Differences between revisions 34 and 40 (spanning 6 versions)
Revision 34 as of 2011-09-29 03:57:02
Size: 6158
Editor: kini
Comment:
Revision 40 as of 2011-09-30 08:22:53
Size: 7070
Editor: SimonKing
Comment:
Deletions are marked like this. Additions are marked like this.
Line 48: Line 48:
 * Thursday: 390 failing tests
Line 91: Line 92:
 * Rebase the patch at [[http://trac.sagemath.org/sage_trac/ticket/4539|#4539 plural wrapper]] to [[http://trac.sagemath.org/sage_trac/ticket/11068|11068 implementation of one and two sided ideals of noncommutative rings]]
 * Preliminary patch that fixed the segfaults: [[http://trac.sagemath.org/sage_trac/attachment/ticket/4539/trac4539_kwds.patch|trac4539_lmul.patch]]. This is not performant and actually works around an rmul/lmut issue. So we need a better fix for that.
 * Fixing missing keyword argument issue by reverting an unnecessary path of the patch: [[http://trac.sagemath.org/sage_trac/attachment/ticket/4539/trac4539_kwds.patch|trac4539_kwds.patch]]
 * [[http://trac.sagemath.org/sage_trac/ticket/4539|#4539 plural wrapper]] is rebased to [[http://trac.sagemath.org/sage_trac/ticket/11068|11068 implementation of one and two sided ideals of noncommutative rings]], so that one has left ideals and two-sided ideals of g-algebras.
 * The left or two-sided Gröbner bases can be used to compute normal forms (that has been missing before).
 * Pickling works (has been missing), the category test suites pass.
 * Since g-algebras are now unique parents (unless uniqueness is destroyed on purpose), `id(self)` is used as hash.
 * Doc tests pass (says Simon), but the last patch does not apply cleanly (says Alexander, but Simon can't confirm it). Apart from that, it is ready for review.
 * ''To do'': Currently, quotients of g-algebras are implemented by the general framework of [[http://trac.sagemath.org/sage_trac/ticket/11068|#11068]]. It should instead wrap what Singular offers, but that shall be on a different ticket.
Line 100: Line 104:
 * It is fixed, combining the old test used in Sage and the test used in Singular.
Line 122: Line 127:
 * add `.saturation()`, `.blow_up_ideal()` methods in the multivariate polynomial ideal class  * [[http://trac.sagemath.org/sage_trac/ticket/11864]] add `.saturation()`, `.blow_up_ideal()` methods in the multivariate polynomial ideal class
Line 146: Line 151:

== Review PolyBoRi 0.8.0 patch ==

'''People''': Martin, Alexander (offsite)

 * [[http://trac.sagemath.org/sage_trac/ticket/11575|#11575]]

Martin reviewed it mathematically. The patch was rebased to 4.7.2alpha3, the spkg got the improved sse2-flags detection from 0.8.1.
(Final) tests are running currently.

== Review linbox for modn dense matrices patch ==

 * [[http://trac.sagemath.org/sage_trac/ticket/4260|#4260]]

Sage/Singular Days Projects

Update Singular spkg

People: Simon, Burcin, Martin

working on prerelease of Sage-4.7.2.alpha3 dated Sept 19, 2012 md5sum: ad021855ae07ecdd1cdcc7b973c9cdba

  • #10903 update Singular

  • #10810 Singular doesn't build on ARM

  • #11769 Singular doesn't build on Ubuntu Oneiric

  • #10902 factorization of multivariate polynomials

Update the Singular interface to work with libpolys

People: Volker, Burcin, Oleksandr

  • There were many changes to libSingular as part of the refactoring effort. The Sage interface needs to be adapted to these. This should be a matter of renaming the functions that are called in Sage and adding the current ring as an argument (modulo bugs in libpolys).
  • create spkg for refactored Singular
  • search/replace calls to libSingular in the Sage library to get the interface to compile & link

Fix the building of spielwiese Singular (libpolys)

People: Volker, Oleksandr

  • parallel building make -j used to fail because of generated sources (e.g. feOpt.inc)

  • make install should install everything needed for make distcheck

  • tests run via make check should be able to find resources (e.g. gftables) - cleate symlink or set Singular environment variables?

Refactored Singular should pass it's test suite

People: Burcin, Oleksandr, Hans, Volker

Bug hunt through Singular kernel code. An adventure no hacker can resist.

Compiler warnings

People:

There are lots of compiler warnings...

Plural/SCA

People: Oleksandr

  • Bug hunt at PLURAL/SCA kernel code: (left) std etc...

Try out some dynamic modules

People: Oleksandr

  • Try building & using dynamic modules

Blackbox objects

People: Oleksandr

  • Try building & using blackbox objects, e.g. coefficient domain as a wrapper for coeffs

Interface to Singular's modules

People: Hiep, Burcin, Simon

  • This shouldn't be so hard since libSingular uses the same type for polynomials and module elements

Review Simon's letterplace patches

People: Simon, Burcin, Alexander (offsite)

  • #7797

  • #11115 make cached methods faster - dependency of #7797 (Volker)

Plural interface

People: Burcin, Simon, Alexander (offsite)

  • #4539 plural wrapper is rebased to 11068 implementation of one and two sided ideals of noncommutative rings, so that one has left ideals and two-sided ideals of g-algebras.

  • The left or two-sided Gröbner bases can be used to compute normal forms (that has been missing before).
  • Pickling works (has been missing), the category test suites pass.
  • Since g-algebras are now unique parents (unless uniqueness is destroyed on purpose), id(self) is used as hash.

  • Doc tests pass (says Simon), but the last patch does not apply cleanly (says Alexander, but Simon can't confirm it). Apart from that, it is ready for review.
  • To do: Currently, quotients of g-algebras are implemented by the general framework of #11068. It should instead wrap what Singular offers, but that shall be on a different ticket.

Exponent overflow

People: Simon

  • Working on the Plural interface, we found that Sage does not raise an overflow error when it should. Example: With P.<x,y>=QQ[], y^2^30 returns a correct result; in P.<x,y,z>=QQ[], an overflow occurs, but instead of raising an error, y^2^30 returns zero.

  • It is fixed, combining the old test used in Sage and the test used in Singular.

Use libpolys for function field arithmetic

People: Burcin

FLINT integration in Singular

People: Fredrik, MartinL, Christian

  • Add missing utility and conversion functions to FLINT
  • Use FLINT in Singular
  • Benchmark against NTL version

Beginner tickets

People: Hiep, Burcin to practice preparing and submitting patches, for people new to the Sage development process:

Fast containers for objects and morphisms of a category

People: Simon

  • See #10667 Morphisms and Objects of Categories

Improve/Review citation library

People: Niels, Burcin, (and somebody else?)

#3317 bitbucket

  • #11851 #11852 split off pybtex package to a separate ticket (submit as an optional package first)

  • Add more examples of citations
  • add sage's citation

  • Open a ticket for matplotlib not to include an extra version of pyparsing.

Fix reference counting of Singular rings

People: Volker, Burcin, Martin

Review PolyBoRi 0.8.0 patch

People: Martin, Alexander (offsite)

Martin reviewed it mathematically. The patch was rebased to 4.7.2alpha3, the spkg got the improved sse2-flags detection from 0.8.1. (Final) tests are running currently.

Review linbox for modn dense matrices patch

days34/projects (last edited 2011-09-30 08:22:53 by SimonKing)