[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: program transform and tests

From: Noah Misch
Subject: Re: program transform and tests
Date: Tue, 9 May 2006 11:09:03 -0700
User-agent: Mutt/1.5.6i

On Tue, May 09, 2006 at 06:59:12PM +0200, Ralf Wildenhues wrote:
> * Paul Eggert wrote on Mon, May 01, 2006 at 07:13:27AM CEST:
> > Do people actually use --program-transform-name for autoconf?
> > If so, it's probably worth fixing now, assuming the fixes are
> > mechanical and are not too dangerous.
> Summary: I've looked at this, and have a half-baked patch.  It's doable,
> but a bit ugly, and I haven't found any non-testsuite changes necessary.
> IMHO not useful to have before the release, since all it's about to
> introduce are potential bugs.


> After all this, I ran
>   make installcheck automake=automake-1.5 aclocal=aclocal-1.5
>   make installcheck automake=automake-1.7 aclocal=aclocal-1.7
> on a GNU/Linux system, and found only one failure with both of them:
> |   94:     Unusual Automake input files
> which fails because the Automake macros in that test use newer syntax;
> the test should not be adjusted because it aims to parallel a test in
> Automake's test suite.

Automake 1.8 introduced the feature in question.  We should skip this test for
older Automake.

>       * tests/ (AUTOTEST): Use `$(MY_AUTOM4TE)' instead of
>       `./autom4te' to create `./testsuite', since the `all' target
>       will ensure its presence, but `installcheck' should not create
>       the uninstalled wrappers.

This is clearly helpful; you could apply it independently.

>       to specify..
>       (autom4te, autoconf, autoheader, autoupdate, autoreconf)
>       (autoscan, ifnames, aclocal, automake, ACLOCAL, AUTOMAKE)
>       (AUTOM4TE): ..these new variables to be passed to the testsuite
>       (check-local, installcheck-local):

With this approach, `./testsuite' will no longer behave reasonably.  Logic to
default these variables should appear in `testsuite' itself.  Autotest does not
currently provide an interface for performing such initializations globally; we
would need to add that.

reply via email to

[Prev in Thread] Current Thread [Next in Thread]