Upgrade pynac to support gcc-11

[mkoeppe]

(from #31786)

pynac fails to compile with gcc-11 (headers porting) pynac/pynac#375

... and ideally to consolidate the various patches added in #31679, #31645 etc.... as also requested in https://groups.google.com/g/sage-packaging/c/_vxPV7aKNjQ/m/b2-EdWP1BgAJ

When making a new release, also care should be taken to use a recent version of libtool that supports macOS Big Sur (see #31920).

CC: @dimpase @DaveWitteMorris @kiwifb @antonio-rojas

Component: packages: standard

Author: Dima Pasechnik, Matthias Koeppe

Branch: c9c50e3

Reviewer: Matthias Koeppe, Dima Pasechnik

Issue created by migration from https://trac.sagemath.org/ticket/31694

[DaveWitteMorris]

comment:1

There are still several pynac bug tickets that may not be hard to fix, so I think I will be posting quite a few more pynac patches in the next few months, so let's wait on this. But it would be good to make an upgrade before the 9.4 release.

[mkoeppe]

comment:2

Aren't you preparing your patches using git? I'd recommend to create pull requests to the pynac repo to make it more transparent.

[dimpase]

comment:7

I am going to cut a new release of pynac with all da patchez

[dimpase]

Commit: e97c1b5

[dimpase]

Branch: u/dimpase/packages/pynac/0.7.28

[dimpase]

Author: Dima Pasechnik

[dimpase]

New commits:

e97c1b5

update pynac to 0.7.28

[kiwifb]

comment:9

Thank you Dima! :)

[antonio-rojas]

comment:10

yah, thanks a lot!

[dimpase]

comment:11

Dave, I've merged all your pynac patches into master, e.g.

pynac/pynac@9d442a8

while the log is clear they are yours, I'm still the author. If you like you can do a PR on github to change the author of these commits.

[mkoeppe]

comment:12

GH Actions run please

[dimpase]

comment:13

Replying to @mkoeppe:

GH Actions run please

running at https://github.com/sagemath/sagetrac-mirror/actions/runs/972993967

[DaveWitteMorris]

comment:14

Replying to @dimpase:

Dave, I've merged all your pynac patches into master, e.g.

pynac/pynac@9d442a8

while the log is clear they are yours, I'm still the author. If you like you can do a PR on github to change the author of these commits.

Thanks, Dima, that's great! I am not concerned about the "author", so nothing needs to be done, as far as I'm concerned.

Copying these patches to the pynac repository has been on my to-do list for a long time, but I never got around to it, so thanks for doing this! In the future, I'll try to save you (and the distro managers) some trouble by making the appropriate PRs.

[dimpase]

comment:15

Replying to @DaveWitteMorris:

Copying these patches to the pynac repository has been on my to-do list for a long time, but I never got around to it, so thanks for doing this! In the future, I'll try to save you (and the distro managers) some trouble by making the appropriate PRs.

I guess you actually might have a git repo with these patches applied (at least this is how I typically do package patching), so it could have been a breeze to convert these into PRs. Anyway, it's done now.

[slel]

comment:16

One can set the author name and email as different from the committer's:

$ git commit --author="Author Name <[email protected]>"

This can be used to amend the last commit done:

$ git commit --amend --author="Author Name <[email protected]>" --no-edit

Reference:

• Answers to Stack Overflow question 3042437 (see more tips there)

[sheerluck]

comment:17

Replying to @dimpase:

Replying to @mkoeppe:

GH Actions run please

running at https://github.com/sagemath/sagetrac-mirror/actions/runs/972993967

gentoo failed bc system eclib wanted #31443

checking whether we can link and run a program using eclib...
eclib version 20210503, using NTL bigints and NTL real and complex multiprecision floating point
yes; use eclib from the system

[dimpase]

comment:18

Replying to @dimpase:

Replying to @mkoeppe:

GH Actions run please

running at https://github.com/sagemath/sagetrac-mirror/actions/runs/972993967

looks like that everything (modulo the usual GH actions suspects, e.g. broken conda on some platforms) is working.
Review?

[mkoeppe]

comment:19

Broken on ubuntu-trusty

[mkoeppe]

comment:56

Dima, could you please make a release without PR 378, so that we can get this blocker ticket done? Just focused on what this ticket says - add support for gcc 11, consolidate existing patches; no breaking/removing of other platform support.

[mkoeppe]

comment:57

I volunteer to make the next pynac releases.

[dimpase]

comment:58

Replying to @mkoeppe:

I volunteer to make the next pynac releases.

let's just drop gcc 4.x. It would minimise the disruption, and it's overdue anyway.

[DaveWitteMorris]

comment:59

As I said on #32074, I am +1 on this proposal to drop support for gcc 4.x. That should solve the immediate problem.

[dimpase]

comment:60

Or even better/quicker, just add AC_CHECK_FUNC(__builtin_smull_overflow,...) to spkg-configure for gcc spkg.

[mkoeppe]

comment:61

It's important to keep tickets focused. In particular for a blocker ticket like this one, which creates an urgency to get it in.

The urgent need to add support for gcc 11 is entirely unconnected to a decision to drop support for the gcc 4 versions used in some of our supported distributions.

It is poor practice to connect these two issues arbitrarily.

[dimpase]

comment:62

our resources are limited;

• making a new pynac release without __builtin_smull_overflow takes time,
• re-doing the branch with __builtin_smull_overflow takes time
• (and God help you finding a workaround not using __builtin_smull_overflow, potentially another time waste to keep a museum compiler in...)

[mkoeppe]

comment:63

It's important to keep tickets focused -- that includes avoiding to give unsolicited advice to other developers regarding time management.

[mkoeppe]

Changed branch from u/dimpase/packages/pynac/0.7.28 to u/mkoeppe/packages/pynac/0.7.28

[mkoeppe]

Changed author from Dima Pasechnik to Dima Pasechnik, Matthias Koeppe

[mkoeppe]

Changed commit from 11c80aa to c9c50e3

[mkoeppe]

New commits:

c9c50e3

build/pkgs/pynac: Upgrade to 0.7.29

[mkoeppe]

Reviewer: Matthias Koeppe, ...

[dimpase]

comment:67

tests run on https://github.com/sagemath/sagetrac-mirror/actions/runs/986545913

[dimpase]

comment:68

lgtm

[dimpase]

Changed reviewer from Matthias Koeppe, ... to Matthias Koeppe, Dima Pasechnik

[mkoeppe]

comment:69

Thanks!

[vbraun]

Changed branch from u/mkoeppe/packages/pynac/0.7.28 to c9c50e3

[dimpase]

comment:71

I don't see this branch in the current develop branch, what's up?

[dimpase]

Changed commit from c9c50e3 to none