Re: mode=link and full path to dependent shared library?

[Ralf Wildenhues]

Hi Tim,

* Tim Mooney wrote on Wed, Oct 12, 2005 at 05:05:15PM CEST:
> In regard to: Re: mode=link and full path to dependent shared library?,...:
> >
> >Seems like an easier fix to use LTLIBINTL than my patch :) Tim, want
> >to submit a patch to the aspell folks?
> 
> :-)  That's more or less what prompted my original question on this list,
> about documentation for how libtool should be handling it.  I wanted to
> know if it was libtool that was doing something incorrectly, or aspell.
> I'm still not sure.  ;-)

Let's put it this way: Not handling absolute-path-shared-deplibs
correctly is an unfortunately undocumented libtool limitation.

We'd like to eliminate this limitation eventually, or alternatively
document why it should not be used.  Given that a better solution
(which is simple at the same time) exists for your issue, that should
be used; it takes pressure from us to take our time to fix this.

So, to conclude: aspell is wrong (because LTLIBINTL is documented),
and libtool could be doing things better.  (But the latter isn't
news to anyone.)

> What I haven't yet tested is what LTLIBINTL looks like on systems that
> have libintl.so but don't have a libintl.la (such as systems that don't
> package libtool .la files).  It could be that switching to LTLIBINTL
> won't work for those systems.

Using LTLIBINTL should work just fine.  The important thing is that
`libtool' gets to eat it, not that it was a libtool library.  Anything
else would be a bug in gettext.m4.

> I'll test in the absence of libintl.la and see what happens.

Cheers,
Ralf