Bas Wijnen wrote: >>> (./ex070.tex >>> ! Missing number, treated as zero. >>> <to be read again> >>> n >>> l.4 \begin{pspicture}(-1,0)(1,nan) >>> >>> ? > > In other words, a TeX error, not a gcc error.
It looks to me like there is a numerical error in the gcc-4.3 compiled version of sketch, which passes an IEEE NAN to printf. I think it must be at least partly an amd64 issue. I have not rebuilt the binary package with gcc-4.3 (needs some chroot wizardry to get right), but cd sketch-0.2.59 make CC=gcc-4.3 sketch -Dex070 ./Doc/manexamples.sk works OK for me on i386 (i.e. no nan in the output). It would be helpful if someone could try a gcc4.3 i386 build to narrow down the possibilities. I will do that as my next step if no-one reports anything before. David PS. new upstream version acknowledged. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]