Release notes Sage 2.8.2
The goal of this release was to squash as many known bugs from our
bug tracker as possible. Only a very limited number of features
have been added and to some extend the portability of Sage to
Solaris has been improved. While we haven't crossed the finish
line for Solaris yet we have gotten very close. Unfortunately
some of the fixes caused problems on other architectues, so we
will revisit those problems in the future, hopefully for 2.9
The total number of tickets closed was 47. Considering that now
there are about 217 tickets open (of which many are enhancements
and not defects) Sage's Bug Day 1 can be considered a tremendous
success. Participation was lively (in total 10 people showed up)
and the whole session went on for about 16 hours. Some of the
people even got together in IRC the next day to sort out some of
the leftovers.
We still think that certain aspects of the way we planned and
handled the bug day can be improved upon (as always) and plan to
do Sage's Bug Day 2 on September 7th 2007 starting at 10 am
Pacific time.
The results from Sage's Bug Day 1 are as follows:
algebraic geometry 4/4
basic arithmetic 12/12
interfaces 5/6
linear algebra 3/3
modular forms 2/2
notebook 5/7
number theory 2/2
packages 7/11
user interface 7/7
x/y denotes that x out of y bugs in that category were fixed.
For the details check out
http://www.sagemath.org:9002/sage_trac/milestone/sage-2.8.2
and
http://www.sagemath.org:9001/bug1/Results
2.8.1:
-----
These are the release notes for sage-2.8.1, though there never was an
official release. The reason it was never officially released was that
2.8.1 was the basis of the first Sage Bug Squashing day that is about
to be wrapped up.
The main goal of this release was to get Sage to compile cleanly on a
wider range of platforms, and fix any build-system bugs we discovered
along the way:
* Linux on Itanium
* Linux on PPC 64
* Solaris on Sparc and Opteron
This involved a whole lot of massaging spkg-install scripts and build
options. William Stein did the port for Linux on Itanium and PPC 64
while William Stein, Michael Abshoff, Didier Deshommes and Kate Minola
did work on the Solaris port. They were helped by Martin Albrecht to
sort out some Singular and libSingular build issues. Robert Bradshaw
coded up a patch to sort out some issues with Cython on *BSD based
systems.
While the Itanium port of Sage is now fully working and supported
neither the Linux on PPC not the Solaris on Sparc/Opteron port made it
to the finish line. While the Linux on PPC is close there are still
some issues left with Solaris, but great progress has been made there.
Detailed changes:
* w stein, m abshoff: add stdint.h to local/include/ to work around
missing defines in Solaris
* r bradshaw: fix coercion code to compile on *BSD, fix Cython to
warn about _[A-Z] PyObjects. [not applied yet.]
* w stein: Make sure that the system-wide gmp doesn't get picked up.
* w stein, d deshommes, m abshoff: linbox build fixes, lround
workaround, build against gsl blas on solaris
* w stein, k minolta: gfortran binary on Solaris, add SAGE_FORTRAN
environment variable.
* w stein: make f2c use gcc instead of cc on Solaris
* w stein: pari build fixes on Solaris and Itanium
* w stein: fix problem with matplotlib when linking in the wrong libz
in /usr/local/lib
* w stein: new mpfi package
* w stein, m abshoff, m. albrecht: many fixes to the singular package
to make it build on Solaris
* w stein, m abshoff, d deshomees: stdint.h fix for Solaris
* w stein: update gmp package, so that the resulting gmp is much more
portable (and possible resolve some other bugs -- old version was
building with certain patches preapplied).
Many of the fixes were a joint effort be several people, some fixes
were discovered twice, i.e. Didier Deshommes had made suggestions on
fixes for Sage 2.5.x on Nexenta which were unfortunately
rediscovered. We will discuss how to deal with that problem in the
report about the first Sage Bug Squashing day.
Cheers,
Michael Abshoff