I was using them 10 years ago or so at the university.

    Until something's done with the generator, I'll skip test_decode. But I have to admit I detected one error by running test_decode so I'll be glad to finish this part of testing. I was returning ZBAR_PARTIAL when I shouldn't do that and it conflicted with other decoders.

     
  • spadix

    spadix - 2011-05-16

    Sorry that you're not able to make the test work.

    We're actually not going to spend much more time developing infrastructure for this test in c; but that's a separate project...

     

Log in to post a comment.

Get latest updates about Open Source Projects, Conferences and News.

Sign up for the SourceForge newsletter:

JavaScript is required for this form.





No, thanks