svn commit: r276962 - head/lib/liblzma

[Baptiste Daroussin]

On Sun, Jan 11, 2015 at 11:27:44PM +0200, Konstantin Belousov wrote:
> On Sun, Jan 11, 2015 at 09:25:02AM +0000, Baptiste Daroussin wrote:
> > Author: bapt
> > Date: Sun Jan 11 09:25:01 2015
> > New Revision: 276962
> > URL: https://svnweb.freebsd.org/changeset/base/276962
> > 
> > Log:
> >   Do not regenerate and install liblzma.pc when only build libraries
> >   aka do not regenerate while generating 32bits libs
> >   
> >   Reported by:	antoine
> > 
> > Modified:
> >   head/lib/liblzma/Makefile
> > 
> > Modified: head/lib/liblzma/Makefile
> > ==============================================================================
> > --- head/lib/liblzma/Makefile	Sun Jan 11 08:40:01 2015	(r276961)
> > +++ head/lib/liblzma/Makefile	Sun Jan 11 09:25:01 2015	(r276962)
> > @@ -148,12 +148,13 @@ CFLAGS+=	-DSYMBOL_VERSIONING
> >  
> >  CLEANFILES+=	liblzma.pc
> >  
> > +.if !defined(LIBRARIES_ONLY)
> >  all: liblzma.pc
> I think that much better define to differentiate on would be COMPAT_32BIT.
> 
> That said, head and stable/10 should be almost clean now for allowing
> ports and third-party software to work with cc -m32.  I think
> that this commit makes it in other direction.
> 
> Could the .pc files for base system provided by some port ?
> 
I think the best would probably be to create a bsd.pc.mk which act a bit like
bsd.incs.mk

Concerning pushing all of those .pc files into a port, I'm not against it,
actually it would be even simpler for ports as it will affect all the ports at
once. But on the other hands that would mean on port with different content
based on OSVERSIONS, I think keeping it in base would be more accurate.

Best regards,
Bapt
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 181 bytes
Desc: not available
URL: <http://lists.freebsd.org/pipermail/svn-src-all/attachments/20150111/b539d4d3/attachment.sig>