yank release 3.0.3

[d-v-b]

we broke a certain indexing incantation with our last release. see #2849. This can lead to data corruption in some cases, so we want to yank release 3.0.3. We will release 3.0.4 today with a fix for the bug. Any objections?

[dstansby]

👍 for yanking if it's messing up written data. Just to check, the bug was introduced in 3.0.3, and isn't in 3.0.{2,1,0}?

[d-v-b]

I think the bug was associated with #2784? @dcherian @ilan-gold please confirm

[ilan-gold]

3.0.2 (with commit in the photo for others to confirm):

git checkout v3.0.2

import zarr
import numpy as np
g = zarr.open_group("foo.zarr", zarr_format=3, mode="w")
a = g.create_array("bar", shape=(10,), chunks=(3,), dtype=int)
data = np.array([7, 8, 9])
a[slice(7, 10)] = data
np.testing.assert_array_equal(a[slice(7, 10)], data)

3.0.3 (same):

git checkout v3.0.3

import zarr
import numpy as np
g = zarr.open_group("foo.zarr", zarr_format=3, mode="w")
a = g.create_array("bar", shape=(10,), chunks=(3,), dtype=int)
data = np.array([7, 8, 9])
a[slice(7, 10)] = data
np.testing.assert_array_equal(a[slice(7, 10)], data)

[ilan-gold]

If you want to check the PR specifically:

git checkout feeb08f4e49f6574d712fe5ceb42ce80ab6ceb3f

And one commit before it:

git checkout c66f32b93f1dfe4b4e6deebb22a5ea45d40c5aa5

[dcherian]

#2854 reverts it and adds extra testing

[dcherian]

That revert PR had the same failures, so I updated #2851 to skip those tests, and opened #2855 .

I don't have rights to do this so someone will need to
a. yank 3.0.3
b. merge #2851 and release 3.0.4

[jhamman]

I don't have rights to do this so someone will need to
a. yank 3.0.3

Neither do I. @alimanfoo, @jakirkham, @rabernat - can we get the full @zarr-developers/python-core-devs team added as maintainers to this project on PyPI?

[jhamman]

We should also mark 3.0.3 as broken in conda-forge. Has anyone done this before? I think these are the relevant docs: https://conda-forge.org/docs/maintainer/updating_pkgs/#removing-broken-packages

[joshmoore]

(Looking into it this morning. If you're pypi differs from your github, please say the word.)

[joshmoore]

Release yanked, in the process of inviting to pypi

[d-v-b]

thanks josh!

[ilan-gold]

Thanks @joshmoore I feel bad causing all this hooplah, but also I just want to say I'm so happy with this project. Y'all are awesome!

[dcherian]

I feel bad causing all this hooplah

you didn't! bad test coverage did!

[d-v-b]

much better to have this hooplah now than later!