[BETA] sagemath: update to 10.5.beta7

[tornaria]

• python3-Cython: alternative fix for 6122
• sagemath: enable all cython warnings
• nauty: update to 2.8.9.

Testing the changes

• I tested the changes in this PR: YES

@dkwo

[dkwo]

I tried to crosscompile sagemath from aarch64 to x86_64: it looks like we should add to hostmakedepends "python3-gmpy2 python3-memory_allocator python3-numpy ecl python3-cypari2 python3-cysignals python3-devel". Makes sense?

[tornaria]

I tried to crosscompile sagemath from aarch64 to x86_64: it looks like we should add to hostmakedepends "python3-gmpy2 python3-memory_allocator python3-numpy ecl python3-cypari2 python3-cysignals python3-devel". Makes sense?

Sure, but we still don't have cross for ntl and others, does it make sense to bother with this?

[dkwo]

True. It's still useful to know in general, and it could be useful later, if we decide e.g. to use qemu to build ntl etc.

[tornaria]

@dkwo is something like this enough to cross-compile:

--- a/srcpkgs/sagemath/template
+++ b/srcpkgs/sagemath/template
@@ -6,7 +6,9 @@ _pypi_version=${version/.beta/b}
 _pypi_version=${_pypi_version/.rc/rc}
 build_style=python3-pep517
 hostmakedepends="pkg-config python3-Cython python3-Jinja2
- python3-pkgconfig python3-setuptools python3-wheel"
+ python3-pkgconfig python3-setuptools python3-wheel
+ python3-gmpy2 python3-memory_allocator python3-numpy ecl
+ python3-cypari2 python3-cysignals python3-devel"
 makedepends="boost-devel brial-devel cliquer-devel ecl eclib-devel
  ecm-devel fflas-ffpack flintlib-devel gap-devel gd-devel giac-devel glpk-devel
  gsl-devel iml-devel lcalc-devel libbraiding-devel libhomfly-devel libmpc-devel

[dkwo]

Yes, that should be enough. I tried to build and check on it aarch64: it fails only one test sage/rings/number_theory/galois_group.py. After installing sage, I cannot run it though, I get

/usr/bin/sage: line 177: /usr/libexec/sagemath/sage-env-config: Permission denied
/usr/bin/sage: line 916: exec: sage-runtests: not found

[tornaria]

Yes, that should be enough.

:+1: I pushed that change here.

I tried to build and check on it aarch64: it fails only one test sage/rings/number_theory/galois_group.py.

Nice, at some point we may want to debug that so tests pass. Can you post the failure?

After installing sage, I cannot run it though, I get

/usr/bin/sage: line 177: /usr/libexec/sagemath/sage-env-config: Permission denied
/usr/bin/sage: line 916: exec: sage-runtests: not found

That's weird. The second error is expected from the first (because sage-env-config sets the correct PATH so if it doesn't run the PATH is incorrect).

What's up with sage-env-config? I don't make it +x but it should not be necessary as line 177 sources it. Am I missing something? Here:

$ ls -l /usr/libexec/sagemath/
total 252
-rwxr-xr-x 1 root root   785 Dec 25 19:20 math-readline
-rwxr-xr-x 1 root root 43267 Dec 25 19:20 sage
-rwxr-xr-x 1 root root   651 Dec 25 19:20 sage-cachegrind
-rwxr-xr-x 1 root root   442 Dec 25 19:20 sage-callgrind
-rwxr-xr-x 1 root root  7788 Dec 25 19:20 sage-cleaner
-rwxr-xr-x 1 root root 10209 Dec 25 19:20 sage-coverage
-rwxr-xr-x 1 root root  1356 Dec 25 19:20 sage-cython
-rwxr-xr-x 1 root root 23416 Dec 25 19:20 sage-env
-rw-r--r-- 1 root root   205 Dec 25 19:20 sage-env-config
-rwxr-xr-x 1 root root   238 Dec 25 19:20 sage-eval
-rwxr-xr-x 1 root root 35314 Dec 25 19:20 sage-fixdoctests
-rwxr-xr-x 1 root root   119 Dec 25 19:20 sage-grep
-rwxr-xr-x 1 root root    70 Dec 25 19:20 sage-grepdoc
-rwxr-xr-x 1 root root   234 Dec 25 19:20 sage-inline-fortran
-rwxr-xr-x 1 root root  1345 Dec 25 19:20 sage-ipynb2rst
-rwxr-xr-x 1 root root   352 Dec 25 19:20 sage-ipython
-rwxr-xr-x 1 root root  3660 Dec 25 19:20 sage-list-packages
-rwxr-xr-x 1 root root   613 Dec 25 19:20 sage-massif
-rwxr-xr-x 1 root root  8463 Dec 25 19:20 sage-notebook
-rwxr-xr-x 1 root root  3138 Dec 25 19:20 sage-num-threads.py
-rwxr-xr-x 1 root root   595 Dec 25 19:20 sage-omega
-rwxr-xr-x 1 root root 10605 Dec 25 19:20 sage-preparse
-rwxr-xr-x 1 root root    28 Dec 25 19:20 sage-python
-rwxr-xr-x 1 root root   699 Dec 25 19:20 sage-run
-rwxr-xr-x 1 root root   245 Dec 25 19:20 sage-run-cython
-rwxr-xr-x 1 root root   121 Dec 25 19:20 sage-runtests
-rwxr-xr-x 1 root root  6219 Dec 25 19:20 sage-startuptime.py
-rwxr-xr-x 1 root root  5646 Dec 25 19:20 sage-update-version
-rwxr-xr-x 1 root root  1201 Dec 25 19:20 sage-valgrind
-rwxr-xr-x 1 root root  1207 Dec 25 19:20 sage-venv-config
-rwxr-xr-x 1 root root   394 Dec 25 19:20 sage-version.sh

