Sage 9.3.rc2 released

[Volker Braun]

As always, you can get the latest beta version from the "develop" git branch. Alternatively, the self-contained source tarball is at http://www.sagemath.org/download-latest.html 

824f9cccaf (tag: 9.3.rc2, trac/develop) Updated SageMath version to 9.3.rc2

4275b275a2 Trac #31530: New subs bug caused by #30378

70711cf13d Trac #31502: GH Actions: Free more disk space

f38d467ae0 Trac #31584: Fix suitesparse/cvxopt /usr/local leakage

22a11999f2 Trac #31552: Update singular to stop /usr/local from leaking into Singular build

24d1bdb786 Trac #31523: Fix dependencies of ipywidgets

ad70e647c5 Trac #31409: Make it possible to disable build of r, rpy2 using ./configure --disable-r

7b44d1cb00 Trac #30888: Make $SAGE_LOCAL/bin/sage work again from any directory, in an environment without SAGE_* variables, following symlinks

8736c424b9 Trac #31538: i386: openssl and tachyon attempt 64-bit build and fail

c7c058bd65 Trac #31395: RuntimeError: can't start new thread (due to memlimit)

9a1e3ccb6b Trac #31558: homomorphism of extension fields do not preserve canonical embeddings of the base

0606dc72e3 Trac #31551: Incorrect conversion from ℚ[-i] to SR

71123d8021 Trac #30319: Upgrade cddlib to fix wrong intersection of polytopes

3da0a29d9f Trac #31536: Fix Sage rich output problems by coupling MathJax with html

a89f81d778 (tag: 9.3.rc1) Updated SageMath version to 9.3.rc1

[Steven Trogdon]

Incremental upgrade from rc1 -> rc2 did not build here with

git pull && make

There was at least one Sage component that was linked against the old Singular. Rebuilding sagelib then worked

./sage -f sagelib && make

Perhaps there was another way to accomplish the build.

[Kwankyu Lee]

Thanks for the release. 

Incremental build from 9.3.rc1 on mac pro with Big Sur succeeded with no problem. All doctests passed. 

[Kenji Iohara]

On my mac with OS 11.2.3, the first built worked perfectly. 

I haven't tried ptestlong yet.

2021年4月7日(水) 3:13 Steven Trogdon <steven....@gmail.com>:

--
You received this message because you are subscribed to the Google Groups "sage-release" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sage-release...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/sage-release/5665e39b-d8a9-4dfa-a4ea-ae82d0599d35n%40googlegroups.com.

[Matthias Köppe]

On Tuesday, April 6, 2021 at 3:33:00 PM UTC-7 Volker Braun wrote:

As always, you can get the latest beta version from the "develop" git branch. Alternatively, the self-contained source tarball is at http://www.sagemath.org/download-latest.html 

824f9cccaf (tag: 9.3.rc2, trac/develop) Updated SageMath version to 9.3.rc2

Thanks, Volker, for the new release candidate! Looks like we are getting closer to a working release.

Tests on various platforms are running at:

- https://github.com/sagemath/sage/actions/runs/724020346 (Linux and macOS)

- https://github.com/sagemath/sage/actions/runs/724020349 (Cygwin)

- https://github.com/sagemath/sage/actions/runs/724020348 (optional packages)

All: We do not have working automatic tests for the Windows Subsystem for Linux platforms (WSL1 and WSL2). If you have access to a machine running Windows 10, testing on these platforms will be very helpful. (https://trac.sagemath.org/ticket/31612 (needs review) may or may not be needed.)

Volker: Repeating my suggestion in https://trac.sagemath.org/ticket/22657#comment:10 (to which I have not received a reaction yet), could you please run the binary building scripts for this (or the next) release candidate so that developers can test whether the built binaries work -- in particular, whether they are portable between processor variants.

Volker: I would like to see a number of "workhorse" tickets merged in the next release candidate:

- https://trac.sagemath.org/ticket/31592 - another fix to the Singular build scripts, fixes a reported build failure

- https://trac.sagemath.org/ticket/31496 - removes a message from the configure output with big potential to confuse users

- https://trac.sagemath.org/ticket/31554 and https://trac.sagemath.org/ticket/31479 - symbolics bugs with blocker flavor

- https://trac.sagemath.org/ticket/31515 - sympy bugfix upgrade (1.7->1.7.1)

- https://trac.sagemath.org/ticket/31532, https://trac.sagemath.org/ticket/30213, https://trac.sagemath.org/ticket/31541 (needs review), https://trac.sagemath.org/ticket/31534 (needs review) - more updates to the GH Actions workflows so that our demonstration of platform support in the release tour will be more impressive

- https://trac.sagemath.org/ticket/31505 (needs review) - no longer refuse to build when MacPorts is active

- https://trac.sagemath.org/ticket/27804 - enables build of the polymake package on platforms where our lrslib package is broken (I know, not a blocker ticket by definition)

All: Thanks to all who have contributed to https://wiki.sagemath.org/ReleaseTours/sage-9.3! More help, also with balancing out the writing, is very welcome. 

All: Also some last minute work on our documentation would be very valuable: https://trac.sagemath.org/ticket/29784, https://trac.sagemath.org/ticket/31101

 

All: In addition to testing the SageMath distribution, please also help testing our new PyPI distribution packages ("pip-installable Sage"). See https://pypi.org/project/sagemath-standard/ and https://pypi.org/project/sage-conf/ . (The pip install of the source packages should work on all platforms; the pre-built wheels are only available for macOS 10.15 Python 3.8 at the moment.)

[Kenji Iohara]

On mac 0S 11.2.3, I could built Sage from the very beginning but make ptestlong failed. 

[Matthias Köppe]

From the log:

- many instances of fatal error: 'gmp.h' file not found

- many instances of 'ld: library not found for -lmpfr'

This is https://trac.sagemath.org/ticket/31348

As a workaround, it may help to run "make ptestlong" in the shell after using ". .homebrew-build-env"

[Samuel Lelièvre]

Thanks Volker for the new release and Matthias for the overview
of remaining work.

On macOS 10.14.6 with many Homebrew packages, from an
existing SageMath 9.3.rc0, after the following commands:
```
$ source .homebrew-build-env && ./bootstrap -q && ./configure -q
```
the extraneous configure message (dealt with at #31496) was
displayed. After that, the following commands
```
$ MAKE='make -j1' && make -s V=0 && make -s V=0 testlong
```
gave "All tests passed!". : ) : ) : ) --Samuel

[Matthias Köppe]

On Wednesday, April 7, 2021 at 10:29:29 AM UTC-7 Matthias Köppe wrote:

On Tuesday, April 6, 2021 at 3:33:00 PM UTC-7 Volker Braun wrote:

As always, you can get the latest beta version from the "develop" git branch. Alternatively, the self-contained source tarball is at http://www.sagemath.org/download-latest.html 

824f9cccaf (tag: 9.3.rc2, trac/develop) Updated SageMath version to 9.3.rc2

Tests on various platforms are running at:

- https://github.com/sagemath/sage/actions/runs/724020346 (Linux and macOS)

- https://github.com/sagemath/sage/actions/runs/724020349 (Cygwin)

- https://github.com/sagemath/sage/actions/runs/724020348 (optional packages)

Here's a summary of these runs. I am comparing to 9.3.beta8, for which the results were summarized in:

- https://groups.google.com/g/sage-release/c/KdSKg6RdZok/m/zM2Bhfm0AwAJ, 

- https://groups.google.com/g/sage-release/c/KdSKg6RdZok/m/75iHB1nYAwAJ,

- https://groups.google.com/g/sage-release/c/KdSKg6RdZok/m/3Fkd7K1aAgAJ, 

- https://groups.google.com/g/sage-release/c/KdSKg6RdZok/m/GXfB1rW1AgAJ:

9.3.beta8> ubuntu-{trusty,xenial,bionic,focal,groovy,hirsute}-{minimal,standard}: Clean

9.3.beta8> debian-{jessie,stretch,buster,bullseye,sid}-{minimal,standard}: Clean

9.3.beta8> ... with the exception of:

9.3.beta8> {debian-bullseye,ubuntu-groovy}-standard: cvxopt testsuite errors, numerics-related sage testsuite errors

9.3.beta8> (is system BLAS feeling OK??)

