Update: I re-did it calling it 3.0.rc2+repack1 since the ~ ended up breaking piuparts on CI. It's just a minor detail in the scheme of things, I assume. I have pushed to the team repo.
It should be "safe" to release now. There is new support for OpenMP and LTO, but without direct access to experimental and buildd's of the various architectures, I wouldn't want to upload a package to unstable at this point with those enabled and risk breaking it. It works fine, if a smidge slower than possible, without openmp and LTO enabled, so I think probably leave it well enough alone. I've added some autopkgtests using the CLI interface as completely as possible without modifying the source tree, which should help too. Probably the most fragile part of the package is the use of three-js (libjs-three) for the HTML trimesh viewer export: if that gets updated and breaks something being used, the export will appear to work fine but will hit a javascript error on usage. I'd be interested to learn how to write up the test to catch this, if anybody has ideas. Thanks! Ryan On 1/19/2021 5:57 PM, Ryan Pavlik wrote: > Anton, > > Thanks for taking the time to update that package! Upstream has released > another RC (which should be identical to 3.0 with the exception of > translation updates). Additionally, we've found a patch from upstream > mimalloc that should fix the armel build (I hope), and which I have > included in the debian package. I have updated to that upstream release, > and updated the package substantially (removed some bundled fonts and > js, etc) and pushed to my salsa repo: I wasn't sure if you wanted me to > push to the main team repo, especially since the existing version is > "3.0.rc1+repack1" which Lintian warns me is not less than 3.0 - I did my > import as "3.0~rc2+repack1" accordingly. > > https://salsa.debian.org/rpavlik/solvespace > > When you have a chance, I'd appreciate a review and sponsoring. I've > done basic testing of the package myself, especially the parts that are > affected by patches/repack, and it seems to work well to me. > > Thanks! > > Ryan > > On 1/16/2021 12:09 PM, Anton Gladky wrote: >> Hi Ryan, >> >> the newer version of solvespace is in sid already. There is a problem >> with armel compilation >> >> selected processor does not support `yield' in ARM mode >> >> So I requested the removal of the package on this platform. >> If somebody has an advice how to fix this special assembler-problem, >> please let me know. >> >> Best regards >> >> Anton >> >> >> Anton >> >> >> Am Do., 14. Jan. 2021 um 00:33 Uhr schrieb Ryan Pavlik >> <ryan.pav...@gmail.com>: >>> On 1/9/2021 2:27 PM, Anton Gladky wrote: >>>> Dear members of Debian Science Team, >>>> >>>> we are approaching to a new stable release freeze. If youhave some >>>> packages to be sponsored, please let us know. >>>> >>>> I have a very limited time, but I will try to review/sponsor someof >>>> packages. >>>> >>>> >>>> Best regards >>>> >>>> Anton >>> I see you were the last to touch the SolveSpace package. Do you expect >>> to have time to look at that soon to update to the new RC? (They/we have >>> very high stability standards, 3.0 could have been released 2 years ago) >>> If not, I'll go ahead and update to the latest tagged upstream this week >>> and request sponsorship. >>> >>> >>> Thanks, >>> >>> >>> Ryan >>> >>>
signature.asc
Description: OpenPGP digital signature
-- debian-science-maintainers mailing list debian-science-maintainers@alioth-lists.debian.net https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-science-maintainers
