Re: [xml] Python as a way to unify libxml2 test on Windows and Unix

[Steve Little <steve icradio com>]

Just to throw a spanner in the works, I've been thinking about how to get
some of the regression tests working under VMS. I realise that this
platform is less widely used, and therefore lower priority for
compatibility, but I thought I'd join in the discussion.

Python might be a better solution for VMS than shell scripts and
makefiles, as it is relatively easily available under VMS.

I did also wonder whether it would be possible to specify at least some of
the tests in a meta format, (XML?), specifying which test executable is
to be used, which input arguments it takes, and which file to diff the
output with. It would then be trivial to generate Unix or Windows (or VMS)
test scripts from this format. In order to avoid a dependency on a
preinstalled xsltproc, they could be built before distribution, as with
some of the automake stuff.

I realise this is a somewhat naive approach, since some tests would be too
complex to fit into this framework, and it would be counterproductive to
attempt to design a complex enough format to specify all possible tests,
but I wondered if some simpler tests could work this way?

Just an idea, and probably very over-simplistic, but I thought I'd throw
it out there...

Thanks,

Steve

--
"In real life, as in Grand Opera, arias only make hopeless situations worse."
                                - Kilgore Trout

On Fri, 7 May 2004, Daniel Veillard wrote:

>   While discussing with William Brack last Tuesday he came up with the
> idea that since libxml2 regression testing on Unix used Python, that
> the problem on Windows was makefile and shell portability and that
> Python works on Windows, the right way to make sure we get full regression
> tests on Windows is to actually make them as python scripts instead of
> relying on the shell and make tools.
>   This sounds like the right idea to unify the tests, the only problem
> would be platforms where UNIX make and shell like tools are available
> but where Python is not available.
>
>   Any comment about this ? I'm afraid that trying to build a portable
> and complete test framework on Windows will be nearly impossible otherwise.
>
> Daniel
>
> --
> Daniel Veillard      | Red Hat Desktop team http://redhat.com/
> veillard redhat com  | libxml GNOME XML XSLT toolkit  http://xmlsoft.org/
> http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/
> _______________________________________________
> xml mailing list, project page  http://xmlsoft.org/
> xml gnome org
> http://mail.gnome.org/mailman/listinfo/xml