Testsuite, Bug Reporting, GNATS

[Mark Mitchell]

>>>>> "Gerald" == Gerald Pfeifer <pfeifer@dbai.tuwien.ac.at> writes:

    Gerald> Mark,

    Gerald> as promised, I am now converting the guidelines you
    Gerald> suggested in
    Gerald> http://gcc.gnu.org/ml/gcc/2000-06/msg00541.html to HTML
    Gerald> for addition to our web site. In the course of this I
    Gerald> noticed an issue that we might want to rethink:

    Gerald>   If you find a bug, but you're not fixing it:

    Gerald>     - Create a test-case.  - Check in the test-case,
    Gerald> marking it as XFAIL.  - Write a bug report, referencing
    Gerald> the test-case.  - Put the bug report in GNATS.

    Gerald> In my opinion, this involves extraneous work in that one
    Gerald> has to submit the bug two times: Once to gcc-bugs, and
    Gerald> once to GNATS. In addition, then there'll be to places one
    Gerald> has to check for responses!

I think we talked about this once before.  The test-case check-in, and
the gcc-bugs posting, are optional.

    Gerald> In addition to gcc-bugs we also have a read-only list
    Gerald> gcc-prs which receives such reports from GNATS. Do these
    Gerald> really have to be separate lists?

No.  I do think it would be a good idea if gcc-prs is fed to gcc-bugs. 

--
Mark Mitchell                   mark@codesourcery.com
CodeSourcery, LLC               http://www.codesourcery.com