That discussion about the Sage distribution etc.

[Matthias Koeppe]

Here's a quick overview on the discussions on this topic since 2021.

(I collected most of this recently in the Issue description of https://github.com/sagemath/sage/issues/32532).

• sage-devel: proposal - remove gcc, gfortran, python building/spkgs (June 2021)

  • Resulted in implementing configure: Change defaults to --with-system-gcc=force, defer error exits after system package info is shown #32060
  • No clear support for removing anything without replacement.
  • Some talk about conda, no action.

• sage-devel: #32532 - removing gcc and gfortran spkgs (September 2021)

  • No clear support for anything.
  • Some talk about conda, no action.

• sage-devel: Re: [sage-release] Sage 10.0.rc0 released (Apr 2023) (TW: abusive comments)

• sage-devel: What was/is/will be the purpose of maintaining the Sage distribution? (Apr 2023)

  • Again talk about conda
  • Resulted in opening PR Install some dummy and optional packages using conda-forge (micromamba) #35585; no reaction

• sage-devel: Policy for disputed PRs: discussion (Nov 2023). 

• CI Linux: Replace use of pkill #36726 (Nov–Dec 2023) (TW: abusive comments)

Pointers to related discussions and facts:

On the git repository structure (monorepo / multirepo): https://groups.google.com/g/sage-devel/c/AaKxoNQAWMg/m/uY1rW5n0BQAJ (October 2021), https://github.com/sagemath/sage/pull/36777 (December 2023).

On what systems we support building Sage from source: https://github.com/sagemath/sage/issues/32074

What versions of Python we are supporting; including a discussion of maintenance costs: https://github.com/sagemath/sage/wiki/NEP-29:-Python-version-strategy

[Dima Pasechnik]

From my perspective, throughout these discussions there is a strong
push against slimming Sage down by removing these unneeded packages
coming from Matthias, often coupled with responses by him (on
trac/github) I consider rude and disrespectful of other's opinions.

See e.g. https://github.com/sagemath/sage/issues/32532
You can see how I am subjected to gaslighting by Matthias there.
(Let's do this/no, stop, don't/just drop this/-1 (without
explaining/... coming up in cycles).
Description of this issues/32532 was edited in a questionable way by
Matthias recently.

> --
> 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+...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/87c9218a-9b6b-4fb9-8f35-78624ba67054n%40googlegroups.com.

[Dima Pasechnik]

On Fri, Jan 12, 2024 at 11:01 AM Dima Pasechnik <dim...@gmail.com> wrote:
>
> From my perspective, throughout these discussions there is a strong
> push against slimming Sage down by removing these unneeded packages
> coming from Matthias, often coupled with responses by him (on
> trac/github) I consider rude and disrespectful of other's opinions.
>
> See e.g. https://github.com/sagemath/sage/issues/32532
> You can see how I am subjected to gaslighting by Matthias there.
> (Let's do this/no, stop, don't/just drop this/-1 (without
> explaining/... coming up in cycles).
> Description of this issues/32532 was edited in a questionable way by
> Matthias recently.

PS. Needless to say, the need for gcc/gfortran in bundled in Sage is
really not there, since years and years.
On macOS one has gfortran from Homebrew, and the argument that
Homebrew is "insecure"
doesn't really fly any more, as Homebrew packages are nowadays
notarized/signed, etc etc.