SageMath 10.7 on macOS 15.5 #40501

GMS103 · 2025-07-28T23:34:29Z

Fixes issue #40500.

Sage 10.7.beta9 does not compile on macOS 15.5 (Sequoia) when doing
./configure --enable-gap_packages --enable-symengine.

The errors are the following.

qhull: Error installing package qhull-2020-src-8.0.2
symengine: Error installing package symengine-0.11.2
gap_packages: Error installing package gap_packages-4.14.0
(the last error includes a problem with libsemigroups)

In this PR we include 4 patches by Marc Culler which allow SageMath to build.

See
https://groups.google.com/g/sage-release/c/RjuHtop_dY0/m/Qwg-ekS9CwAJ
https://groups.google.com/g/sage-release/c/aG3voHcxUn4/m/JXTFZ9W9CwAJ
https://groups.google.com/g/sage-release/c/12-BYGMj4Og/m/MCycYfe9CwAJ
https://groups.google.com/g/sage-release/c/ECrJI6ZiUmQ/m/HjytIFO-CwAJ

📝 Checklist

The title is concise and informative.
The description explains in detail what this PR is about.
I have linked a relevant issue or discussion.
I have created tests covering the changes.
I have updated the documentation and checked the documentation preview.

Tests: make ptestlong gives only one error:
sage -t --long --warn-long 30.0 --random-seed=… src/sage/algebras/quantum_groups/quantum_group_gap.py # 1 doctest failed
related to TikZ.

⌛ Dependencies

None.

GMS103 · 2025-07-28T23:45:54Z

(This is my first PR, please bear with me and correct as you see fit.)

enriqueartal · 2025-07-29T08:48:33Z

I tried Marc Culler patch in Fedora 42 without success. It worked with the patch from Fedora. Is compatible this patch with #40484?

enriqueartal · 2025-07-29T11:18:38Z

I was doing something wrong last time I tried this patch. Now it works in Fedora 42 with gcc-15. Hence I close #40484. I would like to recuperate only this sentence from that PR:
"There is another point concerning gap_packages. I wonder if it is better to have separated packages; now the failure in one package prevents the installation of the whole spkg."

tobiasdiez · 2025-07-29T14:08:08Z

There are now failed runs on debian and fedora, see https://github.com/sagemath/sage/actions/runs/16582710906/job/46902077059?pr=40501. Could you please have a look?

GMS103 · 2025-07-29T15:56:35Z

Looking at the CI logs, this does not seem to be a real error.
The reported error is that the symengine package file could not be downloaded. This could be because the PR upgrades the version of symengine, but no one uploaded the new symengine tarball to the sage repo. So it could not be downloaded from there.
It seems the build script did not find the upstream URL for the tarball. Several build systems (including mine) were able to download the upstream tarball. Could it be that the docker image has some extra IP filtering?
BTW, when building I often had problems downloading from mirrors. Eventually I put the needed files in /upstream to speed things up.

raise FileNotMirroredError('tarball does not exist on mirror network and neither at the upstream URL')
[symengine-0.14.0] sage_bootstrap.tarball.FileNotMirroredError: tarball does not exist on mirror network and neither at the upstream URL
[symengine-0.14.0] ************************************************************************
[symengine-0.14.0] Error downloading tarball of symengine
[symengine-0.14.0] ************************************************************************

(credit: MC)

tobiasdiez · 2025-07-29T16:55:03Z

That part is indeed okay. But there are failing doctests as well, on both systems.

dimpase · 2025-07-29T17:55:54Z

I think libsemigroups should be updated to 2.7.4, and semigroups split into a separate optional package and upgraded to 5.5.3.
This would drop all these (lib)semigroup patches, and allow building with gcc-15, too, I think.
See gentoo/gentoo#43227

dimpase · 2025-07-29T18:00:30Z

Looking at the CI logs, this does not seem to be a real error. The reported error is that the symengine package file could not be downloaded. This could be because the PR upgrades the version of symengine, but no one uploaded the new symengine tarball to the sage repo. So it could not be downloaded from there. It seems the build script did not find the upstream URL for the tarball. Several build systems (including mine) were able to download the upstream tarball. Could it be that the docker image has some extra IP filtering? BTW, when building I often had problems downloading from mirrors. Eventually I put the needed files in /upstream to speed things up.
raise FileNotMirroredError('tarball does not exist on mirror network and neither at the upstream URL')
[symengine-0.14.0] sage_bootstrap.tarball.FileNotMirroredError: tarball does not exist on mirror network and neither at the upstream URL
[symengine-0.14.0] ************************************************************************
[symengine-0.14.0] Error downloading tarball of symengine
[symengine-0.14.0] ************************************************************************
(credit: MC)

symengine is downloadable (or it should be, otherwise it's a bug in itself) from the URL in checksums.ini:
cf. above: "tarball does not exist on mirror network and neither at the upstream URL"

But I checked that it works - so it probably was some intermittent GitHub error.