The numerics failures on debian-bullseye have disappeared (or perhaps it is processor dependent, and we were luckier in this run), but show up in debian-sid this time. 

The numerics failures on ubuntu-groovy are still present. Could someone please help by taking a look at these numerics errors on?

Everything else is unchanged in 9.3.rc2 (with some more results in thanks to #31502). 

9.3.beta8> linuxmint-17-{minimal,standard}: Doctest errors from "setlocale" output

9.3.beta8> linuxmint-{18,19,19.3,20.1}-{minimal,standard}: Clean

Unchanged in 9.3.rc2. GH Actions running out of disk space in linuxmint-20.1-standard - despite #31502.

9.3.beta8> fedora-26-minimal, centos-7-minimal, centos-7-i386-minimal: openssl problem

fedora-26-minimal and centos-7-minimal are clean now, thanks to https://trac.sagemath.org/ticket/31439

9.3.beta8> fedora-{26,27,29,30,31,32,33}-standard: Clean

9.3.beta8> fedora-{27,28,29,30,31,32,33}-minimal: Clean

9.3.beta8> fedora-28-standard: 

9.3.beta8> - cvxopt testsuite errors

9.3.beta8> - src/sage/rings/polynomial/groebner_fan.py # 3 doctests failed

9.3.beta8> - src/sage/rings/real_mpfr.pyx # Killed due to abort

fedora-28-standard: Similar but not the same failures in 9.3.rc2

9.3.beta8> fedora-34-standard:

9.3.beta8>   [eclib-20190909]   /usr/bin/ld: /usr/lib64/libntl.so: undefined reference to `std::__exception_ptr::exception_ptr::_M_release()@CXXABI_1.3.13'

fedora-34-standard eclib problem is unchanged in 9.3.rc2, and also issues with ntl show up in singular

9.3.beta8> centos-7-standard: [dochtml] ImportError: /lib64/libz.so.1: version `ZLIB_1.2.9' not found (required by /sage/local/lib/libpng16.so.16)

Unchanged in 9.3.rc2

9.3.beta8> centos-8-{minimal,standard}: Clean

9.3.beta8> gentoo-{minimal,standard}: Clean

9.3.beta8> gentoo-python3.7-{minimal,standard}: Clean

New doctest failures for all of these from linker messages - https://trac.sagemath.org/ticket/31578

9.3.beta8> archlinux-latest-{minimal,standard}: Clean

Unchanged in 9.3.rc2

9.3.beta8> slackware-14.2-{minimal,standard}: Clean

Unchanged in 9.3.rc2

9.3.beta8> conda-forge-minimal: Error building package python3-3.9.1

9.3.beta8> conda-forge-standard: Sage does not accept system python3

Unchanged in 9.3.rc2

9.3.beta8> ubuntu-{bionic,focal}-i386-minimal: [openssl-3.0.0-alpha11] /usr/include/stdio.h:27:10: fatal error: bits/libc-header-start.h: No such file or directory

9.3.beta8> debian-buster-i386-minimal: likewise

9.3.beta8> debian-buster-i386-standard: Clean except for

9.3.beta8> - src/sage/parallel/decorate.py # 1 doctest failed

9.3.beta8> - src/sage/schemes/elliptic_curves/ell_field.py # 1 doctest failed

 

9.3.beta8> homebrew-macos-10.15-minimal-xcode_11.7 (https://github.com/sagemath/sage/runs/2051746258): Clean except for:

9.3.beta8> sage -t --random-seed=0 src/sage/geometry/polyhedron/base.py  # 1 doctest failed (get_memory_usage)

9.3.beta8> sage -t --random-seed=0 src/sage/interfaces/gap.py  # 1 doctest failed

9.3.beta8> (File "src/sage/interfaces/gap.py", line 482, in sage.interfaces.gap.Gap_generic._read_in_file_command)

9.3.beta8> -standard-default likewise

New errors that appear to be related to Singular!

https://github.com/sagemath/sage/runs/2282411164

sage -t --random-seed=0 src/sage/dynamics/arithmetic_dynamics/projective_ds.py  # 12 doctests failed

sage -t --random-seed=0 src/sage/geometry/polyhedron/base.py  # 1 doctest failed

sage -t --random-seed=0 src/sage/interfaces/gap.py  # 1 doctest failed

sage -t --random-seed=0 src/sage/rings/asymptotic/asymptotics_multivariate_generating_functions.py  # 1 doctest failed

sage -t --random-seed=0 src/sage/rings/function_field/function_field.py  # 5 doctests failed

sage -t --random-seed=0 src/sage/rings/morphism.pyx  # 3 doctests failed

sage -t --random-seed=0 src/sage/rings/polynomial/multi_polynomial_element.py  # 1 doctest failed

sage -t --random-seed=0 src/sage/rings/polynomial/multi_polynomial.pyx  # 1 doctest failed

sage -t --random-seed=0 src/sage/rings/qqbar.py  # 40 doctests failed

sage -t --random-seed=0 src/sage/schemes/affine/affine_subscheme.py  # 1 doctest failed

sage -t --random-seed=0 src/sage/schemes/curves/affine_curve.py  # 1 doctest failed

sage -t --random-seed=0 src/sage/schemes/curves/projective_curve.py  # 1 doctest failed

sage -t --random-seed=0 src/sage/schemes/elliptic_curves/constructor.py  # 1 doctest failed

sage -t --random-seed=0 src/sage/schemes/elliptic_curves/ell_curve_isogeny.py  # 1 doctest failed

sage -t --random-seed=0 src/sage/schemes/plane_conics/con_rational_function_field.py  # 3 doctests failed

sage -t --random-seed=0 src/sage/schemes/projective/projective_morphism.py  # 1 doctest failed

sage -t --random-seed=0 src/sage/schemes/projective/projective_subscheme.py  # 2 doctests failed

9.3.beta8> -standard-python3_xcode (https://github.com/sagemath/sage/runs/2051746451) likewise 

9.3.beta8> -minimal-nokegonly-xcode_11.7 (https://github.com/sagemath/sage/runs/2051746494) likewise

9.3.beta8> -minimal-{xcode_12.2,default=xcode_12.4}

9.3.beta8> Lots of additional doctest failures from unexpected linker messages: "dylib ... was built for newer macOS version (10.15) than being linked (10.9)"

These are fixed in 9.3.rc2.

9.3.beta8> [...]--> https://trac.sagemath.org/ticket/31204 / https://trac.sagemath.org/ticket/18272

9.3.beta8> ...-

9.3.beta8> -standard-python3_xcode likewise

9.3.beta8> -minimal-nokegonly-{xcode_12.2,default,xcode_12.4} likewise

9.3.beta8> homebrew-macos-10.15-standard-nokegonly-python3_xcode-{xcode_11.7,xcode_12.2,default,xcode_12.4} (https://github.com/sagemath/sage/runs/2051746530): fpylll build fails (does not find fplll in /usr/local)

macos (conda-forge-macos, minimal, default, macos-10.15): (same on minimal-macos-11.0, and standard-{macos-10.15,macos-11.0))

docbuild error (https://github.com/sagemath/sage/runs/2282411222?check_suite_focus=true)

[categorie]   File "/Users/runner/work/sage/sage/.tox/local-conda-forge-macos-minimal/local/lib/python3.8/site-packages/sage/rings/qqbar.py", line 5236, in _richcmp_

[categorie]     return bool(self) == (op == op_NE)

[categorie]   File "/Users/runner/work/sage/sage/.tox/local-conda-forge-macos-minimal/local/lib/python3.8/site-packages/sage/rings/qqbar.py", line 3900, in __bool__

[categorie]     c = cmp_elements_with_same_minpoly(left, right, left.minpoly())

[categorie]   File "/Users/runner/work/sage/sage/.tox/local-conda-forge-macos-minimal/local/lib/python3.8/site-packages/sage/rings/qqbar.py", line 2720, in cmp_elements_with_same_minpoly

