Sage 2.10 is out. It wasn't planned this way since we had originally shot for some release early next week, ideally Monday. So what happened? 2.10.alpha4 build perfectly on all of our test platforms and the results from the patches created during
Doc Day 1 would potentially destabilize the build. All was set in motion to release at this point - after all we had gone nearly twelve days since the last release, which is longer than the usual time between releases - but the release was held up by some tricky, long standing but extremely hard to reproduce bug in the Maxima interface of Sage triggered by some last minute merge of an
essential fix. William finally tracked it down and fixed it, which delayed the release about 24 hours. I used that time to catch up on sleep and to prepare myself for
Bug Day 9.
The main changes for Sage 2.10 are:
- Python is now built with ucs4
- FLINT was updated to the 1.0.5 release
- Many bug fixes and also a couple of significant memory leak fixes
- Integrate a fix to the MPFR library so we no longer smash the stack with high precisions. This cause a performance regression when multiplying
- Fix a long standing, hard to hit bug in the Maxima interface
- When launching external executables LD_LIBRARY_PATH is restored for external binaries, i.e. those which do not reside in SAGE_LOCAL/bin. Before the patch many external executables failed due to library conflicts. This was reported many times on Debian based systems
What does the future hold?
Sage 2.10.1 is already getting merges on the way to alpha0. So stay tuned :)
Cheers,
Michael
No comments:
Post a Comment