On Thu, 2011-09-01 at 09:05 +0200, Francois Tigeot wrote: > Hi, > > The latest -master version crashes reliably on NetBSD (i386 and amd64).
We're sure that its the LibO side that changed right, i.e. some earlier version of LibO with same compiler is ok. Seeing as it affects both x86 and x86_64 that would *seem* to rule out a bug in a specific uno bridge seeing as there's separate bridges. Seeing as the test didn't actually say anything about what's wrong, its possibly not something specific that this test "tests", but something more fundamental which is expects to work. Can you build i18npool from scratch ? IIRC there's a saxparser binary there which does far less than the uno test does, but needs certain similar initial working magic smoke to run correctly. Attached is a patch to get the debugger into the testtools tests. With symbols enabled, do "build" in testtools, gdb should wait for you to type run, bt, the usual drill. C.
diff --git a/testtools/source/bridgetest/makefile.mk b/testtools/source/bridgetest/makefile.mk index 3932d04..b5a7043 100644 --- a/testtools/source/bridgetest/makefile.mk +++ b/testtools/source/bridgetest/makefile.mk @@ -143,7 +143,7 @@ runtest : $(DLLDEST)$/uno_types.rdb $(DLLDEST)$/uno_services.rdb makefile.mk \ .IF "$(COM)$(OS)$(CPU)" == "GCCMACOSXP" @echo "Mac OSX PPC GCC fails this test!, likely broken UNO bridge. Fix me." .ELSE - cd $(DLLDEST) && $(AUGMENT_LIBRARY_PATH) $(SOLARBINDIR)/uno \ + cd $(DLLDEST) && $(AUGMENT_LIBRARY_PATH) gdb --args $(SOLARBINDIR)/uno \ -ro uno_services.rdb -ro uno_types.rdb \ -s com.sun.star.test.bridge.BridgeTest -- \ com.sun.star.test.bridge.CppTestObject
_______________________________________________ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice