directory-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Emmanuel Lecharny <elecha...@iktek.com>
Subject RE: Philosophy of failing tests
Date Sun, 27 Feb 2005 17:16:37 GMT
Hi guys !

coming back from a little break, I saw many messages about a patch I
submited two weeks ago.

here are my thought and position :
- as it was a bug in the base code, it deserved a JIRA and possibly a
patch
- as this issue was about a misconception of the ASN.1 decoding, it
appeared that tests were also cursed.
- Unit tests are supposed to validate the correctness of the patch, and
of the code.

So I don't think that applying a patch without correcting the tests is a
good idea. The pb is just to be able to replay the whole tests, not only
in the project where the bug was from, but also from other projects that
have dependences on it (a fixed asn1 bug can have a huge impact on whole
code).

Whatever, as far as I remember, I submitted this patch with some tests
fixe, but it seems to be not enough. I will check it again, and try to
fix the issues, do a full build, and submit all the patches. I just need
a feedback of anybody on the list before applying them, because I don't
want to introduce errors in the existing code ! (call it a mentor...)

Thanks a lot,
Emmanuel







Mime
View raw message