Re: Test binutils packages made...please test quickly
Hi Christopher,
I'm not very familiar with objdump and nm, but I've hopefully put the
correct objdump output at
http://puffin.external.hp.com/~deller/objdump_qgvector
Maybe this can help you to see if it is a ld or g++ bug ?
Thanks
Helge
On Tuesday 16 October 2001 00:44, Christopher C. Chimelis wrote:
> Interesting. Have you looked at the generated object file to see if it's
> actually putting the functions in individual sections? If it is, then
> it's possible that it's the linker. There may be another option that
> could be used, though.
>
> I'm not too familiar with the hppa issues or special options,
> unfortunately.
>
> C
>
> On Mon, 15 Oct 2001, Helge Deller wrote:
> > I have absolutely no idea, if the following bug is related to g++ or to
> > ld from binutils. Maybe you can comment or have an idea ?
> >
> > Ok, I try to compile qt-copy from KDE CVS (local CVS copy of qt) on hppa
> > and dowloaded your package. BTW, this bug is already a few months old,
> > happens with older versions of binutils too and happens not only with qt,
> > but also with normal KDE programs.
> >
> > The compile-log is:
> > ......
> > g++ -c -pipe -fno-exceptions -O0 -ffunction-sections -DQT_MOC
> > -DQT_NO_CODECS -DQT_LITE_UNICODE -DQT_NO_COMPONENT -DQT_NO_STL
> > -I/opt/kde/qt/include -I../tools -I.
> > -I/home/cvs/qt-copy/mkspecs/linux-g++ -o qutfcodec.o
> > ../codecs/qutfcodec.cpp g++ -c -pipe -fno-exceptions -O0
> > -ffunction-sections -DQT_MOC -DQT_NO_CODECS -DQT_LITE_UNICODE
> > -DQT_NO_COMPONENT -DQT_NO_STL -I/opt/kde/qt/include -I../tools -I.
> > -I/home/cvs/qt-copy/mkspecs/linux-g++ -o qfile_unix.o
> > ../tools/qfile_unix.cpp g++ -c -pipe -fno-exceptions -O0
> > -ffunction-sections -DQT_MOC -DQT_NO_CODECS -DQT_LITE_UNICODE
> > -DQT_NO_COMPONENT -DQT_NO_STL -I/opt/kde/qt/include -I../tools -I.
> > -I/home/cvs/qt-copy/mkspecs/linux-g++ -o moc_yacc.o ./moc_yacc.cpp test
> > -d ../../bin/ || mkdir -p ../../bin/
> > g++ --whole-archive -fno-exceptions -o ../../bin/moc qbuffer.o
> > qptrcollection.o qcstring.o qdatastream.o qdatetime.o qfile.o qgarray.o
> > qgdict.o qglist.o qglobal.o qgvector.o qiodevice.o qregexp.o qstring.o
> > qstringlist.o qtextstream.o qbitarray.o qmap.o qgcache.o qtextcodec.o
> > qutfcodec.o qfile_unix.o moc_yacc.o /usr/bin/ld:
> > qgvector.o(.gnu.linkonce.t._ZNK8QGVector2atEj+0x2c): cannot reach
> > 00000395__ZN8QGVector17warningIndexRangeEj+0, recompile with
> > -ffunction-sections /usr/bin/ld:
> > qgvector.o(.gnu.linkonce.t._ZNK8QGVector2atEj+0x2c): cannot handle
> > R_PARISC_PCREL17F for QGVector::warningIndexRange(unsigned)/usr/bin/ld:
> > final link failed: Bad value collect2: ld returned 1 exit status
> > make[1]: *** [../../bin/moc] Error 1
> > make[1]: Leaving directory `/home/cvs/qt-copy/src/moc'
> > make: *** [src-moc] Error 2
> >
> > As you see I compiled with -ffunction-sections, but nevertheless I get
> > the above errors. Is this a bug in ld or in g++ ?
Reply to: