UnixWare 7 g++ testsuite Failures: Ugly std_cwchar.h patch.

rbrown64@csc.com.au rbrown64@csc.com.au
Wed Nov 7 00:54:00 GMT 2001

This ugly patch addresses part of what Robert Lipe describes as
"the ancient problem in libstdc++ of running configure with
liberal flags and then turning on "-ansi" (thereby potentially reducing
the available namespace to that of a conforming implementation) later on
during the testsuite run.".
Note, it has NOT yet survived a full bootstrap - just a make check-c++ .
It won't help if <wchar.h> is included before the C++ headers.

Is it worth it just to silence testsuite failures?

LAST_UPDATED: Mon Oct 29 21:42:45 UTC 2001
Native configuration is i686-unknown-sysv5UnixWare7.1.0

2001-10-07  Rodney Brown  <rbrown64@csc.com.au>
     * include/c_std/bits/std_cwchar.h: #undef __STRICT_ANSI__ before
     including <wchar.h> to avoid warnings for v{f,s,}wscanf on UnixWare 7.

--- libstdc++-v3/include/c_std/bits/std_cwchar.h.orig   Wed Sep  5 07:42:24 2001
+++ libstdc++-v3/include/c_std/bits/std_cwchar.h   Wed Nov  7 17:33:02 2001
@@ -40,7 +40,15 @@

 #pragma GCC system_header
+#if __STRICT_ANSI__
+#undef   __STRICT_ANSI__
 #include <wchar.h>

 // Need to do a bit of trickery here with mbstate_t as char_traits

(Original run)
                === g++ Summary for unix ===

# of expected passes            6553
# of unexpected failures        93
# of unexpected successes       4
# of expected failures          123
# of untested testcases         15

(After fixincluding away - or using the patch above).
FAIL: g++.law/profile1.C (test for excess errors)
FAIL: g++.other/enum5.C (test for excess errors)
XPASS: g++.pt/decl2.C ICE - (test for errors, line 8)
XPASS: g++.robertl/eb17.C  Execution test

                === g++ Summary ===

# of expected passes            6673
# of unexpected failures        2
# of unexpected successes       2
# of expected failures          96
# of untested testcases         15

More information about the Gcc-patches mailing list