numpy 2.0.x has been branched.
Hi All, Numpy 2.0.x has been branched, further work on the release notes should be made against that branch. The main branch is now 2.1.0. Chuck
Hi Chuck, Is there a rough expected landing time for NumPy 2.0? Also does the branching mean that the intention/guarantee is not to make further incompatible changes? SymPy's current master branch is compatible with the current NumPy master branch but SymPy's last release is not and will not be compatible with NumPy 2.0. That means that we want to have a SymPy release before NumPy 2.0 that includes the changes needed in preparation. Ideally I also want that SymPy release to be made after NumPy has finished making breaking changes though. Basically I want to know what the interval of time is between NumPy finishing its incompatible changes and NumPy 2.0 actually being released (to the extent that that is knowable). Oscar On Fri, 8 Mar 2024 at 21:23, Charles R Harris <charlesr.harris@gmail.com> wrote:
Hi All,
Numpy 2.0.x has been branched, further work on the release notes should be made against that branch. The main branch is now 2.1.0.
Chuck _______________________________________________ NumPy-Discussion mailing list -- numpy-discussion@python.org To unsubscribe send an email to numpy-discussion-leave@python.org https://mail.python.org/mailman3/lists/numpy-discussion.python.org/ Member address: oscar.j.benjamin@gmail.com
On Fri, Mar 8, 2024 at 2:48 PM Oscar Benjamin <oscar.j.benjamin@gmail.com> wrote:
Hi Chuck,
Is there a rough expected landing time for NumPy 2.0?
Also does the branching mean that the intention/guarantee is not to make further incompatible changes?
SymPy's current master branch is compatible with the current NumPy master branch but SymPy's last release is not and will not be compatible with NumPy 2.0. That means that we want to have a SymPy release before NumPy 2.0 that includes the changes needed in preparation. Ideally I also want that SymPy release to be made after NumPy has finished making breaking changes though.
Basically I want to know what the interval of time is between NumPy finishing its incompatible changes and NumPy 2.0 actually being released (to the extent that that is knowable).
Oscar
We plan on a beta release, which will probably have a stable API, but not guaranteed, then an rc1. At that point the API should be stable and the final release will wait for the most important downstream projects to make compatible releases. I would hope we are talking about a month in total, but it will depend on how things go. Chuck
On Fri, 8 Mar 2024 at 22:27, Charles R Harris <charlesr.harris@gmail.com> wrote:
We plan on a beta release, which will probably have a stable API, but not guaranteed, then an rc1. At that point the API should be stable and the final release will wait for the most important downstream projects to make compatible releases. I would hope we are talking about a month in total, but it will depend on how things go.
I can understand if it is difficult to answer this but I just want a best estimate answer: You say that it is a month in total but when is the start and end of that month? Are we talking next week or 3 months from now or something else? -- Oscar
How bad are the NumPy 2.0 breakages in SymPy? We could do a backport release if they are serious. Aaron Meurer On Fri, Mar 8, 2024 at 2:46 PM Oscar Benjamin <oscar.j.benjamin@gmail.com> wrote:
Hi Chuck,
Is there a rough expected landing time for NumPy 2.0?
Also does the branching mean that the intention/guarantee is not to make further incompatible changes?
SymPy's current master branch is compatible with the current NumPy master branch but SymPy's last release is not and will not be compatible with NumPy 2.0. That means that we want to have a SymPy release before NumPy 2.0 that includes the changes needed in preparation. Ideally I also want that SymPy release to be made after NumPy has finished making breaking changes though.
Basically I want to know what the interval of time is between NumPy finishing its incompatible changes and NumPy 2.0 actually being released (to the extent that that is knowable).
Oscar
On Fri, 8 Mar 2024 at 21:23, Charles R Harris <charlesr.harris@gmail.com> wrote:
Hi All,
Numpy 2.0.x has been branched, further work on the release notes should be made against that branch. The main branch is now 2.1.0.
