Re: PATCH: Use AC_DEFUN instead of m4_defun in libtool.m4

[Charles Wilson <libtool at cwilson dot fastmail dot fm>]

H.J. Lu wrote:

> I don't get it. Shouldn't libtool in gcc require lt~obsolete.m4 so
> that it won't be missing in the first place? To me, a missing
> lt~obsolete.m4 is a bug in libtool.

As part of Steve Ellcey's toplevel libtool update, he removed a number 
of lt-* files that were obsolete, and explicitly updated or added the 
following files manually:

libtool.m4
ltmain.sh
ltoptions.m4
ltsugar.m4
ltversion.m4

He forget or didn't know to add lt~obsolete.m4 -- and neither I nor 
anybody monitoring noticed that (a) it was missing, and (b) it 
apparently is required, not optional.

The libtool team can't control when other projects grab the wrong set of 
 files to manually import -- we can only control what libtoolize does, 
if the end-developer uses that tool to manage libtool in their project.


If there is a bug, it is that libtool /mostly/ worked -- instead of 
whining loudly and being totally broken -- when one of its constituent 
files was, by design or accident, missing.

Hence, the patch earlier in this thread which induces 20070701 CVS HEAD 
libtool to do exactly that: whine loudly if lt~obsolete is missing. 
However, gcc is using 20070318 libtool; all gcc needs to do is import 
lt~obsolete.m4 from that date into toplevel -- with associated regen of 
automake/autoconf files in the affected directories.

--
Chuck