Michael, to make things short: The discussion is around these changes to the `pyproject.toml` file https://github.com/sagemath/sage/pull/38227/files#diff-0acbedc0b174ebec342997fdca9442a5486917ab04644da866213e97d2543604 which he claims harmonize the constraints with scipy (although scipy actually declares a different numpy constraint). I'd argue that these are incorrect as build-time requirements of scipy should not influence what build-time requirements we declare for sage-the-lib. Matthias claims these are needed for the system-site-package feature.
The claim by Matthias that such version constraint updates are standard procedure seems to be invalid as well. Currently, the constraints for cython and numpy are not synced with the built-time requirements of scipy 1.12 and the last upgrade PR of scipy (only a couple of months ago) also didn't touch any of these versions https://github.com/sagemath/sage/pull/37266. On Thursday, September 19, 2024 at 9:41:13 PM UTC+8 dim...@gmail.com wrote: > > > On 19 September 2024 03:20:52 BST, Michael Orlitzky <mic...@orlitzky.com> > wrote: > >On Wed, 2024-09-18 at 16:23 -0700, Matthias Koeppe wrote: > >> > >> I would appreciate feedback from the developers working on this feature > >> (Dima?, Michael?) regarding how to specify these constraints now that > >> sage-lib defines its own constraints in pyproject.toml. > >> > >> The change is purely syntactic -- that's all that your PR (and the > >> subsequent fixups of it) did. There has been no semantic change > whatsoever. > >> Previously all version constraints were set centrally in the > >> build/pkgs/*/install-requires.txt files (recently renamed to > >> version_requirements.txt), now some are set in src/pyproject.toml. > That's > >> all. > > > >I don't know what happened, but a lot of the messages that I receive > >from sage-devel now have no quoting. (One of the paragraphs above is > >NOT mkoeppe). The quoting is there in HTML, but not in the plain-text > >version that I read. > > > >If I am failing to respond to questions, it's because I secretly have > >no idea what's going on in 75% of these conversations any more. > > > > Yes, I see this too. Another "improvement" of groups by GOOG. > > -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/b39dd187-a7d4-4293-9d8c-d9966b853bcen%40googlegroups.com.