 $ wget https://github.com/symengine/symengine/releases/download/v0.14.0/symengine-0.14.0.tar.gz
--2025-07-29 13:02:08--  https://github.com/symengine/symengine/releases/download/v0.14.0/symengine-0.14.0.tar.gz
Resolving github.com... 140.82.116.4
Connecting to github.com|140.82.116.4|:443... connected.
HTTP request sent, awaiting response... 302 Found
Location: https://release-assets.githubusercontent.com/github-production-release-asset/42376/d7c0633b-fd61-4a6d-9059-ea999b244a49?sp=r&sv=2018-11-09&sr=b&spr=https&se=2025-07-29T18%3A51%3A21Z&rscd=attachment%3B+filename%3Dsymengine-0.14.0.tar.gz&rsct=application%2Foctet-stream&skoid=96c2d410-5711-43a1-aedd-ab1947aa7ab0&sktid=398a6654-997b-47e9-b12b-9515b896b4de&skt=2025-07-29T17%3A50%3A33Z&ske=2025-07-29T18%3A51%3A21Z&sks=b&skv=2018-11-09&sig=%2B18dQrFCxf4oe1sTKJcqaR86xdsVTICP%2FVoW4FCL7gQ%3D&jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJpc3MiOiJnaXRodWIuY29tIiwiYXVkIjoicmVsZWFzZS1hc3NldHMuZ2l0aHVidXNlcmNvbnRlbnQuY29tIiwia2V5Ijoia2V5MSIsImV4cCI6MTc1MzgxMjQzMCwibmJmIjoxNzUzODEyMTMwLCJwYXRoIjoicmVsZWFzZWFzc2V0cHJvZHVjdGlvbi5ibG9iLmNvcmUud2luZG93cy5uZXQifQ.Mo0hNBiZCP9f4Vqf3LaMkCxnPSdBILzjP8NSJ-VERco&response-content-disposition=attachment%3B%20filename%3Dsymengine-0.14.0.tar.gz&response-content-type=application%2Foctet-stream [following]
--2025-07-29 13:02:09--  https://release-assets.githubusercontent.com/github-production-release-asset/42376/d7c0633b-fd61-4a6d-9059-ea999b244a49?sp=r&sv=2018-11-09&sr=b&spr=https&se=2025-07-29T18%3A51%3A21Z&rscd=attachment%3B+filename%3Dsymengine-0.14.0.tar.gz&rsct=application%2Foctet-stream&skoid=96c2d410-5711-43a1-aedd-ab1947aa7ab0&sktid=398a6654-997b-47e9-b12b-9515b896b4de&skt=2025-07-29T17%3A50%3A33Z&ske=2025-07-29T18%3A51%3A21Z&sks=b&skv=2018-11-09&sig=%2B18dQrFCxf4oe1sTKJcqaR86xdsVTICP%2FVoW4FCL7gQ%3D&jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJpc3MiOiJnaXRodWIuY29tIiwiYXVkIjoicmVsZWFzZS1hc3NldHMuZ2l0aHVidXNlcmNvbnRlbnQuY29tIiwia2V5Ijoia2V5MSIsImV4cCI6MTc1MzgxMjQzMCwibmJmIjoxNzUzODEyMTMwLCJwYXRoIjoicmVsZWFzZWFzc2V0cHJvZHVjdGlvbi5ibG9iLmNvcmUud2luZG93cy5uZXQifQ.Mo0hNBiZCP9f4Vqf3LaMkCxnPSdBILzjP8NSJ-VERco&response-content-disposition=attachment%3B%20filename%3Dsymengine-0.14.0.tar.gz&response-content-type=application%2Foctet-stream
Resolving release-assets.githubusercontent.com... 185.199.109.133, 185.199.111.133, 185.199.110.133, ...
Connecting to release-assets.githubusercontent.com|185.199.109.133|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 944589 (922K) [application/octet-stream]
Saving to: ‘symengine-0.14.0.tar.gz’

symengine-0.14.0.tar.gz                                100%[============================================================================================================================>] 922.45K  --.-KB/s    in 0.1s    

2025-07-29 13:02:09 (8.15 MB/s) - ‘symengine-0.14.0.tar.gz’ saved [944589/944589]

$ sha256sum symengine-0.14.0.tar.gz 
11c5f64e9eec998152437f288b8429ec001168277d55f3f5f1df78e3cf129707  symengine-0.14.0.tar.gz
$ sha1sum symengine-0.14.0.tar.gz 
e53aa2677bcad5a5dc075a94d4a4f9ef95b5fe07  symengine-0.14.0.tar.gz

dimpase · 2025-07-29T19:53:40Z

please see #40509 for a better way to deal with (lib)semigroups blues.

