Sage 9.5.beta7 ?

59 views
Skip to first unread message

Emmanuel Charpentier

unread,
Nov 19, 2021, 12:32:27 PM11/19/21
to sage-release

A recent refresh of my Sahge tree gave me this :

charpent@zen-book-flip:/usr/local/sage-9$ sage --version
SageMath version 9.5.beta7, Release Date: 2021-11-18

But as far as I know, this version has been neither released nor announced…

What is this ?

William Stein

unread,
Nov 19, 2021, 12:47:30 PM11/19/21
to sage-release
I think the develop branch is regularly changing with new commits each
day (hopefully) as Volker merges positively reviewed tickets. You can
browse pages and pages of commits that came after 9.5.beta6 to the
develop branch here:

https://github.com/sagemath/sage/commits/develop

It looks like he committed a new 9.5.beta7 version 18 hours ago, but
just hasn't sent out an email yet. Presumably he always does a tagged
commit like this *before* he sends out an email, since it would be
annoying if he did that after sending out the announcement email.

William
> --
> 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/049c7011-2101-4b34-9874-9c3685e68939n%40googlegroups.com.



--
William (http://wstein.org)

Emmanuel Charpentier

unread,
Nov 19, 2021, 1:39:25 PM11/19/21
to sage-release

FWIW, here is what I observe after this (possibly premature) upgrade :

On Debian testing running on core i7 + 16 GB RAM, upgrading 9.5.beta6(configured to use all available system packages) to 9.5.beta7 andrunning ptestlong results in a (transient) timeout and five permanent failures :

Doctest Result P/T
src/sage/rings/polynomial/polynomial_zmod_flint.pyx Timed out T
src/sage/interfaces/fricas.py 1 doctest failed P
src/sage/rings/complex_arb.pyx 3 doctests failed P
src/sage/functions/gamma.py 1 doctest failed P
src/sage/symbolic/function.pyx 1 doctest failed P
src/sage/rings/real_arb.pyx 12 doctests failed P

The same failures (one format difference in fricas an a bunch of small numerical differences) have been reported for the upgrade of this machine to 9.5.beta5, but the numerical differences were not observed when upgrading another Debian testing to 9.5.beta5.

HTH,

Samuel Lelièvre

unread,
Nov 19, 2021, 4:47:22 PM11/19/21
to sage-release
Hi,

Regarding the release process, my understanding is Volker works
on a release branch in a different repo; once we see Sage 9.5.beta7 at


it can be considered as released. It sometimes takes the announcement
a few hours or a day to get posted by Volker after that; that step is not
automated.

After fetching that branch (with tags), one gets the list of merged tickets
(such as posted in Volker's announcement emails) from an appropriate
`git log` invocation:

    $ git fetch origin develop --tags -q
    $ git log --oneline --first-parent 9.5.beta6^...9.5.beta7

Building on two machines I observed the same failures
as reported by Emmanuel Charpentier, except the timeout
in `src/sage/rings/polynomial/polynomial_zmod_flint.pyx`.

I also got a few other failures, see details below along with
references to tickets fixing them.

On Debian 10 buster with FriCAS installed, I get the same
FriCAS doctest failure, which is already tracked at #32886.

On an Early 2014 13-inch MacBook Air, in macOS 10.14.6,
with many Homebrew packages, here is how things went.

No trouble in these preliminary steps:

```
$ git switch develop
$ git pull origin develop --tags -q
$ git branch -vv
* develop   cc60cfebc4 Updated SageMath version to 9.5.beta7
$ export MAKE='make -j5'
$ source .homebrew-build-env
$ ./bootstrap -q
```

Running configure gives an import error regarding lzma:

```
$ ./configure -q
Traceback (most recent call last):
  File "<string>", line 1, in <module>
ImportError: No module named lzma
```

This might have to do with

- #30948: build/pkgs: Split out spkg liblzma from xz

and maybe `build/pkgs/lzma/distro/homebrew.txt` needs adjusting.

Sage builds fine but there are doctest failures in 7 files:

```
$ make -s V=0 ptestlong
```

Random seed 264523619537841110147717491410595481298
was used and doctests failed in the following files:

- (a) src/sage/env.py  # 1 doctest failed
- (b) src/sage/lfunctions/pari.py  # 1 doctest failed
- (c) src/sage/functions/gamma.py  # 1 doctest failed
- (d) src/sage/rings/complex_arb.pyx  # 3 doctests failed
- (e) src/sage/rings/real_arb.pyx  # 12 doctests failed
- (f) src/sage/symbolic/function.pyx  # 1 doctest failed
- (g) src/sage/interfaces/expect.py  # 2 doctests failed

```
Total time for all tests: 11732.3 seconds
    cpu time: 25929.6 seconds
    cumulative wall time: 45148.8 seconds
Pytest is not installed, skip checking tests that rely on it.
make: *** [ptestlong] Error 1
```

All failures but (g) were reproduced with random seed 0.

- (a) was already reported at ticket:29024#comment:168,
      and has a proposed fix at #32880 (needs review)
- (b) already has a positively reviewed fix at #32797
- for (c-d-e-f): see proposed fix at #32905 (needs review)

Here is failure (g), in case it rings a bell for someone.

```
sage -t --long --random-seed=264523619537841110147717491410595481298 src/sage/interfaces/expect.py
**********************************************************************
File "src/sage/interfaces/expect.py", line 915, in sage.interfaces.expect.Expect._eval_line
Failed example:
    singular.interrupt()
Expected:
    True
Got:
    False
**********************************************************************
File "src/sage/interfaces/expect.py", line 921, in sage.interfaces.expect.Expect._eval_line
Failed example:
    singular('2+3')
Expected:
    Singular crashed -- automatically restarting.
    5
Got:
    5
**********************************************************************
1 item had failures:
   2 of  16 in sage.interfaces.expect.Expect._eval_line
    [102 tests, 2 failures, 15.77 s]
```

Hopefully the tickets mentioned above make it into Sage 9.5.beta8.
And maybe some more -- currently there are 23 positively reviewed
tickets with milestone sage-9.5:


and many more tickets with that milestone are at various
other stages of completion.  Lots to do!   --Samuel

Volker Braun

unread,
Nov 20, 2021, 5:28:18 AM11/20/21
to sage-release
I'm often starting the release/upload script before going to sleep at night, in those cases you'll see the announcement the next day or so. No worries though, if you see the code then its released!
Reply all
Reply to author
Forward
0 new messages