Re: [PATCH]: Cross building and Cygwin fixes.

[Jan Nieuwenhuizen]

Marius Vollmer <address@hidden> writes:

> Thanks!  I have applied it (with some changes, see below) to the
> 'unstable' branch.  It is too late for 1.6.

Ah, that's very nice.  Being too late is only a small setback, I'm
glad not having to maintain a patch against CVS; as I'm maintaining
Guile officially for Cygwin now.

> I changed this to just
>
>     GUILE_FOR_BUILD='$(preinstguile)'
>
> but did not change am/preinstguile.  This keeps the roles of the
> 'pre-inst-guile' separate from the 'guile-for-build'.  pre-inst-guile
> is used during "make check" for example, and guile-for-build is used
> for the doc snarfing.

Ok.

>> +test -n "$with_guile_for_build" && GUILE_FOR_BUILD="$with_cc_for_build"
>
> "GUILE" instead of "CC", no?

Sure, thanks.

>> +2002-07-03  Jan Nieuwenhuizen  <address@hidden>
>> +
>> +    * pre-inst-guile (preinstguile): Initialize with
>> +    $(GUILE_FOR_BUILD) configure parameter.
>
> I didn't make this change, see above.  Instead, I defined
> snarf2checkedtexi as
>
>   GUILE="$(GUILE_FOR_BUILD)" $(top_srcdir)/scripts/snarf-check-and-output-texi
Ok, looks good.  I'll have a look at CVS today.

>> -#ifdef HAVE_WINSOCK2_H
>> +#if defined (HAVE_WINSOCK2_H) \
>> +  && !(defined (__CYGWIN32__) || defined (__CYGWIN__))
>>  #include <winsock2.h>
>>  #endif
>
> What is the reason for this change?

The reason is that Cygwin has the winsock header, but it also has unix
sockets.  We don't want winsock.
  
> Wouldn't it be better to fix the check for <winsock2.h> in
> configure.in?

Possibly.  But then it's not a HAVE_WINSOCK2, but rather a
WANT_WINSOCK2.  If configure.in only changes 'have' into want if not
Cygwin, that would maybe be cleaner.

Jan.

-- 
Jan Nieuwenhuizen <address@hidden> | GNU LilyPond - The music typesetter
http://www.xs4all.nl/~jantien       | http://www.lilypond.org