Differences between revisions 1 and 39 (spanning 38 versions)
Revision 1 as of 2007-11-28 11:31:24
Size: 1264
Comment: initial version
Revision 39 as of 2009-05-14 11:44:57
Size: 2217
Comment: Add Sage 4.0 page for Solaris
Deletions are marked like this. Additions are marked like this.
Line 3: Line 3:
This page tracks the Solaris port of Sage. This page tracks the Solaris port of Sage currently being done by Michael Abshoff.
Line 5: Line 5:
mabshoff is currently working on a 32 bit build on Solaris 9/Sparc. It should run on Solaris 10/Sparc. == Goal ==
Line 7: Line 7:
== History of efforts == The goal of this port is to get Sage compiling and passing all doctests on x86/x86-64 as well as 32 and 64 bit Sparc on Solaris 10 and higher. The only compiler supported for this port will be gcc 4.3.2, but as the gcc port matures we will look into full support for the Sun Forte compiler toolchain.
Line 9: Line 9:
 * 2.8.14:
  * new bug in matrix code that provides a special class for multivariate polynomials as entries causes segfaults all over the map.
  * int_fast64_t needs to be 4 bytes large, otherwise Sage throws an exception at startup. This needs fixing and actually causes doctest failures with the modp code
  * libfplll miscompiles, fix clear, but needs to be merged upstream
  * FLINT miscompiles, fixed upstream
  * lcalc miscompiles, fixes need to be integrated
  * clisp compiles, but segfaults. Solution: use 2.39 binary until issue is solved
  * cvxopt doesn't work (complex.h missing)
  * 2-3 patches are needed to sage.spkg - easy fixes, need to be merged upstream
 * 2.8.3:
  * worked reasonably well after workarounds, about 10-12 doctest failures
  * work arounds needed: clisp, lcalc
  * missing/failed spkgs: cvxopt
 * pre 2.8:
  * various people (David Kirby) worked on this, but I am not sure if Sage ever passed 100% of doctests
== Status ==
 
 * As of Sage 3.3.alpha4 the compile works on Solaris x86 with the custom toolchain. When using a binary clisp-2.47.spkg on Solaris/Sparc and the custom toolchain the compile also finishes. Note that several patches have to be applied for most of the doctests to pass. See the latest status page for more info below.
 * 64 bit support for Solaris 10 on x86-64 as well as Sparc is planned for Sage 3.4, but will take a couple releases to be fully merged.

Please see also the open issues section at the bottom of the page for known problems.

 * [[solaris/sage-4.0|Sage 4.0 build issues and doctest failures]] (current)
 * [[solaris/sage-3.4.2|Sage 3.4.2 build issues and doctest failures]] (outdated)
 * [[solaris/sage-3.4.1|Sage 3.4.1 build issues and doctest failures]] (outdated)
 * [[solaris/sage-3.3|Sage 3.3 build issues and doctest failures]] (outdated)
 * [[solaris/sage-3.2.3|Sage 3.2.3 build issues and doctest failures]] (outdated)
 * [[solaris/sage-3.1.3|Sage 3.1.3 build issues and doctest failures]] (outdated)
 * [[solaris/sage-3.0.5|Sage 3.0.5 build issues and doctest failures]] (outdated)
 * [[solaris/sage-3.0.4-and-earlier|Sage 3.0.4 and earlier build issues and doctest failures]] (outdated)

== Open Issues ==
 * clisp fails to compile on Sparc/Solaris up to clisp 2.47 with gcc 3.3.6 to 4.3.2. The switch to ecl will fix this.
 * Symmetrica detects internal inconsistency issues on Solaris. This is likely a bug in the interface code that just happens to cause trouble on Solaris only.
 * Sympow needs to have the FPU mode initialized properly on Solaris/x86. High optimization level for the compiler also break Sympow on Solaris/x86
Line 27: Line 32:
 * make doctests pass 100%
 * open track tickets for all known issue
 * make doctests pass 100% :)
 * make building toolchain automatic

The Solaris Port of Sage

This page tracks the Solaris port of Sage currently being done by Michael Abshoff.

Goal

The goal of this port is to get Sage compiling and passing all doctests on x86/x86-64 as well as 32 and 64 bit Sparc on Solaris 10 and higher. The only compiler supported for this port will be gcc 4.3.2, but as the gcc port matures we will look into full support for the Sun Forte compiler toolchain.

Status

  • As of Sage 3.3.alpha4 the compile works on Solaris x86 with the custom toolchain. When using a binary clisp-2.47.spkg on Solaris/Sparc and the custom toolchain the compile also finishes. Note that several patches have to be applied for most of the doctests to pass. See the latest status page for more info below.
  • 64 bit support for Solaris 10 on x86-64 as well as Sparc is planned for Sage 3.4, but will take a couple releases to be fully merged.

Please see also the open issues section at the bottom of the page for known problems.

Open Issues

  • clisp fails to compile on Sparc/Solaris up to clisp 2.47 with gcc 3.3.6 to 4.3.2. The switch to ecl will fix this.
  • Symmetrica detects internal inconsistency issues on Solaris. This is likely a bug in the interface code that just happens to cause trouble on Solaris only.
  • Sympow needs to have the FPU mode initialized properly on Solaris/x86. High optimization level for the compiler also break Sympow on Solaris/x86

ToDo

  • make doctests pass 100% :)

  • make building toolchain automatic