Sage 10.2.rc3 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

fa5c9395d3e (tag: 10.2.rc3, github/develop) Updated SageMath version to 10.2.rc3
705a0ee9bf3 gh-36711: CI macOS: Build following stages even when some job failed; adjust platforms
7f6a2a7e6a8 gh-36708: CI Linux: Fixes for tests of optional packages
dccb4127165 gh-36707: `build/pkgs/nauty`: Handle `SAGE_FAT_BINARY`
82a3102d058 gh-36699: CI Linux: Follow-up fix for conda-forge-python3.11
4d2dd497f5f gh-36696: Fix SSL timeout in doctest / internet feature
613775f17bd gh-36692: Make doc-pdf separate from doc-html
606c816a9d9 gh-36670: .github/workflows/docker.yml: Interrupt the build before the 6 hour cancellation
9953321a2d2 gh-36660: .github/workflows/ci-linux.yml: Fine-tune max-parallel
4430e4cc37c (tag: 10.2.rc2) Updated SageMath version to 10.2.rc2

[John Cremona]

Is it expected that doing "git pull upstream develop" on top of the rc1 commit should be a merge?

--
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/8593234a-b974-40f0-be55-8ec14c975231n%40googlegroups.com.

[Eric Gourgoulhon]

On Ubuntu 22.04 running on Xeon E5-2623 + 16 GB RAM:

- incremental build (-j16) from 10.2.rc2, with system python (3.10.12): OK
- make ptestlong --> "All tests passed!"

Eric.

[Matthias Köppe]

You may have to force-pull the release tags:

git fetch --tags -f upstream 

[John Cremona]

I am seeing failures building from rc3 (commit fa5c9395d3).  To make sure I made a fresh git clone of this and did nothing more than

./bootstrap
./configure
 make -j16

On thre machines, make build worked OK, but not make: it is sagemath_doc_html-none which fails.  The short logfile is attached.  It's a problem with libntl linking: it wants libntl.so.44 .   In configure, Sage did decide that it was OK to use a system NTL.  I have two (which may be  the problem):  in /usr and in /usrlocal.  the relevant library files are /usr/lib/x86_64-linux-gnu/libntl.so -> libntl.so.43.0.1 and /usr/local/lib/libntl.so -> libntl.so.44.0.1.   From the error message in the attached log it could not find libntl.so.44 but it exists in /usr/local/lib and has the correct permissions.

On two of the three machines I ran "make" again and that resulted in crashing python -- you can see the traceback file if you want, it ends with 

No symbol table info available.
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
[Inferior 1 (process 3558942) detached]
27      ../sysdeps/unix/sysv/linux/wait4.c: No such file or directory.
Traceback (most recent call last):
  File "<string>", line 25, in <module>
ModuleNotFoundError: No module named 'Cython'
Error while executing Python code.
Saved trace to /home/jec/.sage/crash_logs/crash_y7tokr_q.log
------------------------------------------------------------------------
Unhandled SIGABRT: An abort() occurred.
This probably occurred because a *compiled* module has a bug
in it and is not properly wrapped with sig_on(), sig_off().
Python will now terminate.
------------------------------------------------------------------------
Aborted (core dumped)
make[6]: *** [Makefile:28: doc-inventory--reference-combinat] Error 134
make[5]: *** [Makefile:45: doc-inventory-reference] Error 2

To view this discussion on the web visit https://groups.google.com/d/msgid/sage-release/db4ffd07-b437-4d4c-bf79-3f4f7928a084n%40googlegroups.com.

[Dima Pasechnik]

On Wed, Nov 15, 2023 at 5:12 PM John Cremona <john.c...@gmail.com> wrote:
>
> I am seeing failures building from rc3 (commit fa5c9395d3).  To make sure I made a fresh git clone of this and did nothing more than
>
> ./bootstrap
> ./configure
>  make -j16
>
> On thre machines, make build worked OK, but not make: it is sagemath_doc_html-none which fails.  The short logfile is attached.  It's a problem with libntl linking: it wants libntl.so.44 .   In configure, Sage did decide that it was OK to use a system NTL.  I have two (which may be  the problem):  in /usr and in /usrlocal.  the relevant library files are /usr/lib/x86_64-linux-gnu/libntl.so -> libntl.so.43.0.1 and /usr/local/lib/libntl.so -> libntl.so.44.0.1.   From the error message in the attached log it could not find libntl.so.44 but it exists in /usr/local/lib and has the correct permissions.

