Re: [PATCH 1/4] {branch "tests-init"} Tests defs: $srcdir is unconditionally substituted.

[Stefano Lattarini]

On Monday 06 September 2010, Ralf Wildenhues wrote:
> * Stefano Lattarini wrote on Mon, Sep 06, 2010 at 01:05:25AM CEST:
> > On Sunday 05 September 2010, Ralf Wildenhues wrote:
> > > * Stefano Lattarini wrote on Fri, Sep 03, 2010 at 12:14:18AM 
CEST:
> > > > -srcdir=${srcdir-'@abs_srcdir@'}
> > > > +srcdir='@abs_srcdir@'
> > > 
> > > Can you explain why this is not needed conditionally any more?
> > 
> > For consistency with the variable `$top_scrdir' introduced
> > in the PATCH 2/4, which is unconditonally defined.  Also, while
> > in my endless tweaking an experimentation with the testsuite
> > I've wanted to and/or resorted to override many things, I not
> > once needed or wished to override $(srcdir).  And the
> > uncondional definition let us remove # the code below for srcdir
> > "absolutization".
> > 
> > > I mean, was there a reason we didn't already do that last time?
> > 
> > Which last time?
> 
> Well, in v1.11-139-g6fee87c and before that in v1.11-138-gc6bbc45
> the above line was move resp. introduced.  Was that a buglet in the
> latter patch?
No, there's no bug to be fixed that I know of.  It's just that we
really need $testsrcdir to be the *real* absolute path of $(srcdir),
so there's no need to read it from the environment when we can just 
AC_SUBST it (also, currently we have to absolutize $testsrcdir, which 
add a fork for every test).

> If yes, then this patch is fine.
"Unfortunately", there is no bug fixed by this patch.  I can only 
offer the above argumentation (plus consistency of $testsrcdir and
$top_testsrcdir).

> The other updated patches 2 and 3 in this series are fine anyway.
OK, thanks.

Regards,
  Stefano