Chuck _______________________________________________ NumPy-Discussion mailing list -- numpy-discussion@python.org To unsubscribe send an email to numpy-discussion-leave@python.org https://mail.python.org/mailman3/lists/numpy-discussion.python.org/ Member address: oscar.j.benjamin@gmail.com
_______________________________________________ NumPy-Discussion mailing list -- numpy-discussion@python.org To unsubscribe send an email to numpy-discussion-leave@python.org https://mail.python.org/mailman3/lists/numpy-discussion.python.org/ Member address: asmeurer@gmail.com
On Fri, 8 Mar 2024 at 23:31, Aaron Meurer <asmeurer@gmail.com> wrote:
How bad are the NumPy 2.0 breakages in SymPy? We could do a backport release if they are serious.
I don't remember exactly what needed changing but some basic things in sympy needed changing because of basic changes in numpy. I don't see what purpose a backport release would serve. I would like to get out a sympy release right now. I would also like to know when is the right time to put out a release that is tested with all new numpy changes and that should be expected to be compatible with numpy 2.0. I am asking for a timeline: is numpy 2.0 coming out in 1 week, 1 month, 2 months, 3 months, 6 months, 12 months? I understand if it is difficult to be precise but right now I have no idea what timeline anyone is talking about and I am pretty sure that numpy folks have a much clearer idea: please share it with me! -- Oscar
On Fri, Mar 8, 2024 at 4:45 PM Oscar Benjamin <oscar.j.benjamin@gmail.com> wrote:
On Fri, 8 Mar 2024 at 23:31, Aaron Meurer <asmeurer@gmail.com> wrote:
How bad are the NumPy 2.0 breakages in SymPy? We could do a backport release if they are serious.
I don't remember exactly what needed changing but some basic things in sympy needed changing because of basic changes in numpy.
I don't see what purpose a backport release would serve. I would like to get out a sympy release right now. I would also like to know when is the right time to put out a release that is tested with all new numpy changes and that should be expected to be compatible with numpy 2.0.
I am asking for a timeline: is numpy 2.0 coming out in 1 week, 1 month, 2 months, 3 months, 6 months, 12 months?
I understand if it is difficult to be precise but right now I have no idea what timeline anyone is talking about and I am pretty sure that numpy folks have a much clearer idea: please share it with me!
About a month from now. Chuck
On Sat, 9 Mar 2024 at 00:44, Charles R Harris <charlesr.harris@gmail.com> wrote:
About a month from now.
What will happen about a month from now? It might seem obvious to you but I can interpret this in different ways. To be clear numpy 2.0 is expected to be released in full to the public in about one month's time from today? -- Oscar
On Sat, Mar 9, 2024 at 2:03 AM Oscar Benjamin <oscar.j.benjamin@gmail.com> wrote:
On Sat, 9 Mar 2024 at 00:44, Charles R Harris <charlesr.harris@gmail.com> wrote:
About a month from now.
What will happen about a month from now? It might seem obvious to you but I can interpret this in different ways.
To be clear numpy 2.0 is expected to be released in full to the public in about one month's time from today?
Let me give the optimistic and pessimistic timelines. Optimistic: - 2.0.0b1 later today - 2.0.0rc1 (ABI stable) in 7-10 days - 2.0.0 final release in 1 month Pessimistic: - 2.0.0b1 within a few days - 2.0.0rc1 (ABI stable) in 2 weeks - 2.0.0rc2 in 4 weeks - 2.0.0rc3 in 6 weeks - 2.0.0 final release in 8 weeks For projects which have nontrivial usage of the NumPy API (and especially if they also use the C API), I'd recommend: 1. Check whether things work with 2.0.0b1, ideally asap so if there is anything we missed we can catch it before rc1. Perhaps do a pre-release of your own package 2. Do a final release after 2.0.0rc1 - ideally as soon as possible after, and definitely before the final 2.0.0 release For (2), note that there are a ton of packages that do not have correct upper bounds, so if you haven't done your own new release that is compatible with both 2.0.0 and 1.26.x *before* 2.0.0 comes out, the users of your project are likely to have a hard time. Cheers, Ralf
Maybe someone composes another preemptive blog post about this? I think this needs some banging on the pans to soften its landing since NumPy 2.0 is a big deal. The community will take care of its circulation in typical places. We always get the grilling no matter the outcome :) But I think it is important that at least package maintainers see this and react to it. On Sat, Mar 9, 2024 at 11:12 AM Ralf Gommers <ralf.gommers@gmail.com> wrote:
On Sat, Mar 9, 2024 at 2:03 AM Oscar Benjamin <oscar.j.benjamin@gmail.com> wrote:
On Sat, 9 Mar 2024 at 00:44, Charles R Harris <charlesr.harris@gmail.com> wrote:
About a month from now.