This setup is asking for trouble. We don't have a fine-tune way to deal with such cases.

Consider this a bug. 

Basically,  ./configure sees libntl available and good enough, but goes with whatever default ordering of libraries' locations it has.

Then the linker, possibly using another ordering, tries to resolve -lntl...

> To view this discussion on the web visit https://groups.google.com/d/msgid/sage-release/CAD0p0K6587td__1MvJFHv3a9GH9hJLfR3jvg63NSYMhDyBKBLA%40mail.gmail.com.

[John Cremona]

On Wed, 15 Nov 2023, 17:43 Dima Pasechnik, <dim...@gmail.com> wrote:

On Wed, Nov 15, 2023 at 5:12 PM John Cremona <john.c...@gmail.com> wrote:
>
> I am seeing failures building from rc3 (commit fa5c9395d3).  To make sure I made a fresh git clone of this and did nothing more than
>
> ./bootstrap
> ./configure
>  make -j16
>
> On thre machines, make build worked OK, but not make: it is sagemath_doc_html-none which fails.  The short logfile is attached.  It's a problem with libntl linking: it wants libntl.so.44 .   In configure, Sage did decide that it was OK to use a system NTL.  I have two (which may be  the problem):  in /usr and in /usrlocal.  the relevant library files are /usr/lib/x86_64-linux-gnu/libntl.so -> libntl.so.43.0.1 and /usr/local/lib/libntl.so -> libntl.so.44.0.1.   From the error message in the attached log it could not find libntl.so.44 but it exists in /usr/local/lib and has the correct permissions.

This setup is asking for trouble. We don't have a fine-tune way to deal with such cases.

Consider this a bug. 

Basically,  ./configure sees libntl available and good enough, but goes with whatever default ordering of libraries' locations it has.

Then the linker, possibly using another ordering, tries to resolve -lntl...

I know, I cannot expect magic.

 As far as I can see, the GCC compiler always looks in /usr/local/include before /usr/include, and you cannot change that, while the linker looks in /usr/lib before /usr/local/lib; the order there is set in /etc/ld.so.conf so could be changed perhaps, though that would breaking something else.

I'll just uninstall or hide one of the ntls.

It is strange though that Sage's configure can clearly find a working library, while one step in building documentation also needs it (why?) and fails to do so despite it being known to ld.

John

To view this discussion on the web visit https://groups.google.com/d/msgid/sage-release/CAAWYfq22m0tmJFeqN4s3MTCoBuThBnAavFYbTbMyYx_Da4geJA%40mail.gmail.com.

[Matthias Köppe]

On Tuesday, November 14, 2023 at 3:13:46 PM UTC-8 Volker Braun wrote:

fa5c9395d3e (tag: 10.2.rc3, github/develop) Updated SageMath version to 10.2.rc3

Thanks, Volker, for the latest rc. I think we're in pretty good shape now, but I think we need one more round for the following:

- https://github.com/sagemath/sage/pull/36722 (openblas upgrade, appears to fix build failures on archlinux-latest and fedora-38; needs review)

I would also like to get the following PR in:

- https://github.com/sagemath/sage/pull/36031 (LatticePolytope.normal_form fixes; needs review)

[Dima Pasechnik]

This package produces incorrect maths results, it has to be demoted to experimental, IMHO.

[Matthias Köppe]

Which package, Dima?

[Dima Pasechnik]

On Thu, 16 Nov 2023, 01:40 Matthias Köppe, <matthia...@gmail.com> wrote:

Which package, Dima?

sorry, wrong thread

--
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/1136a803-7a38-433f-b9ea-b7d57a939b42n%40googlegroups.com.