wrote: > Hi all! > > First of all, this bug is ages old, it's been around at least from 4.3. > onwards. > > Second, it is amd64-only; I can reproduce it with 100% certainty on all > amd64-machines, but nowhere else (including, btw, one 64-bit alpha). > > Third, dx works fine nevertheless: executive works, program runs etc, it's > simply not possible to edit the program due to the layout bug.
I have the exact same symptoms here. Interestingly I ran dxui through valgrind and got a few warnings about using uninitalised memory: ==7745== Source and destination overlap in memcpy(0xBCB2720, 0xBCB2740, 96) ==7745== at 0x4C2323A: memcpy (mc_replace_strmem.c:402) ==7745== by 0x5A0330E: _XmBuildResources (in /usr/lib/libXm.so.2.0.1) ==7745== by 0x59EF1FC: (within /usr/lib/libXm.so.2.0.1) ==7745== by 0x7E51350: (within /usr/lib/libXt.so.6.0.0) ==7745== by 0x7E51350: (within /usr/lib/libXt.so.6.0.0) ==7745== by 0x7E51835: XtInitializeWidgetClass (in /usr/lib/libXt.so.6.0.0) ==7745== by 0x7E527D8: _XtCreateWidget (in /usr/lib/libXt.so.6.0.0) ==7745== by 0x7E809A8: (within /usr/lib/libXt.so.6.0.0) ==7745== by 0x7E80BD3: XtVaCreateManagedWidget (in /usr/lib/libXt.so.6.0.0) ==7745== by 0x47B234: EditorWindow::createWorkArea(_WidgetRec*) (EditorWindow.C:1290) ==7745== by 0x5484AF: MainWindow::initialize() (MainWindow.C:282) ==7745== by 0x544148: IBMMainWindow::initialize() (IBMMainWindow.C:194) ==7745== ==7745== Conditional jump or move depends on uninitialised value(s) ==7745== at 0x59DEDA2: _XmCalcLabelDimensions (in /usr/lib/libXm.so.2.0.1) ==7745== by 0x59DF730: (within /usr/lib/libXm.so.2.0.1) ==7745== by 0x7E51428: (within /usr/lib/libXt.so.6.0.0) ==7745== by 0x7E51E64: (within /usr/lib/libXt.so.6.0.0) ==7745== by 0x7E5279B: _XtCreateWidget (in /usr/lib/libXt.so.6.0.0) ==7745== by 0x7E809A8: (within /usr/lib/libXt.so.6.0.0) ==7745== by 0x7E80BD3: XtVaCreateManagedWidget (in /usr/lib/libXt.so.6.0.0) ==7745== by 0x4D4A9B: PageSelector::buildSelector() (PageSelector.C:261) ==7745== by 0x4D5D2A: PageSelector::PageSelector(EditorWindow*, _WidgetRec*, Network*) (PageSelector.C:174) ==7745== by 0x47B3BF: EditorWindow::createWorkArea(_WidgetRec*) (EditorWindow.C:1323) ==7745== by 0x5484AF: MainWindow::initialize() (MainWindow.C:282) ==7745== by 0x544148: IBMMainWindow::initialize() (IBMMainWindow.C:194) ==7745== ==7745== Conditional jump or move depends on uninitialised value(s) ==7745== at 0x59DE07D: _XmLabelGetPixmapSize (in /usr/lib/libXm.so.2.0.1) ==7745== by 0x59DED54: _XmCalcLabelDimensions (in /usr/lib/libXm.so.2.0.1) ==7745== by 0x59DF730: (within /usr/lib/libXm.so.2.0.1) ==7745== by 0x7E51428: (within /usr/lib/libXt.so.6.0.0) ==7745== by 0x7E51E64: (within /usr/lib/libXt.so.6.0.0) ==7745== by 0x7E5279B: _XtCreateWidget (in /usr/lib/libXt.so.6.0.0) ==7745== by 0x7E809A8: (within /usr/lib/libXt.so.6.0.0) ==7745== by 0x7E80BD3: XtVaCreateManagedWidget (in /usr/lib/libXt.so.6.0.0) ==7745== by 0x4D4A9B: PageSelector::buildSelector() (PageSelector.C:261) ==7745== by 0x4D5D2A: PageSelector::PageSelector(EditorWindow*, _WidgetRec*, Network*) (PageSelector.C:174) ==7745== by 0x47B3BF: EditorWindow::createWorkArea(_WidgetRec*) (EditorWindow.C:1323) ==7745== by 0x5484AF: MainWindow::initialize() (MainWindow.C:282) ^C==7745== ==7745== ERROR SUMMARY: 4 errors from 3 contexts (suppressed: 12 from 2) ==7745== malloc/free: in use at exit: 1,127,265 bytes in 17,721 blocks. ==7745== malloc/free: 28,373 allocs, 10,652 frees, 5,401,815 bytes allocated. ==7745== For counts of detected errors, rerun with: -v ==7745== searching for pointers to 17,721 not-freed blocks. ==7745== checked 7,251,808 bytes. The ones in _XmLabelGetPixmapSize sound like they could be to blame maybe, but I can't quite figure out the source of it... > Also, the layout bug can be fixed by recompiling against libmotif-dev instead > of lesstif2-dev, so I guess blame can be squarely placed on lesstif2 or its > use by dx. That's handy to see a work-around at least! Alan -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
