Hi Don (and others :)
There seems to be a problem with the rpm version of me, first to
let you know, the problem with my version which you was unable
to reproduce disappeared when I compiled it with the following
options: g77 -O2 -Wall pmxab.f -o pmxab
Also Andreas sent a file called EHLERTPC.pmx to someone which
crashes at bar 50 with my version. I then created a version
using f2c which did not crash, but gave me a strange output
though it worked fine:
Bar 48 Bar 49 Bar 50 Bar 51 Bar 52 Bar 53 Bar 54 R? , ? not "d","r",or"b";
rptfq2:D
Writing ./EHLERTPC.tex
Well the result shows an open slur which never seems to get
closed. I have had similar "mistakes" in my pieces thought they
never lead to a crash of pmx, here is the bar where my version
compiled with the above options crashes (seems to be a g77 related
problem):
%Bar 50
a4 o>-18 b8d o>-17 s-6 c1s d8 s-6 r8b r4b | //
\cchar{26}{\ppff{ff}}\ a2 o> s+6 a8 o_+9 s+7 r8-0 r4-0 | /
\zchar{19}{\ppff{ff}}\ c2s o> s5 d8 s5 r8 r4 | /
e2n o>-18 s1 d8 s1 r8b r4b | //
a2 s2 o>-3 d8- s2 r8-0 r4-0 | /
The "wrong" slur starts on this bar:
%Bar 14
r0 | //
e1 f1 g1 a1 b4 s+6 g4 s+6 r4 | /
g4 a4 b4u s6 r4 | /
e4 f4 e4 s-5 r4 | //
b8 e8 e8 d8s e4 s9+5 r4 | /
To me it seems to be a memory (access) problem, but maybe I'm wrong.
At least this could explain the strange output above as well. It's
perhabs simply luck that the above version does not crash.
On Wed, Jan 26, 2000 at 12:26:31PM +0100, Zdenek Pytela wrote:
> Andreas Deininger pise:
> > Hi all,
> >
> > I don't want to make any more confusion, I just want to tell my experiences
> > once more for those of you who couldn't read my German mail.
> >
> > >If there are any problems with the binaries produced by f2c+gcc, the
> > >easiest solution might be to put
> > >the generated C source into the main directory of my Unix/C
> > >package and compile it there using the patched libraries.
> > >(Or put the fortran sources in the pmx-orig/ subdirectory and run
> > >the convert.sh script there)
> >
> > Yes, there are problems with the rpm-pmx-binaries that were distributed at
> > Werner's site. Obviously the package was produced by Reinhard Katzmann using
> > g77 (http://www.gmd.de/Mail/mutex-archive-tail/1093.html)
> >
> > I encountered two problems with these binaries:
> > a)Have a look at the problem described at
> > http://www.gmd.de/Mail/mutex-archive-tail/1113.html
> > You can reproduce it with the Katzmann-rpm-binary, but you won't get that
> > strange behaviour with Stefan's "C" -version
> > b) I attach a pmx-file that crashes with "Katzmann-rpm-pmx" but works fine
> > with the "Stefan-C-pmx"
> >
> > I don't have the debian-rpm-pmx, but maybe one of you can check, whether is
> > works fine or not. Maybe one of the "experts" can even find out what's going
> > wrong when using the "Katzmann-rpm".
> I have pmx-2.10-1 rpm package and the enclosed EHLERTPC.PMX
> doesn't compile, it ends with SIGSEGV and core dump,
> Program received signal SIGSEGV, Segmentation fault.
> 0x80751bd in make2bar_ ()
> This is the g77 version of pmxab.
>
> --
>
> --Zdenek Pytela, <[EMAIL PROTECTED]>
--
E-Mail: [EMAIL PROTECTED]
Current Projects: Linux-Installation RH6.1 (90% finished)
Java-DB and -Interface for Pincity
Bach & H�ndel Trio Sonatas
