https://bugzilla.redhat.com/show_bug.cgi?id=840244 --- Comment #9 from pcpa <paulo.cesar.pereira.de.andrade@xxxxxxxxx> --- (In reply to comment #8) > According to http://sagemath.org/packages/experimental/, Sage is using surf > version 1.1! Where did that version come from? Do you know? Looking at the spkg contents: [[[[ $ cat surf-1.1/get_from_cvs echo "Enter a blank password" cvs -d:pserver:anonymous@xxxxxxxxxxxxxxxxxxx:/cvsroot/surf login cvs -z3 -d:pserver:anonymous@xxxxxxxxxxxxxxxxxxx:/cvsroot/surf co -P surf ]]]] But it looks seriously outdated: [[[[ $ cat surf-1.1/SAGE.txt This is the 2006-02-12 CVS version of surf 1.1.0, obtained using cvs -d:pserver:anonymous@xxxxxxxxxxxxxxxxxxx:/cvsroot/surf login cvs -z3 -d:pserver:anonymous@xxxxxxxxxxxxxxxxxxx:/cvsroot/surf co -P surf I'm using the cvs version since I couldn't get the download version from 2003 to compile (it's just too out of date). I then ran autogen.sh to create the configure script. ---------------------------------------------- I had to change the source code in two files to get surf to compile with GCC 4.0.2 on my system: * Deleted "IO::" in three places in surf/misc/IO.cc lines 207, 218, and 228 * Also in misc/IO.cc, changed line 121 to #ifdef XHAVE_LIBREADLINE so that chunk of code that uses readline isn't used. (It seems to be out of date. Since we're only using surf from singular, not having readline somewhere shouldn't be a big problem.) PACKAGE MAINTAINER: William Stein ]]]] I actually did never check on actual upstream sagemath binaries. Maybe the singular examples using surf are not working in recent sagemath; I know it works in my Mandriva and current work in progress Fedora sagemath rpm package, using the package proposed here. Yes, BTW the sagemath version looks quite bogus... > The conflict with the existing surf package can cause problems. Can we > rename the binary in this package instead so they don't conflict? Again, I > realize that this makes us nonstandard and that we can't talk to upstream > about it because upstream mostly isn't there anymore, but that seems to be a > better path to me. Renaming the binary should be trivial, I was not much happy with needing to patch Singular scripts to match the new name, but should also not be a big deal. But then, the "surf" package, for the single webkit window appears to be quite outdated. > I noticed that the configure script reports that it cannot find tiffio.h, > even though that file is in libtiff-devel. Do you know what's going on > there? Looks like something bogus, and should be corrected, (or just ignore tiff support :-) [[[[ dnl check for tiff library and header file (FreeBSD 3.0 has tiff34.so dnl instead of tiff.so and the header files are in /usr/local/lib/tiff34): AC_CHECK_LIB(tiff, main,, AC_CHECK_LIB(tiff34, main,, AC_MSG_ERROR([Sorry: can't find libtiff]))) AC_CHECK_HEADER(tiffio.h,, [AC_CHECK_HEADER(tiff34/tiffio.h,AC_DEFINE(TIFF_HEADER_34))], [AC_MSG_ERROR(["Sorry: cannot find header file tiffio.h"])] ) ]]]] Thanks for spotting it. I will check if autoreconf does work, otherwise, could patch the configure script instead. Actually, the "surfex" java interface in Singular-surfex should be good for most if not all uses, just that the surf binary and the associated sagemath tutorial examples expect to use Singular's surf.lib that uses the surf binary. As noted in #c2 but link here for easier access, this is how it looks like in the sagemath tutorial http://fedorapeople.org/~pcpa/rawhide-sage-notebook4.png -- You are receiving this mail because: You are on the CC list for the bug. _______________________________________________ package-review mailing list package-review@xxxxxxxxxxxxxxxxxxxxxxx https://admin.fedoraproject.org/mailman/listinfo/package-review
