This is the mail archive of the
mauve-discuss@sources.redhat.com
mailing list for the Mauve project.
Re: using configure to tune tests?
- To: Peter Naulls <peter at erble dot freeserve dot co dot uk>
- Subject: Re: using configure to tune tests?
- From: Brian Jones <cbj at gnu dot org>
- Date: 09 Feb 2001 15:19:46 -0500
- Cc: mauve-discuss at sourceware dot cygnus dot com
- References: <d901b4a4a%peter@erble.freeserve.co.uk>
Peter Naulls <peter@erble.freeserve.co.uk> writes:
> In message <m3r918fo5i.fsf@lyta.haphazard.org>
> Brian Jones <cbj@gnu.org> wrote:
>
> > This is an idea (not a patch I want applied at the moment) that uses
> > configure to determine some property of the host environment and tunes
> > a specific test for that host. Is this acceptable? I didn't include
> > an updated configure here, nor the things autogenerated by
> > aclocal;autoheader, etc. Why are those even checked into mauve?
>
> I think it'd be better on the whole to have any tests in Java.
> But that's just me.
That's a nice sentiment, but how would you test File.listRoots() in
Java? I mean here, how would you ensure that the list of roots given
is indeed correct for the given host, again in Java, without
hardcoding Mauve? Either you're using a known good VM+classlib vs
something else or you indeed need this sort of thing. I haven't been
able to thinking of another alternative.
Brian
--
Brian Jones <cbj@gnu.org>