| Summary: | -fvisibility, with some GCC's causes problems with STL pools due to missing visibility push/pop pragmas in STL includes | ||
|---|---|---|---|
| Product: | [Unmaintained] arts | Reporter: | Iskren Stoyanov <i.stoyanov> |
| Component: | general | Assignee: | Stefan Westerfeld <stefan> |
| Status: | RESOLVED FIXED | ||
| Severity: | grave | CC: | alexandergleich, anthony, belred, bjorn, bugskde, caleb, ciocile, dehaven-radio, firelight951, fpitt_83, geekboy, hephooey_dev, i.stoyanov, ibc, ivo, kdebugsystem, madman, maksim, marcelo, mschiff, mueller, nilsarne, nitro.120, pbernstein, post, raphink, rauge, raul.redondo, sdkovacs, sjnovick, stefan.asserhall, stefan.vunckx, steveahart, tdfischer, Timorgenius, vesa.hytonen |
| Priority: | NOR | ||
| Version: | unspecified | ||
| Target Milestone: | --- | ||
| Platform: | Compiled Sources | ||
| OS: | Linux | ||
| Latest Commit: | Version Fixed In: | ||
| Sentry Crash Report: | |||
| Attachments: | testcase sources + makefile | ||
I forgot to mention arts version it is 1.5 from svn Try with a sane gcc version, that is 3.x. *** Bug 113365 has been marked as a duplicate of this bug. *** *** Bug 114650 has been marked as a duplicate of this bug. *** *** Bug 113364 has been marked as a duplicate of this bug. *** *** Bug 114706 has been marked as a duplicate of this bug. *** mpyne and uga both confirmed it. I think this one is a showstopper, will try to get a test for this on the weekend *** Bug 114802 has been marked as a duplicate of this bug. *** which distribution do you use? please make your distribution vendor apply the patch from http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19664 which adds visibility support to libstdc++. without that your toolchain is broken. Dirk: kubuntu is one of the affected distros. Created attachment 13110 [details]
testcase sources + makefile
Here is a testcase for the problem. It is a bit brittle, as it relies on
non-public problems, but if someone can find a static inline function in the
STL that's spec-mandated, that can be fixed. I haven't fully confirmed this on
a patched system, though, posting it for a wider dessimination..
I guess std::numeric_limits<short>::epsilon will do. *** Bug 114916 has been marked as a duplicate of this bug. *** I'd filed <a href="http://bugs.kde.org/show_bug.cgi?id=114916"><s>Bug 14196<s></a> which was apparantly a duplicate of this, however 10 minutes of searching failed to get me here, hence the duplicate But I digress... There's supposed to be a workaround here? What is it? Sorry, thought it was here. Please see http://bugs.kde.org/show_bug.cgi?id=114650#c7 *** Bug 115083 has been marked as a duplicate of this bug. *** SVN commit 474432 by mueller:
disable -fvisibility=hidden for all those broken distro's out there
CCBUG: 111115
M +20 -1 acinclude.m4.in
--- branches/KDE/3.4/kde-common/admin/acinclude.m4.in #474431:474432
@@ -3286,7 +3286,26 @@
[
AC_BEFORE([AC_PATH_QT_1_3], [KDE_ENABLE_HIDDEN_VISIBILITY])
+ AC_MSG_CHECKING([grepping for visibility push/pop in headers])
+
if test "x$GXX" = "xyes"; then
+ AC_LANG_SAVE
+ AC_LANG_CPLUSPLUS
+ AC_EGREP_CPP(
+ [GCC visibility push],
+ [ #include <cassert>
+ ],
+ [
+ AC_MSG_RESULT(yes)
+ kde_stdc_visibility_patched=yes ],
+ [
+ AC_MSG_RESULT(no)
+ AC_MSG_WARN([Your libstdc++ doesn't appear to be patched for
+ visibility support. Disabling -fvisibility=hidden])
+
+ kde_stdc_visibility_patched=no ])
+ AC_LANG_RESTORE
+
kde_have_gcc_visibility=no
KDE_CHECK_COMPILER_FLAG(fvisibility=hidden,
[
@@ -3314,7 +3333,7 @@
]
)
- if test x$kde_cv_val_qt_gcc_visibility_patched = "xyes"; then
+ if test x$kde_stdc_visibility_patched = "xyes" && test x$kde_cv_val_qt_gcc_visibility_patched = "xyes"; then
