[sane-devel] Code-freeze for sane-1.0.9 active

Major A andras@users.sourceforge.net
Sun, 20 Oct 2002 14:13:07 +0100


> The above mentioned (real) times are from an Athlon XP1600+ (1.4 GHz
> real frequency). I'm not sure, which RAM is used. It's remote, so I
> can't just open it.
> 
> Here is the complete time output:
> real 1m52.472s
> user 1m38.540s
> sys  0m13.400s

OK, that makes sense. Your real time is shorter than mine, that's
probably because you have more memory (256MB is quite tight for a
machine running apache2, MySQL, samba, etc.)

> > ccc -o scanimage scanimage.o stiff.o  ../backend/.libs/libsane.a -ldl ../lib/liblib.a -lnsl -lm -ljpeg
> > /usr/lib/compaq/ccc-6.2.9.002-2/alpha-linux/bin/ld: scanimage: Not enough room for program headers (allocated 6, need 7)
> > /usr/lib/compaq/ccc-6.2.9.002-2/alpha-linux/bin/ld: final link failed: Bad value
> > make[1]: *** [scanimage] Error 1
> > 
> > Not good, not sure what we can do about it. This error occurs after
> > 47s though, so we have a candidate...
> 
> What does the linker mean by "program header"?

I have no idea, must be something related to the binary format. I've
seen this before, but I can't remember a solution at all.

Commercial software is unpredictable, SANE 1.0.8 failed on the same
SunOS machine on which 1.0.9-pre2 did just fine -- some weird i/o
error (!) in the middle of one of the backends. If I performed the
same steps as make does by hand, it worked (as far as I
remember). That was 100% reproducible.

  Andras

===========================================================================
Major Andras
    e-mail: andras@users.sourceforge.net
    www:    http://andras.webhop.org/
===========================================================================