Re: multilib improvement?

[Peter O'Gorman]

On Oct 21, 2006, at 8:52 PM, Ralf Wildenhues wrote:

> Hello Peter, Bob,
>
> * Bob Friesenhahn wrote on Wed, Oct 18, 2006 at 06:13:39PM CEST:
> > On Wed, 18 Oct 2006, Peter O'Gorman wrote:
> > > This is what I came up with to take -m64/-m32 into account when  
> > > using gcc.
> >
> > I am not able to spot a problem with it.  Extensive testing will  
> > prove
> > if it works as expected.
>
> Except of course that no test case was added to confirm this, so there
> is no way you can know it's fixed except for hoping that somebody will
> test it before the release (or you do it yourself).
>
> Taking the idealistic hat off, how can we test something like this
> usefully?  I think one idea may be like this: Try to link against an
> installed library in one of the search paths with -L$path.  If that
> succeeds, then omitting -L$path should succeed as well.  It's quite
> a weak test; too weak for most possible errors, but at least I think
> it should not have false negatives.  Do you have better ideas?

No, the way I "tested" it was to run ./configure; make with CC set to  
gcc -m32 and gcc -m64 on a few systems and look at  
sys_lib_search_path_spec manually after each run to see if it looked  
sane. Your patch is better than that :)


> And of course the patch is missing a NEWS addition, this having been
> a bug that bothered many people.
>
> OK to apply (NEWS also to branch-1-5)?
> I haven't had a chance to test your actual patch yet, by the way.

Looks okay on visual inspection. Please apply. Would appreciate  
testing of the patch too when you get around to it.

Thanks,
Peter