Andreas Schwab <[EMAIL PROTECTED]> wrote:
> What is gl_ACL_GET_FILE supposed to test? It doesn't look like the test
> is matching its description.
Thanks. That was a bug.
It's supposed to do what the comment says:
# gl_ACL_GET_FILE(IF-WORKS, IF-NOT)
# -------------------------------------
# If `acl_get_file' works (does not have a particular bug),
# run IF-WORKS, otherwise, IF-NOT.
# This tests for a Darwin 8.7.0 bug, whereby acl_get_file returns NULL,
# but sets errno = ENOENT for an existing file or directory.
I've fixed it like this in gnulib:
2006-12-12 Jim Meyering <[EMAIL PROTECTED]>
* m4/acl.m4 (gl_ACL_GET_FILE): Fix logic error.
Reported by Andreas Schwab <[EMAIL PROTECTED]>.
Index: m4/acl.m4
===================================================================
RCS file: /sources/gnulib/gnulib/m4/acl.m4,v
retrieving revision 1.8
diff -u -p -r1.8 acl.m4
--- m4/acl.m4 26 Sep 2006 10:30:06 -0000 1.8
+++ m4/acl.m4 12 Dec 2006 15:29:43 -0000
@@ -53,7 +53,7 @@ AC_DEFUN([gl_ACL_GET_FILE],
#include <errno.h>
]],
[[return !! (!acl_get_file (".", ACL_TYPE_ACCESS)
- || errno == ENOENT);]])],
+ && errno == ENOENT);]])],
[gl_cv_func_working_acl_get_file=yes],
[gl_cv_func_working_acl_get_file=no],
[gl_cv_func_working_acl_get_file=cross-compiling])])
_______________________________________________
Bug-coreutils mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/bug-coreutils
