Bug 1192464
| Summary: | Qt: FTBFS with gcc5 | ||
|---|---|---|---|
| Product: | [Fedora] Fedora | Reporter: | Rex Dieter <rdieter> |
| Component: | gcc | Assignee: | Jakub Jelinek <jakub> |
| Status: | CLOSED RAWHIDE | QA Contact: | Fedora Extras Quality Assurance <extras-qa> |
| Severity: | unspecified | Docs Contact: | |
| Priority: | unspecified | ||
| Version: | 22 | CC: | davejohansen, dvratil, itamar, jakub, jreznik, kevin, law, ltinkl, rdieter, rnovacek, robatino, satellitgo, smparrish, than |
| Target Milestone: | --- | ||
| Target Release: | --- | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | gcc-5.0.0-0.13.fc22 | Doc Type: | Bug Fix |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2015-02-15 15:49:25 UTC | Type: | Bug |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
| Bug Depends On: | |||
| Bug Blocks: | 1043121, 1135100 | ||
|
Description
Rex Dieter
2015-02-13 12:55:37 UTC
I'll try this recent fix for Qt5: https://qt.gitorious.org/qt/qtbase/commit/9fb4c2c412621b63c06dbbd899f44041b2e126c2 %changelog * Fri Feb 13 2015 Rex Dieter <rdieter> - 1:4.8.6-23 - Qt: FTBFS with gcc5 (#1192464) - Make Adwaita the default theme for applications running in the GNOME DE (#1192453) Darn, that wasn't enough, was worth a try. Proposed as a Blocker for 22-alpha by Fedora user rdieter using the blocker tracking app because: Since gcc5 landed, Qt(4) builds fail. We still ship quite a few Qt4 (and KDE4) based applications on the kde spin. so my initial look,
qdrawhelper_p.h references qt_memfill32 alright, and
src/gui/painting/qdrawhelper.cpp references a bunch of optimized versions, including:
src/gui/painting/qdrawhelper_x86_p.h
painting.pri contains too:
if(mmx|3dnow|sse|sse2|iwmmxt) {
HEADERS += painting/qdrawhelper_x86_p.h \
painting/qdrawhelper_mmx_p.h \
painting/qdrawhelper_sse_p.h \
painting/qdrawingprimitive_sse2_p.h
MMX_SOURCES += painting/qdrawhelper_mmx.cpp
MMX3DNOW_SOURCES += painting/qdrawhelper_mmx3dnow.cpp
SSE3DNOW_SOURCES += painting/qdrawhelper_sse3dnow.cpp
SSE_SOURCES += painting/qdrawhelper_sse.cpp
SSE2_SOURCES += painting/qdrawhelper_sse2.cpp
SSSE3_SOURCES += painting/qdrawhelper_ssse3.cpp
IWMMXT_SOURCES += painting/qdrawhelper_iwmmxt.cpp
}
so it's possible that sse2 (and/or friends) are no longer being detected properly.
build.log contains: Configuration .......... system-sqlite release largefile stl mmx 3dnow sse sse2 sse3 ssse3 sse4_1 sse4_2 avx ... ... MMX/3DNOW/SSE/SSE2/SSE3. yes/yes/yes/yes/yes SSSE3/SSE4.1/SSE4.2..... yes/yes/yes AVX..................... yes So that part looks ok. Qt Devs investigated and determined to be a gcc bug, re-assigning. Per QTBUJG-44466 comment: > It's a GCC bug. One file defines a symbol "qt_memfill32" and the other > searches for "_Z12qt_memfill32". > > GCC 4.9, Clang and ICC don't have this problem. Therefore, it's a GCC bug. Reported: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65062 Testcase: template <class> void tf() { extern void (*qt_memfill32)(); qt_memfill32(); } void f() { tf<int>(); } Restting default assignee(s), see also https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65062 already dup'd to upstream bug, https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64898 (and fixed upstream) I guess we're just waiting for a fix gcc build downstream |