The included regex cannot support equivalence classes and multibyte
collation symbols properly. On the other hand it supports 64-bit
regoff_t, which glibc cannot provide without breaking the ABI.
We currently favor the latter, but this is no longer correct since
there's clearly no hope of ever passing the test.
* m4/regex.m4: Disable test for regoff_t size.
---
ChangeLog | 5 +++++
m4/regex.m4 | 11 ++++++++---
2 files changed, 13 insertions(+), 3 deletions(-)
diff --git a/ChangeLog b/ChangeLog
index 836f553..3174aa9 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,8 @@
+2010-09-10 Paolo Bonzini <bonz...@gnu.org>
+
+ regex: Pass the system regex if its only problem is 32-bit regoff_t.
+ * m4/regex.m4: Disable test for regoff_t size.
+
2010-08-26 Erik Faye-Lund <kusmab...@gmail.com>
poll: return immediately on POLLHUP.
diff --git a/m4/regex.m4 b/m4/regex.m4
index 95784e4..c138eb0 100644
--- a/m4/regex.m4
+++ b/m4/regex.m4
@@ -158,12 +158,17 @@ AC_DEFUN([gl_REGEX],
if (! REG_STARTEND)
return 1;
- /* Reject hosts whose regoff_t values are too narrow.
- These include glibc 2.3.5 on hosts with 64-bit ptrdiff_t
- and 32-bit int. */
+#if 0
+ /* It would be nice to reject hosts whose regoff_t values are too
+ narrow (including glibc on hosts with 64-bit ptrdiff_t and
+ 32-bit int), but glibc cannot make this change without breaking
+ its ABI. If this test was left on, support for equivalence
+ classes and multibyte collation symbols would always be broken
+ except when compiling --without-included-regex. */
if (sizeof (regoff_t) < sizeof (ptrdiff_t)
|| sizeof (regoff_t) < sizeof (ssize_t))
return 1;
+#endif
return 0;]])],
[gl_cv_func_re_compile_pattern_working=yes],
--
1.7.1