What will happen about a month from now? It might seem obvious to you but I can interpret this in different ways.
To be clear numpy 2.0 is expected to be released in full to the public in about one month's time from today?
Let me give the optimistic and pessimistic timelines. Optimistic:
- 2.0.0b1 later today - 2.0.0rc1 (ABI stable) in 7-10 days - 2.0.0 final release in 1 month
Pessimistic:
- 2.0.0b1 within a few days - 2.0.0rc1 (ABI stable) in 2 weeks - 2.0.0rc2 in 4 weeks - 2.0.0rc3 in 6 weeks - 2.0.0 final release in 8 weeks
For projects which have nontrivial usage of the NumPy API (and especially if they also use the C API), I'd recommend: 1. Check whether things work with 2.0.0b1, ideally asap so if there is anything we missed we can catch it before rc1. Perhaps do a pre-release of your own package 2. Do a final release after 2.0.0rc1 - ideally as soon as possible after, and definitely before the final 2.0.0 release
For (2), note that there are a ton of packages that do not have correct upper bounds, so if you haven't done your own new release that is compatible with both 2.0.0 and 1.26.x *before* 2.0.0 comes out, the users of your project are likely to have a hard time.
Cheers, Ralf
_______________________________________________ NumPy-Discussion mailing list -- numpy-discussion@python.org To unsubscribe send an email to numpy-discussion-leave@python.org https://mail.python.org/mailman3/lists/numpy-discussion.python.org/ Member address: ilhanpolat@gmail.com
On Sat, 9 Mar 2024 at 10:16, Ralf Gommers <ralf.gommers@gmail.com> wrote:
On Sat, Mar 9, 2024 at 2:03 AM Oscar Benjamin <oscar.j.benjamin@gmail.com> wrote:
On Sat, 9 Mar 2024 at 00:44, Charles R Harris <charlesr.harris@gmail.com> wrote:
About a month from now.
What will happen about a month from now? It might seem obvious to you but I can interpret this in different ways.
To be clear numpy 2.0 is expected to be released in full to the public in about one month's time from today?
Let me give the optimistic and pessimistic timelines. Optimistic:
- 2.0.0b1 later today - 2.0.0rc1 (ABI stable) in 7-10 days - 2.0.0 final release in 1 month
Pessimistic:
- 2.0.0b1 within a few days - 2.0.0rc1 (ABI stable) in 2 weeks - 2.0.0rc2 in 4 weeks - 2.0.0rc3 in 6 weeks - 2.0.0 final release in 8 weeks
Thanks Ralf and Chuck. Sorry, I meant to reply to this earlier but got distracted with other things. We are now 16 days into the future so NumPy 2.0 would be 2 weeks time for the optimistic timescale. I assume that now the beta release is out the intention is no further breaking changes. Then if SymPy master is currently compatible with NumPy 2.0.0b1 a good time for a SymPy release with accumulated fixes is... ASAP! Presumably now that NumPy 2.0 is branched downstream projects should test in CI against the prereleases (pip install --pre numpy) rather than the nightly wheels. (SymPy does not usually test against the nightly wheels. I added that for NumPy 2.0 but maybe we should keep it...) -- Oscar
On Mon, Mar 25, 2024 at 5:54 PM Oscar Benjamin <oscar.j.benjamin@gmail.com> wrote:
On Sat, 9 Mar 2024 at 10:16, Ralf Gommers <ralf.gommers@gmail.com> wrote:
On Sat, Mar 9, 2024 at 2:03 AM Oscar Benjamin <
oscar.j.benjamin@gmail.com> wrote:
On Sat, 9 Mar 2024 at 00:44, Charles R Harris <
charlesr.harris@gmail.com> wrote:
About a month from now.
What will happen about a month from now? It might seem obvious to you but I can interpret this in different ways.
To be clear numpy 2.0 is expected to be released in full to the public in about one month's time from today?