[categorie]     raise RuntimeError('a = {}\nb = {}\np = {}'.format(a, b, p))

[categorie] RuntimeError: a = 2.917983368234029?e15

[categorie] b = 2.917983368234029?e15

[categorie] p = x^6 - 5835966736468056*x^5 + 2658455991569831813923974927379870544*x^4 + 2658455991569831814199401327502424704*x^3 + 2658455991569831809001982851260652032*x^2 + 253512790118400*x + 4096

9.3.beta8> local-macos (homebrew-macos, minimal, default, macos-11.0) and other variants: scipy build fails

Fixed in 9.3.rc2.

9.3.beta8> local-macos (homebrew-macos-python3_pythonorg, standard, default, macos-11.0) - likewise

9.3.beta8> local-macos (homebrew-macos, standard, default, macos-11.0) - mostly clean (times out during make ptest)

local-macos-nohomebrew (macos-10.15, macos-nobootstrap, minimal, default) and other variants: Clean

local-macos-nohomebrew (macos-10.15, macos-nobootstrap-python3_pythonorg, minimal, default) and other variants: Does not accept provided python3

[Justin C. Walker]

Just finished up compiling and testing the three .rc releases for sage 9.3.

I did this on my three macOS systems: 10.13.6 (2017 iMac Pro, 18-core Xeon W), 10.14.6 (2017 MBP, quad-core Core i7), and 10.15.7 (2019 MBP, 8-core Core i9).

> On Apr 6, 2021, at 15:33 , Volker Braun <vbrau...@gmail.com> wrote:
>
> As always, you can get the latest beta version from the "develop" git branch. Alternatively, the self-contained source tarball is at http://www.sagemath.org/download-latest.html

For RC0 and RC1, none of the builds completed, due to a failure for Pynac (the same in all 6 cases: “/usr/local” leakage).

For RC2, all builds completed without problem, and all tests (‘ptestlong’) passed.

Justin

--
Justin C. Walker
Curmudgeon-at-large
--
Network, n., Difference between work
charged for and work done

[Matthias Köppe]

I have opened tickets for some but not all of these issues:

#31624
fedora-34-standard: eclib, singular build failures with system ntl

#31623
homebrew-macos-10.15-standard-default: Doctest errors related to singular

#31622
conda-forge-macos: RuntimeError from qqbar cmp_elements_with_same_minpoly

#31621
ubuntu-groovy-standard: numerics-related sage testsuite errors

Help is needed...

[axio...@yahoo.de]

Unfortunately, one of the recent changes breaks the emacs interface sage-shell-mode, see https://trac.sagemath.org/ticket/31630.

I think I recall that some display change was announced, I'll try to find out when it happened.

Martin

[axio...@yahoo.de]

Indeed, SageMath version 9.3.rc1, Release Date: 2021-03-29 works just fine.

[axio...@yahoo.de]

I'm very sorry, it seems that the problem was elsewhere.  Rebuilding solved the issue.

[Emmanuel Charpentier]

On Debian testing, upgrading from 9.3.rc0 to 9.3.rc2 fails at documentation building. From install.log :