[dkwo]

For me, that file is only readable to root:

$ ls -l /usr/libexec/sagemath/ 
total 252
-rwxr-xr-x 1 root root   785 Dec 26 17:58 math-readline
-rwxr-xr-x 1 root root 43267 Dec 26 17:58 sage
-rwxr-xr-x 1 root root   651 Dec 26 17:58 sage-cachegrind
-rwxr-xr-x 1 root root   442 Dec 26 17:58 sage-callgrind
-rwxr-xr-x 1 root root  7788 Dec 26 17:58 sage-cleaner
-rwxr-xr-x 1 root root 10209 Dec 26 17:58 sage-coverage
-rwxr-xr-x 1 root root  1356 Dec 26 17:58 sage-cython
-rwxr-xr-x 1 root root 23416 Dec 26 17:58 sage-env
-rw------- 1 root root   205 Dec 26 17:58 sage-env-config
-rwxr-xr-x 1 root root   238 Dec 26 17:58 sage-eval
-rwxr-xr-x 1 root root 35314 Dec 26 17:58 sage-fixdoctests
-rwxr-xr-x 1 root root   119 Dec 26 17:58 sage-grep
-rwxr-xr-x 1 root root    70 Dec 26 17:58 sage-grepdoc
-rwxr-xr-x 1 root root   234 Dec 26 17:58 sage-inline-fortran
-rwxr-xr-x 1 root root  1345 Dec 26 17:58 sage-ipynb2rst
-rwxr-xr-x 1 root root   352 Dec 26 17:58 sage-ipython
-rwxr-xr-x 1 root root  3660 Dec 26 17:58 sage-list-packages
-rwxr-xr-x 1 root root   613 Dec 26 17:58 sage-massif
-rwxr-xr-x 1 root root  8463 Dec 26 17:58 sage-notebook
-rwxr-xr-x 1 root root  3138 Dec 26 17:58 sage-num-threads.py
-rwxr-xr-x 1 root root   595 Dec 26 17:58 sage-omega
-rwxr-xr-x 1 root root 10605 Dec 26 17:58 sage-preparse
-rwxr-xr-x 1 root root    28 Dec 26 17:58 sage-python
-rwxr-xr-x 1 root root   699 Dec 26 17:58 sage-run
-rwxr-xr-x 1 root root   245 Dec 26 17:58 sage-run-cython
-rwxr-xr-x 1 root root   121 Dec 26 17:58 sage-runtests
-rwxr-xr-x 1 root root  6219 Dec 26 17:58 sage-startuptime.py
-rwxr-xr-x 1 root root  5646 Dec 26 17:58 sage-update-version
-rwxr-xr-x 1 root root  1201 Dec 26 17:58 sage-valgrind
-rwxr-xr-x 1 root root  1207 Dec 26 17:58 sage-venv-config
-rwxr-xr-x 1 root root   394 Dec 26 17:58 sage-version.sh

[dkwo]

If I chmod it by hand, then everything works and that test fails like this:

Running doctests with ID 2024-12-27-21-07-26-dc66dad2.
Running with SAGE_LOCAL='/usr' and SAGE_VENV='/usr'
Using --optional=pip,sage
Features to be detected: 4ti2,SAGE_SRC,benzene,bliss,buckygen,conway_polynomials,coxeter3,csdp,cvxopt,cvxopt,database_cremona_ellcurve,database_cremona_mini_ellcurve,database_cubic_hecke,database_ellcurves,database_graphs,database_jones_numfield,database_knotinfo,dot2tex,dvipng,ecm,fpylll,fricas,gap_package_atlasrep,gap_package_design,gap_package_grape,gap_package_guava,gap_package_hap,gap_package_polenta,gap_package_polycyclic,gap_package_qpa,gap_package_quagroup,gfan,giac,glucose,graphviz,imagemagick,info,ipython,jmol,jupymake,jupyter_sphinx,kenzo,kissat,latte_int,lrcalc_python,lrslib,mathics,matroid_database,mcqd,meataxe,mpmath,msolve,nauty,networkx,numpy,palp,pandoc,pdf2svg,pdftocairo,pexpect,phitigra,pillow,plantri,polytopes_db,polytopes_db_4d,pplpy,primecountpy,ptyprocess,pycosat,pycryptosat,pynormaliz,pyparsing,python_igraph,requests,rpy2,rubiks,sage.combinat,sage.geometry.polyhedron,sage.graphs,sage.groups,sage.libs.braiding,sage.libs.ecl,sage.libs.flint,sage.libs.gap,sage.libs.giac,sage.libs.homfly,sage.libs.linbox,sage.libs.m4ri,sage.libs.ntl,sage.libs.pari,sage.libs.singular,sage.misc.cython,sage.modular,sage.modules,sage.numerical.mip,sage.plot,sage.rings.complex_double,sage.rings.finite_rings,sage.rings.function_field,sage.rings.number_field,sage.rings.padics,sage.rings.polynomial.pbori,sage.rings.real_double,sage.rings.real_mpfr,sage.sat,sage.schemes,sage.symbolic,sage_numerical_backends_coin,sagemath_doc_html,scipy,singular,sirocco,sloane_database,sphinx,symengine_py,sympy,tdlib,threejs,topcom
Doctesting 1 file.
sage -t --warn-long 5.0 --random-seed=338122519044229990874824510482622064820 /usr/lib/python3.13/site-packages/sage/rings/number_field/galois_group.py
**********************************************************************
File "/usr/lib/python3.13/site-packages/sage/rings/number_field/galois_group.py", line 997, in sage.rings.number_field.galois_group.GaloisGroup_v2.artin_symbol
Failed example:
    [G.artin_symbol(P) for P in K.primes_above(7)]
Expected:
    [(1,5)(2,6)(3,7)(4,8), (1,5)(2,6)(3,7)(4,8),
     (1,4)(2,3)(5,8)(6,7), (1,4)(2,3)(5,8)(6,7)]
Got:
    [(1,4)(2,3)(5,8)(6,7),
     (1,4)(2,3)(5,8)(6,7),
     (1,5)(2,6)(3,7)(4,8),
     (1,5)(2,6)(3,7)(4,8)]
**********************************************************************
1 item had failures:
   1 of   8 in sage.rings.number_field.galois_group.GaloisGroup_v2.artin_symbol
    [295 tests, 1 failure, 0.80s wall]
----------------------------------------------------------------------
sage -t --warn-long 5.0 --random-seed=338122519044229990874824510482622064820 /usr/lib/python3.13/site-packages/sage/rings/number_field/galois_group.py  # 1 doctest failed
----------------------------------------------------------------------
Total time for all tests: 1.6 seconds
    cpu time: 0.8 seconds
    cumulative wall time: 0.8 seconds
Features detected for doctesting: sage.groups,sage.rings.number_field,sage.symbolic

[dkwo]

it seems that's just an ordering issue?

[tornaria]

-rw------- 1 root root 205 Dec 26 17:58 sage-env-config

This probably means that the file in your void-packages repo is already 600 and we cp the file. There's a fair number of templates doing that, but most use vinstall.

I think we can do something like:

--- a/srcpkgs/sagemath/template
+++ b/srcpkgs/sagemath/template
@@ -51,7 +51,7 @@ post_install() {
        mv -T ${DESTDIR}/usr/bin ${DESTDIR}/usr/libexec/sagemath
 
        # copy configuration
-       cp ${FILESDIR}/sage-env-config ${DESTDIR}/usr/libexec/sagemath
+       vinstall ${FILESDIR}/sage-env-config 644 usr/libexec/sagemath
 
        # symlink main binary
        vmkdir usr/bin

to make sure the permissions are 644 regardless.

[tornaria]

it seems that's just an ordering issue?

Yes, I'll try to adjust so the ordering doesn't affect the test.

[tornaria]

@dkwo both issues should be fixed now :crossed_fingers:

[dkwo]

Thanks a lot! Everything works for me now on aarch64. cc @leahneukirchen