Sparc Solaris lacks an int64_t data type. I added a macro to
detect/correct this problem.
Jon Shiring
Btw, in my quick tests, Vorbis successfully decodes ogg files on sparc
solaris.
The following should be added to configure.in after the calls to set
SIZE64
----------------------
dnl this is a test to see if int64_t is defined
dnl this is because AC_CHECK_TYPE breaks on "long long"
AC_MSG_CHECKING(for int64_t)
AC_CACHE_VAL(has_int64_t,
[AC_TRY_RUN([
#include <sys/types.h>
int64_t foo;
int main() {return 0;}
],
has_int64_t=yes,
has_int64_t=no,
has_int64_t=no
)])
AC_MSG_RESULT($has_int64_t)
if test x$has_int64_t = "xno" ; then
AC_DEFINE_UNQUOTED(int64_t, $SIZE64)
fi
--- >8 ----
List archives: http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
> Sparc Solaris lacks an int64_t data type. I added a macro to > detect/correct this problem.Existing autoconf checks should have been dealing with this. I'd rather fix the bug in the existing checks... Monty --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/
> Sparc Solaris lacks an int64_t data type. I added a macro to > detect/correct this problem. > > Jon Shiring > Btw, in my quick tests, Vorbis successfully decodes ogg files on sparc > solaris. > > The following should be added to configure.in after the calls to set > SIZE64Having just looked at this on Solaris 2.5,6 and 7, I can't duplicate the problem. GCC on Solaris and the SunSoft CC installs both deal with int64_t just fine on these boxen... Monty --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-dev-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
> Sparc Solaris lacks an int64_t data type. I added a macro to > detect/correct this problem.OK! After taking the time to look closely at this, now I see why you wrote your own test. Some days I hate autoconf, but at least you provided the answer. Thanks Jon. I've committed a modified version of your autoconf patch (modified to work with BeOS needing a different include) and changed configure to propagate the types by substituting into os_types.h, and finally updated the various tools to use the right defines. I'd appreciate it if folks on different platforms could test build the current CVS mainline to make sure I didn't break things in the cleanup. Monty --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-dev-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
> One quickie that I forgot to mention (I learned this the hard way) -- > AC_PROG_CC likes to automatically set CFLAGS to "-g" if it isn't already > set (I'm not sure what autoconf's rationale is here). Solaris cc > complains that -g conflicts with one of the meta options in -fast, for > example. > > You might want to surround the call to AC_PROG_CC with: > > ---- > cflags_save="$CFLAGS" > AC_PROG_CC > CFLAGS="$cflags_save" > ---- > > This shouldn't be necessary, but it is. :-\Ah, that's probably the reson I was explicity setting CFLAGS to "". I'll add that change too. Monty --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-dev-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.