In message <[EMAIL PROTECTED]>, Tom Tromey writes:
> >>>>> "Erez" == Erez Zadok <[EMAIL PROTECTED]> writes:
>
> Erez> Akim, you could help folks like me by distributing some sort of
> Erez> wrapper script that does the following:
>
> Erez> - run make test
>
> It should be "make check".
Right. My bad.
> Erez> - for every failed test, run the debug*.sh script and record its output
> Erez> - remember to remove config.log etc. before running each debug*.sh test
> Erez> - create a tar file of everything you need: configure, config.cache,
> Erez> config.log's for each debug script, the debug scripts themselves, and the
> Erez> full output of "make test".
> Erez> - add anything else as needed
> Erez> - uuencode the tar file
>
> I already have a script which I use to automatically test GNU packages
> which come my way. I'd like it if we could standardize on a way to do
> this across packages. Then I could add a bit of code to my script(s)
> and be sure I was generating the report that the package author
> wanted.
>
> For instance, we could add a new Makefile target, like "check-report",
> whose output would be the report.
That would be even better, b/c it'll extend to every package. Question: is
the "make check" interface and behavior expected to be the same for all
packages that use GNU auto*? For example, I've never seen these debug*.sh
scripts generated for libtool and automake?
However, if you guys are going to standardize this, then please do so soon.
Major testing on autoconf will begin shortly, so anything semi-working would
be of help to me. In the mean time, I'd be happy with a simple shell script
that I'd have to execute manually. Thanks.
> Tom
Erez.