GMS103 · 2025-07-29T20:11:26Z

Tobias said:

That part is indeed okay. But there are failing doctests as well, on both systems.

Dima said:

please see #40509 for a better way to deal with (lib)semigroups blues.

Which one should I try to deal with first?

GMS103 · 2025-07-29T20:14:37Z

Would it be OK if I merge #40509 with this PR (after testing it of course)?

enriqueartal · 2025-07-29T20:35:42Z

The three tests in fedora 42 are OK if run on their own.

dimpase · 2025-07-29T20:36:41Z

sure, just don't forget to remove the (lib) Semigroups-related changes/patches here.

dimpase · 2025-07-30T05:39:13Z

@GMS103 - you can merge my PR to your repo to rebase this over #40509

New 40501

enriqueartal · 2025-07-30T11:07:11Z

Likely you merged before @dimpase's last change (line 115 in spk-install.in in gap_packages); patches for semigroups and libsemigroups are no more needed.

GMS103 · 2025-07-30T13:01:24Z

Sorry for my ignorance, I made a mistake.
How do I put https://github.com/GMS103/sage here again?

dimpase · 2025-07-30T14:23:48Z

reopen this PR?

GMS103 · 2025-07-30T16:01:52Z

Please see #40512.
And feel free to sort things out.

GMS103 added 2 commits July 29, 2025 00:11

Changes for clang 17 on macOS

1a2245d

Changes for clang 17 on macOS

ac4ca26

GMS103 mentioned this pull request Jul 28, 2025

SageMath 10.7 on macOS 15.5 #40500

Open

2 tasks

GMS103 added p: blocker / 1 s: needs review t: bug labels Jul 28, 2025

GMS103 requested a review from vbraun July 28, 2025 23:44

GMS103 requested a review from jhpalmieri July 29, 2025 00:35

vbraun added the s: positive review label Jul 29, 2025

enriqueartal mentioned this pull request Jul 29, 2025

gap_packages and gcc15 #40484

Closed

5 tasks

tobiasdiez added s: needs work and removed s: needs review s: positive review labels Jul 29, 2025

dimpase added 2 commits July 29, 2025 13:23

update libsemigroups to 2.7.4

758bb59

sping GAP's semigroups package into a separate spkg

2c075bb

enriqueartal mentioned this pull request Jul 29, 2025

split Semigroups into a separate spkg, bump versions #40509

Merged

5 tasks

bump symengine to 0.14.0

affc125

patch ghull for Apple clang 17

b8e907e

Merge pull request #1 from dimpase/new40501

8785d91

New 40501

github-actions bot added s: needs review and removed s: needs work labels Jul 30, 2025

GMS103 closed this by deleting the head repository Jul 30, 2025

GMS103 mentioned this pull request Jul 30, 2025

SageMath 10.7 on macOS 15.5 (again) #40512

Closed

Uh oh!

SageMath 10.7 on macOS 15.5 #40501

SageMath 10.7 on macOS 15.5 #40501

Uh oh!

Conversation

GMS103 commented Jul 28, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

📝 Checklist

⌛ Dependencies

Uh oh!

GMS103 commented Jul 28, 2025

Uh oh!

enriqueartal commented Jul 29, 2025

Uh oh!

enriqueartal commented Jul 29, 2025

Uh oh!

tobiasdiez commented Jul 29, 2025

Uh oh!

GMS103 commented Jul 29, 2025

Uh oh!

tobiasdiez commented Jul 29, 2025

Uh oh!

dimpase commented Jul 29, 2025

Uh oh!

dimpase commented Jul 29, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

dimpase commented Jul 29, 2025

Uh oh!

GMS103 commented Jul 29, 2025

Uh oh!

GMS103 commented Jul 29, 2025

Uh oh!

enriqueartal commented Jul 29, 2025

Uh oh!

dimpase commented Jul 29, 2025

Uh oh!

dimpase commented Jul 30, 2025

Uh oh!

enriqueartal commented Jul 30, 2025

Uh oh!

GMS103 commented Jul 30, 2025

Uh oh!

dimpase commented Jul 30, 2025

Uh oh!

GMS103 commented Jul 30, 2025

Uh oh!

Uh oh!

GMS103 commented Jul 28, 2025 •

edited

Loading

dimpase commented Jul 29, 2025 •

edited

Loading