Problem detecting libxslt

[Vincent Béron]

Hi Mike,

Doing ./configure on a RH8 box I get the following:
checking libxslt/transform.h presence... yes
configure: WARNING: libxslt/transform.h: present but cannot be compiled
configure: WARNING: libxslt/transform.h:     check for missing
prerequisite headers?
configure: WARNING: libxslt/transform.h: see the Autoconf documentation
configure: WARNING: libxslt/transform.h:     section "Present But Cannot
Be Compiled"
configure: WARNING: libxslt/transform.h: proceeding with the
preprocessor's result
configure: WARNING: libxslt/transform.h: in the future, the compiler
will take precedence
configure: WARNING:     ## ------------------------------------ ##
configure: WARNING:     ## Report this to [e-mail] ##
configure: WARNING:     ## ------------------------------------ ##

The problem looks to come from my transform.h needing xsltInternals.h
being included first.

I added an #ifdef HAVE_LIBXSLT_PATTERN_H include to the AC_CHECK_HEADERS
which defines HAVE_LIBXSLT, as that file does #include
"xsltInternals.h". I don't think we should directly include
xsltInternals.h. Could you please comment if the attached patch is
correct or not? I'll send it to wine-patches if acceptable.

Thank you,
Vincent

[Mike McCormack]

Vincent Béron wrote:
> --- configure.ac	3 Sep 2005 15:43:53 -0000	1.403
> +++ configure.ac	8 Sep 2005 23:08:04 -0000
> @@ -469,7 +469,10 @@ then
>              [AC_DEFINE(HAVE_LIBXSLT, 1, [Define if you have the libxslt library])
>               XSLTLIBS="$ac_xslt_libs"
>               XSLTINCL="$ac_xslt_cflags"],,$ac_xslt_libs)
> -
>    ])
> -
>    ],,
>
> +[#ifdef HAVE_LIBXSLT_PATTERN_H
> +# include <libxslt/pattern.h>
> +#endif])
>      CPPFLAGS="$ac_save_CPPFLAGS"
>  fi

It doesn't seem exactly right to me, but I'm not really sure of a better 
way of fixing it.  Maybe an AC_TRY_COMPILE()?   Alexandre understands 
configure better than I do, so you might try asking him.

Mike

[Alexandre Julliard]

Mike McCormack  writes:
> It doesn't seem exactly right to me, but I'm not really sure of a better 
> way of fixing it.  Maybe an AC_TRY_COMPILE()?   Alexandre understands 
> configure better than I do, so you might try asking him.

Looks good to me.