From: Alan W. Irwin <airwin@us...> - 2006-08-07 18:58:20
Update of /cvsroot/plplot/plplot/cmake/modules
In directory sc8-pr-cvs1.sourceforge.net:/tmp/cvs-serv15459
Improve commentary about the caddr_t situation.
RCS file: /cvsroot/plplot/plplot/cmake/modules/plplot.cmake,v
retrieving revision 1.25
retrieving revision 1.26
diff -u -d -r1.25 -r1.26
--- plplot.cmake 5 Aug 2006 22:05:20 -0000 1.25
+++ plplot.cmake 7 Aug 2006 18:58:18 -0000 1.26
@@ -129,12 +129,17 @@
-# Some X11 headers require "caddr_t" even on systems that claim POSIX.1
-# compliance, which is illegal. This makes it impossible to compile
-# programs that include X11 headers if _POSIX_SOURCE is defined. I work
-# around this potential problem by just defining caddr_t to 'char *' on all
-# systems (unless it is set already), whether it will be needed or not.
+# In the past, some X11 headers required "caddr_t" even on systems that
+# claimed POSIX.1 compliance, which was illegal. This made it impossible
+# to compile programs that included X11 headers if _POSIX_SOURCE was
+# defined. We used to work around this potential problem by just defining
+# caddr_t to 'char *' on all systems (unless it is set already), whether
+# it was needed or not. Now we ignore the issue because we don't expect
+# such broken X behaviour any more and because this kind of argument list
+# for AC_CHECK_TYPE is now deprecated in the autoconf documentation.
+# Do not implement the equivalent of this since commented out in the ABS
# AC_CHECK_TYPE(caddr_t, char *)
# Test signal handler return type (mimics AC_TYPE_SIGNAL)
Get latest updates about Open Source Projects, Conferences and News.