Re: [PATCH] tests: add AM_PROG_AR to help losing archivers

[Peter Rosin]

Stefano Lattarini skrev 2012-01-31 14:31:
> On 01/31/2012 01:55 PM, Peter Rosin wrote:
>> Hi,
>>
>> Here's a couple of missing AM_PROG_AR lines, ok for maint?
>>
> Alas no, because msvc is *not* merged into maint, so that we don't even
> have AM_PROG_AR in maint  :-(

Arrgh.  Good catch!

> You might instead want to merge maint into msvc, then apply this patch
> to msvc, then merge msvc back into branch-1.11 and master.
> 
> (I know, the present organization of branches sucks in some respects;
> we might rethink it after the 1.11.3 release, OK?)

I would love to send maint, msvc and branch-1.11 to some dump
somewhere far away with a one-way ticket.  They have simply
diverged too much.  At least for me, I'm only somewhat up to speed
on the ar-lib mess, but you might be in a different position having
written most (all?) of the other diverging changes.

>> Subject: [PATCH] tests: add AM_PROG_AR to help losing archivers
>>
>> * tests/extradep.test (configure.in): Add AM_PROG_AR.
>> * tests/extradep2.test (configure.in): Likewise.
>>
> By only reading this, I'm not sure whether the change is needed to pacify
> some automake warning, to improve coverage, or to make the tests runnable
> with Microsoft lib as the archiver.  Could you add a paragraph after the
> summary line that explicitly specifies what is the patch motivation?
> 
> ACK with that addressed.

Turns out these AM_PROG_ARs are already on master.  So, the merge from
maint into msvc should probably include them with a manual adjustment.
I tried doing the merges you describe above, but it's too rich for my
stomach.  I get something that works, sort of, by I don't feel good
about it and I have this feeling that some changes leaked into branch-1.11
that are not supposed to be there.  I simply don't feel qualified and
can't assess if my conflict resolutions are good or bad without further
digging.  Which I don't have time for, so I'm leaving those merges for
someone else<tm>.  Sorry.

Regarding the requested extra paragraph in the commit message, is that
really needed?  I think it would be quite a bit of extra work to get an
accurate description of the various failure cases, as I haven't kept the
test suite results and would need to rerun the tests with/without
patches and with varying compiler settings etc.  It's a trivial bug, and
as I said, the code is already there on master.

Cheers,
Peter