make[3]: rien à faire pour « all-sage ».
[dochtml] Traceback (most recent call last):
[dochtml]   File "/usr/lib/python3.9/runpy.py", line 188, in _run_module_as_main
[dochtml]     mod_name, mod_spec, code = _get_module_details(mod_name, _Error)
[dochtml]   File "/usr/lib/python3.9/runpy.py", line 147, in _get_module_details
[dochtml]     return _get_module_details(pkg_main_name, error)
[dochtml]   File "/usr/lib/python3.9/runpy.py", line 111, in _get_module_details
[dochtml]     __import__(pkg_name)
[dochtml]   File "/usr/local/sage-9/local/lib/python3.9/site-packages/sage_docbuild/__init__.py", line 57, in <module>
[dochtml]     import sage.all
[dochtml]   File "/usr/local/sage-9/local/lib/python3.9/site-packages/sage/all.py", line 142, in <module>
[dochtml]     from sage.rings.all      import *
[dochtml]   File "/usr/local/sage-9/local/lib/python3.9/site-packages/sage/rings/all.py", line 83, in <module>
[dochtml]     from .polynomial.all import *
[dochtml]   File "/usr/local/sage-9/local/lib/python3.9/site-packages/sage/rings/polynomial/all.py", line 42, in <module>
[dochtml]     from sage.rings.polynomial.laurent_polynomial_ring import LaurentPolynomialRing
[dochtml]   File "/usr/local/sage-9/local/lib/python3.9/site-packages/sage/rings/polynomial/laurent_polynomial_ring.py", line 50, in <module>
[dochtml]     from sage.rings.polynomial.laurent_polynomial import LaurentPolynomial_mpair, LaurentPolynomial_univariate
[dochtml]   File "sage/rings/polynomial/laurent_polynomial.pyx", line 1, in init sage.rings.polynomial.laurent_polynomial (build/cythonized/sage/rings/polynomial/laurent_polynomial.c:40841)
[dochtml]   File "sage/matrix/matrix0.pyx", line 28, in init sage.matrix.matrix0 (build/cythonized/sage/matrix/matrix0.c:44271)
[dochtml]   File "/usr/local/sage-9/local/lib/python3.9/site-packages/sage/modules/free_module.py", line 167, in <module>
[dochtml]     import sage.matrix.matrix_space
[dochtml]   File "/usr/local/sage-9/local/lib/python3.9/site-packages/sage/matrix/matrix_space.py", line 55, in <module>
[dochtml]     from . import matrix_mpolynomial_dense
[dochtml]   File "sage/matrix/matrix_mpolynomial_dense.pyx", line 1, in init sage.matrix.matrix_mpolynomial_dense (build/cythonized/sage/matrix/matrix_mpolynomial_dense.cpp:8652)
[dochtml] ImportError: libSingular-4.2.0p1+2021-03-13+sage.so: cannot open shared object file: No such file or directory
make[3]: *** [Makefile:2280 : doc-html] Erreur 1
make[2]: *** [Makefile:2160 : all-start] Erreur 2
make[2] : on quitte le répertoire « /usr/local/sage-9/build/make »

However :

charpent@zen-book-flip:/usr/local/sage-9$ ls -l local/lib/libSingular*
-rwxr-xr-x 1 charpent charpent 4618600 10 avril 21:18 local/lib/libSingular-4.2.0p1+2021-03-24+sage-2.so
lrwxrwxrwx 1 charpent charpent      40 10 avril 21:18 local/lib/libSingular.so -> libSingular-4.2.0p1+2021-03-24+sage-2.so

It seems that something (what ?) is not updated in the build process scripts and still points to the Singular version used in 9.3.rc0. I suppose that rebuildig would fix the problem, but is  a bit of a pain in the *ss...

Suggestions ?

​

[Matthias Köppe]

On Saturday, April 10, 2021 at 1:23:53 PM UTC-7 emanuel.c...@gmail.com wrote:

It seems that something (what ?) is not updated in the build process scripts and still points to the Singular version used in 9.3.rc0. I suppose that rebuildig would fix the problem [...]

It suffices to rebuild sagelib. Use "make sagelib-clean" and then proceed with a normal build.

 

[Emmanuel Charpentier]

Thanks. This seems to have done it.

One more question ; how should I have divined this ? I can't remember reading anything about this in the build instructions nor in the developer's guide...

[Matthias Köppe]

It's an annoying shortcoming of our build system. This particular problem is explained in https://trac.sagemath.org/ticket/29711, on which you are already cc'd. Overall I agree that possible remedies for failures in incremental builds are not sufficiently explained anywhere in our documentation. 

[Emmanuel Charpentier]

FWIW, on Debian testing, once solved the above problem of upgrading from `9.3.rc0` to 9.3.rc2`, thanks to Matthias, ptestlong gave exactly the same results as for the last two previous releases : one timeout (parigp.py), one cosmetic bug (gap_packages.py), one real bug (division by zero in abelian_group_gap.py).

HTH,

​