CXXFLAGS="$CXXFLAGS -fvisibility=hidden"
KDE_CHECK_VISIBILITY_GCC_BUG
HAVE_GCC_VISIBILITY=1
SVN commit 474433 by mueller:
disable -fvisibility=hidden for all broken distro's out there
CCBUG: 111115
M +20 -1 acinclude.m4.in
--- branches/KDE/3.5/kde-common/admin/acinclude.m4.in #474432:474433
@@ -3353,7 +3353,26 @@
[
AC_BEFORE([AC_PATH_QT_1_3], [KDE_ENABLE_HIDDEN_VISIBILITY])
+ AC_MSG_CHECKING([grepping for visibility push/pop in headers])
+
if test "x$GXX" = "xyes"; then
+ AC_LANG_SAVE
+ AC_LANG_CPLUSPLUS
+ AC_EGREP_CPP(
+ [GCC visibility push],
+ [ #include <cassert>
+ ],
+ [
+ AC_MSG_RESULT(yes)
+ kde_stdc_visibility_patched=yes ],
+ [
+ AC_MSG_RESULT(no)
+ AC_MSG_WARN([Your libstdc++ doesn't appear to be patched for
+ visibility support. Disabling -fvisibility=hidden])
+
+ kde_stdc_visibility_patched=no ])
+ AC_LANG_RESTORE
+
kde_have_gcc_visibility=no
KDE_CHECK_COMPILER_FLAG(fvisibility=hidden,
[
@@ -3381,7 +3400,7 @@
]
)
- if test x$kde_cv_val_qt_gcc_visibility_patched = "xyes"; then
+ if test x$kde_stdc_visibility_patched = "xyes" && test x$kde_cv_val_qt_gcc_visibility_patched = "xyes"; then
CXXFLAGS="$CXXFLAGS -fvisibility=hidden"
KDE_CHECK_VISIBILITY_GCC_BUG
HAVE_GCC_VISIBILITY=1
SVN commit 474434 by mueller:
disable -fvisibility=hidden if distribution is broken
BUG: 111115
M +30 -8 acinclude.m4.in
--- trunk/KDE/kde-common/admin/acinclude.m4.in #474433:474434
@@ -2997,16 +2997,38 @@
[
AC_REQUIRE([AC_PATH_QT])
+ AC_MSG_CHECKING([grepping for visibility push/pop in headers])
+
if test "x$GXX" = "xyes"; then
+ AC_LANG_SAVE
+ AC_LANG_CPLUSPLUS
+ AC_EGREP_CPP(
+ [GCC visibility push],
+ [ #include <cassert>
+ ],
+ [
+ AC_MSG_RESULT(yes)
+ kde_stdc_visibility_patched=yes ],
+ [
+ AC_MSG_RESULT(no)
+ AC_MSG_WARN([Your libstdc++ doesn't appear to be patched for
+ visibility support. Disabling -fvisibility=hidden])
+
+ kde_stdc_visibility_patched=no ])
+ AC_LANG_RESTORE
+
kde_have_gcc_visibility=no
- KDE_CHECK_COMPILER_FLAG(fvisibility=hidden,
- [
- kde_have_gcc_visibility=yes
- CXXFLAGS="$CXXFLAGS -fvisibility=hidden"
- KDE_CHECK_VISIBILITY_GCC_BUG
- HAVE_GCC_VISIBILITY=1
- AC_DEFINE_UNQUOTED(__KDE_HAVE_GCC_VISIBILITY, "$HAVE_GCC_VISIBILITY", [define to 1 if -fvisibility is supported])
- ])
+
+ if test x$kde_stdc_visibility_patched = xyes ; then
+ KDE_CHECK_COMPILER_FLAG(fvisibility=hidden,
+ [
+ kde_have_gcc_visibility=yes
+ CXXFLAGS="$CXXFLAGS -fvisibility=hidden"
+ KDE_CHECK_VISIBILITY_GCC_BUG
+ HAVE_GCC_VISIBILITY=1
+ AC_DEFINE_UNQUOTED(__KDE_HAVE_GCC_VISIBILITY, "$HAVE_GCC_VISIBILITY", [define to 1 if -fvisibility is supported])
+ ])
+ fi
fi
])
this is the upstream bug: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21243 *** Bug 115299 has been marked as a duplicate of this bug. *** *** Bug 115422 has been marked as a duplicate of this bug. *** *** Bug 115430 has been marked as a duplicate of this bug. *** I have patch gcc for pr19664, but amarok still crash when startup, and the backtrack says: Using host libthread_db library "/lib/libthread_db.so.1". `system-supplied DSO at 0xffffe000' has disappeared; keeping its symbols. [Thread debugging using libthread_db enabled] [New Thread -1243863344 (LWP 27618)] [KCrash handler] #4 0x081baf40 in __gnu_cxx::__mt_alloc<PluginManager::StoreItem, __gnu_cxx::__common_pool_policy<__gnu_cxx::__pool, true> >::allocate () #5 0x081bb44a in std::vector<PluginManager::StoreItem, std::allocator<PluginManager::StoreItem> >::_M_insert_aux () #6 0x081b9cf1 in PluginManager::createFromService () #7 0x0811edda in EngineController::loadEngine () #8 0x0811fab5 in EngineController::loadEngine () #9 0x0808d055 in App::applySettings () #10 0x0808fad8 in App::App () #11 0x082246c4 in