On Mon, 2011-06-27 at 20:05 +0200, Julien Nabet wrote:
> Le 27/06/2011 16:57, Caolán McNamara a écrit :
> > On Fri, 2011-06-24 at 20:02 +0200, Julien Nabet wrote:
> >> But when i use the test file, nothing as you can see below :
> >>
> >> $ g++ -D_GLIBCXX_DEBUG attachment.cxx
> >> $ echo hello world | ./a.out
> >> Line is: "hello world"
> > Hmm, hard to track down a reproducible test-case then. Anything in
> > echo hello world | valgrind ./a.out ?
> $ cat test.c
> #include <iostream>
> #include <string>
>
> using namespace std;
>
> int main (int argc, char * const argv[]) {
> string line;
> getline(cin, line);
> cout << "Line is: \"" << line << "\"" << endl;
> return 0;
> }
> $ g++ -D_GLIBCXX_DEBUG test.c
> configure:14449: result: internal
> Are there other tests I could do ?
I reckon this is basically a libstdc++ bug, so we're just looking for
either a reproducer we can submit a bug for, and/or a hackaround for
ourselves in the interim.
So, for a hackaround, I guess we might be looking at a mismatch between
how the internal cppunit is built and the outside world, so to test that
theory, try this patch to the internal cppunit makefile.mk and rm -rf
cppunit/unxlng* and rebuild it and rebuild those sal tests as well,
easiest to just blow them all away.
This might all be futile, be much nicer if we had a small standalone
test-case.
C.
diff --git a/cppunit/makefile.mk b/cppunit/makefile.mk
index 54c55c2..0e588b4 100644
--- a/cppunit/makefile.mk
+++ b/cppunit/makefile.mk
@@ -40,6 +40,10 @@ PATCH_FILES = solarisfinite.patch warnings.patch windows.patch ldflags.patch aix
# warnings.patch: see <https://sourceforge.net/tracker/?func=detail&;
# aid=2912630&group_id=11795&atid=311795>
+.IF "$(ENABLE_DEBUG_STL)" == "TRUE"
+EXTRA_CFLAGS+=-D_GLIBCXX_DEBUG
+.ENDIF
+
.IF "$(OS)" == "WNT"
.IF "$(COM)" == "MSC"
_______________________________________________
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice
