-
-
Notifications
You must be signed in to change notification settings - Fork 328
2.3 release planning #339
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Comments
I'd be happy to revise the list of issues currently tagged in 2.3 - it was
ambitious - and go for a 2.3 release sooner.
…On Tue, 20 Nov 2018, 05:05 Ryan Abernathey ***@***.*** wrote:
We have a new feature--consolidated metadata--that some of us (me!) are
eager to get into production.
Zarr releases quite infrequently compared to other OS projects I'm
involved in. (Our last release was 8 months ago.) The list of things to
fix for 2.3 <https://github.com/zarr-developers/zarr/milestone/7>
contains some hard items which potentially involve revising the spec.
@alimanfoo <https://github.com/alimanfoo> & @jakirkham
<https://github.com/jakirkham>, are all of the things on that list "must
haves" for a new release? Or perhaps we might consider a more frequent and
less ambitious release cycle?
Or maybe, although this is not a bugfix, we do a micro release?
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#339>, or mute the thread
<https://github.com/notifications/unsubscribe-auth/AAq8Quu35soci6hJNo0LBVkRXWIrHRtCks5uw42dgaJpZM4YqY0g>
.
|
Well personally I try not to release before going on vacation (US Thanksgiving). 😉 That said, not opposed to moving towards a release here. Would be nice if we could get Numcodecs out the door first though as we've been working towards a release there recently and are pretty close. Perhaps as a good first step towards a Zarr release it would be nice to have some testing of Zarr currently (particularly new features). Also it would be good to re-triage the 2.3 milestone and look to see what are critical bugs (need fixes) vs. cosmetic things (nice to haves). We can probably re-milestone the latter (unless they are particularly important to us). |
Took the liberty to rename this issue "2.3 release planning", suggest we use issue this to discuss what remaining work is in scope for 2.3. |
Agreeing with @jakirkham, in an ideal world it would be nice to get numcodecs 0.6 out first, then finalise and release zarr 2.3. Part of the reason for this is that there's a fix in the new version of numcodecs for a bug in the current msgpack codec, which is currently worked around in the zarr tests (in fact the zarr tests assert the broken behaviour). So updating numcodecs would enable the zarr tests to get fixed, and this would also unlock a problem in another piece of work in zarr. See zarr-developers/numcodecs#98 and #324 for details. However, the two releases are not strictly tied, so if there was pressure to release zarr and we get blocked on numcodecs, we could be flexible. FWIW for Zarr 2.3 there are three bugs would be good to fix: #263, #272 and #253. Two of these (#272, #263) have a mature PR, the other (#253) hasn't been worked on yet but shouldn't be too complicated. The other issues currently tagged in the 2.3 milestone I could live with postponing. |
Some issues cropped up related to |
Thanks for entertaining my suggestion. I'm not saying this is super urgent, just that we might want to consider a release whenever we get a significant new feature or bug fix. Sounds like that is doable. We are planning a bunch of pangeo demos in early December. So it would be great to have the release before then. But we can always set up environments that use the latest github master, so again, not super urgent. |
FWIW I think it would be good to push for numcodecs and zarr releases
before the end of the year if not sooner. There is good momentum so a bit
of extra impetus to get something out the door is helpful!
…On Thu, 22 Nov 2018, 01:53 Ryan Abernathey ***@***.*** wrote:
Thanks for entertaining my suggestion. I'm not saying this is super
urgent, just that we might want to consider a release whenever we get a
significant new feature or bug fix. Sounds like that is doable.
We are planning a bunch of pangeo demos in early December. So it would be
great to have the release before then. But we can always set up
environments that use the latest github master, so again, not super urgent.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#339 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AAq8QluPC2MffXeaf8qHkyR8GNP4rzJgks5uxgOngaJpZM4YqY0g>
.
|
At this point, I think we just need more eyes on the PRs for v2.3. Nearly all of them already pass CI, but need a review and/or decision. |
Alright, I think we should start thinking about the 2.3 release again. My suggestion would be the following remain on the milestone and everything else gets pushed off until the next release.
These all have PRs and AFAICT could be merged and then released today if we wanted to. The other issues are about tricky Jupyter UI issues, changing getsize's behavior, supporting pathlib, and handling read-only stores through some APIs respectively. All of these will require more work and more thought. So I'd suggest that they be punted on and handled in a later release. Thoughts @zarr-developers/core-devs? Edit: Should add everything in @alimanfoo's scoped down list is either included in this or resolved except for getsize. |
I favor a release asap. Xarray 0.12.0 is coming soon (pydata/xarray#2776) which has several zarr-related features (e.g. consolidated metadata). It would be nice if these happened around the same time. |
Not sure if I'm too late to the party, but I'd love to see
make it into the release. |
Personally I'd also like to see PR ( #309 ) integrated into the release as there are some people internally that are eagerly awaiting it. That said, I'd be ok sticking with my previous list (without any additions) to allow a faster turnaround. |
Would be great to get the Azure blob store into the release, IIRC it is
basically ready to go.
…On Sat, 23 Feb 2019, 23:05 jakirkham, ***@***.***> wrote:
Personally I'd also like to see PR ( #309
<#309> ) integrated into the
release as there are some people internally that are eagerly awaiting it.
That said, I'd be ok sticking with my previous list
<#339 (comment)>
(without any additions) to allow a faster turnaround.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#339 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AAq8QtKvmuGUBp-surEn-8n-3NIDrGtrks5vQclUgaJpZM4YqY0g>
.
|
Bumping this again. What are our thoughts on making a release? |
I think we are all good to go for a release. All the PR that have been brought up in the discussion have been merged. Xarray 0.12 just came out. Let's go! |
Thanks so much @jakirkham for keeping up the momentum on this. I am all in favour of making a release. Would you like to do the honours? |
Sure happy to move this forward. Will update if there are any issues. |
Packages for 2.3.1 are now available on PyPI and conda-forge. |
Thanks so much @jakirkham for dealing with the release friction! This is thankless but important work! |
Thanks also from me, fantastic to see this go out, 🍻 |
Thanks all for contributing! This is a very nice release 😄 |
Would anyone care to write a very short blog post about the release? This could go on Alistair's blog or we would be happy to host it on the Pangeo medium blog. Then we will have something to tweet about. :) |
A blog post would be very cool. I started sketching something, contributions/comments very welcome: https://docs.google.com/document/d/14qvMzK3HL20E63BqI3Q5EG_imoJ7h7pXhKBc_vVxnfY/edit?usp=sharing @jakirkham you were the mastermind of this release, do you have a blogging channel you like to use? |
No strong preferences on where we post it. Maybe it makes sense to have a blog for the Zarr project itself? |
That would be good. Use medium (like pangeo)? Vanilla jekyll site on github pages (i.e., zarr-developers.github.io)? I think I'd prefer jekyll and github pages just because it might be a bit easier to collaborate on writing blog posts via PRs, but no strong preference. Btw I have two working days before going on leave for 3 weeks, so have a short window in which I could help. Also very happy for someone else to run with this if it's not done before I go away. |
If you are comfortable setting up the framework for the repo, that would be very helpful. |
I've created a new repo https://github.com/zarr-developers/zarr-developers.github.io which holds source for the blog served at https://zarr-developers.github.io. Creating a post just needs a new markdown file in the |
We have a new feature--consolidated metadata--that some of us (me!) are eager to get into production.
Zarr releases quite infrequently compared to other OS projects I'm involved in. (Our last release was 8 months ago.) The list of things to fix for 2.3 contains some hard items which potentially involve revising the spec. @alimanfoo & @jakirkham, are all of the things on that list "must haves" for a new release? Or perhaps we might consider a more frequent and less ambitious release cycle?
Or maybe, although this is not a bugfix, we do a micro release?
The text was updated successfully, but these errors were encountered: