Re: [lmi] PATCH: Submodule changes

[Vadim Zeitlin]

On Tue, 13 Oct 2020 00:32:32 +0000 Greg Chicares <gchicares@sbcglobal.net> 
wrote:

GC> On 2020-10-09 16:14, Vadim Zeitlin wrote:
GC> > On Fri, 9 Oct 2020 15:46:20 +0000 Greg Chicares <gchicares@sbcglobal.net> 
wrote:
GC> [...]
GC> > GC> * posix_fhs.make:
GC> > GC> 
GC> > GC> -platform_gnome_xml_libraries := \
GC> > GC> +platform_xml_libraries := \
GC> > GC> +  $(shell xmlwrapp-config --libs) \
GC> > GC>    -lexslt \
GC> > GC>    $(shell xslt-config --libs) \
GC> > GC>    $(shell xml2-config --libs) \
GC> > GC> 
GC> > GC> The unchanged '-lexslt' line seems incongruous. Do you happen to know
GC> > GC> whether it's superfluous? I can look into it; I just thought you might
GC> > GC> know off the top of your head.
GC> > 
GC> >  Strangely enough, I don't think it's superfluous, i.e. xslt-config
GC> > behaviour doesn't seem to have changed since the post from October 2001
GC> > linked from a comment just above: it still doesn't output -lexslt.
GC> 
GC> Does xsltwrapp actually use anything provided by '-lexslt',

 At least formally, it does, because it calls exsltRegisterAll(), provided
by libexslt, when initializing the library. I can't be absolutely sure if
calling this function is required, but I think it is.

GC> > GC> * 
https://github.com/vadz/lmi/pull/162/commits/be12e2e2621189d0d90a3387cfd4c807b8b0dc68
GC> > GC> 
GC> > GC> "Consistently use the same config.guess in all build scripts"
GC> > GC> 
GC> > GC> Later, I'll want to look for all other uses of 'config.guess'.
GC> > GC> I remember wrestling with this problem and choosing something like
GC> > GC>   $(/usr/share/libtool/build-aux/config.guess)
GC> > GC> elsewhere because it seemed like the least bad way,
GC> > 
GC> >  I think we've discussed it and I thought /usr/share/misc/config.guess
GC> > might be a better choice on a Debian system, but it's probably also a less
GC> > portable one.
[...]
GC> /usr/share/misc/ would seem to be the right place according to
GC>   https://refspecs.linuxfoundation.org/FHS_3.0/fhs/ch04s11.html
GC> and, on my debian system at least, it's the referent of the
GC> libtool/build-aux symlink:
GC> 
GC> $ls -l /usr/share/libtool/build-aux/config.guess 
/usr/share/misc/config.guess
GC> lrwxrwxrwx 1 root root    23 Aug 28  2019 
/usr/share/libtool/build-aux/config.guess -> ../../misc/config.guess
GC> -rwxr-xr-x 1 root root 44283 Feb 24  2018 /usr/share/misc/config.guess
GC> 
GC> so I went ahead and changed it everywhere in commit 1f4ed3ccf5.

 Thanks, I agree it's tidier to use the one here.

GC> >  Summary: I don't think there is anything to do here urgently right now,
GC> > but I could make the suggested changes to msw_common.make slightly later.
GC> > Please let me know if I should.
GC> 
GC> Actually, it's probably easier for me to do it: the change is
GC> trivial, while testing it on an ever-changing corporate server
GC> is the part that takes time--but I'm already trying to get that
GC> server to work again, so it's a good time to make little changes.
GC> My only question is whether 'libexslt' is even wanted.

 I'm almost sure that it is. I could try looking into it more and checking
what exactly breaks if I remove exsltRegisterAll() call if you'd like me
to, but, again, something surely will.

 Regards,
VZ

pgpsyVWAWVNNm.pgp
Description: PGP signature