Let me give the optimistic and pessimistic timelines. Optimistic:
- 2.0.0b1 later today - 2.0.0rc1 (ABI stable) in 7-10 days - 2.0.0 final release in 1 month
Pessimistic:
- 2.0.0b1 within a few days - 2.0.0rc1 (ABI stable) in 2 weeks - 2.0.0rc2 in 4 weeks - 2.0.0rc3 in 6 weeks - 2.0.0 final release in 8 weeks
Thanks Ralf and Chuck. Sorry, I meant to reply to this earlier but got distracted with other things.
We are now 16 days into the future so NumPy 2.0 would be 2 weeks time for the optimistic timescale.
I assume that now the beta release is out the intention is no further breaking changes. Then if SymPy master is currently compatible with NumPy 2.0.0b1 a good time for a SymPy release with accumulated fixes is... ASAP!
Presumably now that NumPy 2.0 is branched downstream projects should test in CI against the prereleases (pip install --pre numpy) rather than the nightly wheels. (SymPy does not usually test against the nightly wheels. I added that for NumPy 2.0 but maybe we should keep it...)
The rc1 release is now waiting on pybind11, so there is an (uncertain) delay. Chuck
It looks like the pybind11 release is now done (https://github.com/pybind/pybind11/releases/tag/v2.12.0)? Any more blockers? (We're eagerly awaiting -rc1 so we can release new wheels for our project...)
On Fri, Mar 29, 2024 at 2:07 PM Peter Hawkins <peter.hawkins@gmail.com> wrote:
It looks like the pybind11 release is now done ( https://github.com/pybind/pybind11/releases/tag/v2.12.0)? Any more blockers?
No more blockers - CI is running on the last backport that we need I believe, so it's very close. Hours to max a few days away I think.
(We're eagerly awaiting -rc1 so we can release new wheels for our project...)
Great! Ralf
On Tue, Mar 26, 2024 at 1:20 AM Charles R Harris <charlesr.harris@gmail.com> wrote:
On Mon, Mar 25, 2024 at 5:54 PM Oscar Benjamin <oscar.j.benjamin@gmail.com> wrote:
On Sat, 9 Mar 2024 at 10:16, Ralf Gommers <ralf.gommers@gmail.com> wrote:
On Sat, Mar 9, 2024 at 2:03 AM Oscar Benjamin <
oscar.j.benjamin@gmail.com> wrote:
On Sat, 9 Mar 2024 at 00:44, Charles R Harris <
charlesr.harris@gmail.com> wrote:
About a month from now.
What will happen about a month from now? It might seem obvious to you but I can interpret this in different ways.
To be clear numpy 2.0 is expected to be released in full to the public in about one month's time from today?
Let me give the optimistic and pessimistic timelines. Optimistic:
- 2.0.0b1 later today - 2.0.0rc1 (ABI stable) in 7-10 days - 2.0.0 final release in 1 month
Pessimistic:
- 2.0.0b1 within a few days - 2.0.0rc1 (ABI stable) in 2 weeks - 2.0.0rc2 in 4 weeks - 2.0.0rc3 in 6 weeks - 2.0.0 final release in 8 weeks
Thanks Ralf and Chuck. Sorry, I meant to reply to this earlier but got distracted with other things.
We are now 16 days into the future so NumPy 2.0 would be 2 weeks time for the optimistic timescale.
I assume that now the beta release is out the intention is no further breaking changes. Then if SymPy master is currently compatible with NumPy 2.0.0b1 a good time for a SymPy release with accumulated fixes is... ASAP!
Presumably now that NumPy 2.0 is branched downstream projects should test in CI against the prereleases (pip install --pre numpy) rather than the nightly wheels. (SymPy does not usually test against the nightly wheels. I added that for NumPy 2.0 but maybe we should keep it...)
The rc1 release is now waiting on pybind11, so there is an (uncertain) delay.
Pybind11 2.12.0 was released earlier today, so everything is unblocked. We should see a numpy 2.0.0rc1 very soon - https://github.com/numpy/numpy/pull/26149 is the one PR we still need I believe. Cheers, Ralf
participants (6)
-
Aaron Meurer -
Charles R Harris -
Ilhan Polat -
Oscar Benjamin -
Peter Hawkins -
Ralf Gommers