Re: [PATCH] tests: avoid spurious failure when gcj is not installed

[Stefano Lattarini]

Hi Jim.

On 02/25/2012 12:44 PM, Jim Meyering wrote:
> I noticed the gcj4 test failing on master and wrote this patch,
> but figured it belonged on the maint branch.
>
Nope, it's a regression introduced by me on master recently.  The maint
branch should unaffected.

BTW, I'm sure I had tested the 'gcj*.test' scripts before pushing, and
didn't find any error on my system...  Nor am I seeing it now...  Oh,
got it: since $GCJ is exported to the empty string in the affect tests,
configure tries to automatically find a gcj compiler, which succeeds
on my system, but fails on systems without a gcj installation.

> Humph.  It doesn't even apply there, due to lack of defs-static.in,
> and in fact the gcj4 test doesn't fail on maint, either.
> 
> Is there a schedule for merging maint into master?
>
The latest maint is already merged into master.  For what concerns the
near future, I just plan not to touch maint anymore since the 1.12
release, at which point we'll just merge master into maint (and fork
off a new master from that).

> From 39c13b5bc88dc2a3e28301e1d1a089bd13089c48 Mon Sep 17 00:00:00 2001
> From: Jim Meyering <address@hidden>
> Date: Sat, 25 Feb 2012 12:37:25 +0100
> Subject: [PATCH] tests: avoid spurious failure when gcj is not installed
> 
> * tests/defs-static.in (GNU_GCJ, GNU_GCJFLAGS): Define.
> * tests/Makefile.am (do_subust): Add those to the list of substituted symbols.
> ---
>  tests/Makefile.am    |    2 ++
>  tests/defs-static.in |    2 ++
>  2 files changed, 4 insertions(+)
> 
The patch is OK, of course.  Extra kudos if you add a reference in the
commit message to the commit where I broke the tests.

Thanks,
  Stefano