main () Is this because a vector<mytype> bug which the export patch may introduced? *** Bug 115648 has been marked as a duplicate of this bug. *** *** Bug 115655 has been marked as a duplicate of this bug. *** *** Bug 115797 has been marked as a duplicate of this bug. *** *** Bug 115825 has been marked as a duplicate of this bug. *** *** Bug 115881 has been marked as a duplicate of this bug. *** *** Bug 115961 has been marked as a duplicate of this bug. *** *** Bug 116041 has been marked as a duplicate of this bug. *** *** Bug 114582 has been marked as a duplicate of this bug. *** *** Bug 116221 has been marked as a duplicate of this bug. *** *** Bug 116283 has been marked as a duplicate of this bug. *** *** Bug 116352 has been marked as a duplicate of this bug. *** *** Bug 116373 has been marked as a duplicate of this bug. *** *** Bug 116307 has been marked as a duplicate of this bug. *** *** Bug 116515 has been marked as a duplicate of this bug. *** *** Bug 116348 has been marked as a duplicate of this bug. *** *** Bug 116459 has been marked as a duplicate of this bug. *** *** Bug 116943 has been marked as a duplicate of this bug. *** *** Bug 117058 has been marked as a duplicate of this bug. *** *** Bug 117058 has been marked as a duplicate of this bug. *** *** Bug 117235 has been marked as a duplicate of this bug. *** *** Bug 117318 has been marked as a duplicate of this bug. *** *** Bug 117350 has been marked as a duplicate of this bug. *** *** Bug 117417 has been marked as a duplicate of this bug. *** *** Bug 117606 has been marked as a duplicate of this bug. *** *** Bug 118137 has been marked as a duplicate of this bug. *** *** Bug 118607 has been marked as a duplicate of this bug. *** *** Bug 118613 has been marked as a duplicate of this bug. *** *** Bug 118685 has been marked as a duplicate of this bug. *** *** Bug 120905 has been marked as a duplicate of this bug. *** *** Bug 120541 has been marked as a duplicate of this bug. *** *** Bug 122563 has been marked as a duplicate of this bug. *** *** Bug 117972 has been marked as a duplicate of this bug. *** *** Bug 117972 has been marked as a duplicate of this bug. *** |
Version: (using KDE Devel) Installed from: Compiled sources Compiler: gcc 4.0.2 OS: Linux My sound card is soundblaster live 24bit,i'm using alsa 1.0.9a. here is the debug info: Using host libthread_db library "/lib/tls/libthread_db.so.1". [Thread debugging using libthread_db enabled] [New Thread -1216502080 (LWP 6522)] [KCrash handler] #3 0xb7a7233e in __gnu_cxx::__pool<true>::_M_reclaim_block () from /usr/lib/libstdc++.so.6 #4 0xb7fb21b7 in __gnu_cxx::__mt_alloc<std::string, __gnu_cxx::__common_pool_policy<__gnu_cxx::__pool, true> >::deallocate (this=0xbf9f6e28, __p=0x81067b0, __n=1) at mt_allocator.h:746 #5 0xb7fb21eb in std::_Vector_base<std::string, std::allocator<std::string> >::_M_deallocate (this=0xbf9f6e28, __p=0x81067b0, __n=1) at stl_vector.h:123 #6 0xb7fb2224 in ~_Vector_base (this=0xbf9f6e28) at stl_vector.h:109 #7 0xb7fb2290 in ~vector (this=0xbf9f6e28) at stl_vector.h:273 #8 0xb7fb22c8 in ~ObjectReference (this=0xbf9f6e1c) at core.h:117 #9 0xb7fa3883 in Arts::SoundServerStartup_base::_fromString ( objectref=@0xbf9f6ed0) at soundserver.cc:2545 #10 0x08067ff1 in SoundServerStartup (this=0xbf9f6eb0, r=@0xbf9f6eb8) at soundserver.h:1376 #11 0x080676e8 in Arts::SoundServerStartup_impl::cleanReference ( this=0x8163a08) at soundserverstartup_impl.cc:54 #12 0x08067819 in Arts::SoundServerStartup_impl::lock (this=0x8163a08) at soundserverstartup_impl.cc:78 #13 0x080637bd in Arts::SoundServerStartup::lock (this=0xbf9f7058) at soundserver.h:2082 #14 0x08062cf6 in main (argc=14, argv=0xbf9f7164) at artsd.cc:293