Release 1.6.1 without release candidate?

24 views
Skip to first unread message

Oscar Benjamin

unread,
Jun 26, 2020, 8:14:50 PM6/26/20
to sympy
Hi all,

I'm preparing the SymPy 1.6.1 bugfix release and I'm considering
skipping the release candidate and releasing directly and just wanted
to see if anyone has any opinions on that. In the past I haven't seen
much evidence that anyone has actually tested the pre-releases
(release candidate or beta) and it is a bunch of work producing them
so I'm asking here to see if anyone appreciates having them.

You can see the fixes included for 1.6.1 here:
https://github.com/sympy/sympy/wiki/Release-Notes-for-1.6.1

The big change from 1.6 is the partial revert of some changes in
matrix operations. SymPy 1.6 introduced automatic simplification of
some calculations as part of matrix operations which gives a big speed
up in some simple cases but also a big slowdown in some other cases
for matrices with very complicated entries. There is a longer term
plan that can (hopefully) get the best of both worlds and speed up the
simple cases without affecting the others but that won't be ready for
some time. This change introduced in 1.6 is to be reverted for 1.6.1.

There are also some other smaller bugfixes to be included in the 1.6.1 release.

Does anyone have any strong desire for a release candidate? Otherwise
I'll release 1.6.1 ASAP.

Oscar

Aaron Meurer

unread,
Jun 26, 2020, 9:15:45 PM6/26/20
to sympy
My feedback here isn't as important as users of SymPy who would be
testers of a release candidate, but I would say that as long as we're
willing to make bugfix releases like 1.6.1 for major regressions
identified from the major release, it is less important to do release
candidates.

I also would like to see evidence of people using release candidates.
If you have tested a recent SymPy release candidate, even if you
didn't find any issues with it, please let us know here. We can also
take a look at the download stats for the release candidates.

Aaron Meurer
> --
> You received this message because you are subscribed to the Google Groups "sympy" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to sympy+un...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/sympy/CAHVvXxTgk2WUo6w3qEbty_F_3yEqs-MXMx%3DfjGM8AErwbJCHpw%40mail.gmail.com.

David Bailey

unread,
Jun 27, 2020, 10:19:03 AM6/27/20
to sy...@googlegroups.com
On 27/06/2020 01:15, Oscar Benjamin wrote:
Hi all,

I'm preparing the SymPy 1.6.1 bugfix release and I'm considering
skipping the release candidate and releasing directly and just wanted
to see if anyone has any opinions on that. In the past I haven't seen
much evidence that anyone has actually tested the pre-releases
(release candidate or beta) and it is a bunch of work producing them
so I'm asking here to see if anyone appreciates having them.

As a user, I agree. If you identify something as a "release candidate" or "beta" people tend to skip it and wait. This is because a new release of any complex software that is used to build other systems takes time and effort to validate, and people wait for stable code. I'd suggest that instead, perhaps you should informally update those who have reported problems, to see if they are satisfied and then put out a general release.

As a last resort it is easy for a user to revert to a previous version of SymPy.

David

Reply all
Reply to author
Forward
0 new messages