Release version 0.3.0 final

[embray]

No description provided.

[astrofrog]

Main issues that need to be resolved for 0.3.0:

• Allow sets of units to be enabled in astropy.units (Allow sets of units to be enabled #1268) [very close]
• Refactor astropy.coordinates to support arrays (Implement support for arrays in astropy.coordinates #884, Distance to quantity #1472) [close]
• Make any required changes in astropy.coordinates API while it is still 'in development'
• Change SOFA to ERFA (Change SOFA to ERFA #1195, Change SOFA by ERFA #1293)
• Finalize (as much as possible) API for astropy.models, update intro docs (Simplify/clean up 'Getting started' section for astropy.modeling #1473)
• Fix convolution kernel docs, move convolution to top level (Fixes to new kernel documentation #1451) [close]

More minor issues that don't all have to be done by 0.3.0 but would be nice:

• Refactor internals of Table/NDData I/O (Refactor Table readers and writers to use classes rather than free functions... #962, Refactor/rethink aspects of the Table I/O #1046)
• Improve interoperability of packages, including more support for quantities/units (e.g. Set column units when initializing Column or Table from a Quantity #732)

And of course any other issues tagged as 0.3.0 are a bonus.

[pllim]

I see some VO stuff tagged for v0.3.0 but not listed above. Does that mean they will not be in v0.3.0? #552, #933, #945. Just need to know if there is a need for me to get to them before the release.

[astrofrog]

@pllim - the above are 'big picture' things that aren't done yet, but anything merged before the 0.3.0 release will be in 0.3.0. We should try and get as much as possible that is tagged 0.3.0 implemented and merged over the next month, but the main list above are things that are critical to the release.

#552 is already part of 0.3.0 by default since it's already merged into master. The two other issues you mention would be nice to have for 0.3.0, so feel free to work on them!

[astrofrog]

@embray - just to check, is the only thing remaining to do merge #1770 and update the changelog? Is there anything I can do to help with the changelog?

[embray]

Yes, hopefully this will finally be done today. I'm working on a script that will help massively with the changelog. Should be able to get it done in a couple hours....

[astrofrog]

@embray - sounds great!

[astrofrog]

@eteq @embray - we don't want this to drag on for too long, so how about setting a firm deadline of Wednesday for the final tarball release?

[embray]

That works for me. Let's make sure a fix for #1804 gets merged and that the tests pass with it. Unless anything else comes up that should give plenty of time to do it by tomorrow.

That said, do we think the release candidate has gotten enough testing, given that there were issues installing it?

[astrofrog]

I personally think that we should release tomorrow unless any major issues come up - to some extent we understand what the issues were with e.g. pip and conda, and some people did manage to install it since we did get a few bug reports (such as #1804). But we could send an email to the list asking for any final testing ahead of the planned final release tomorrow?

[eteq]

I think we got quite a few installs based on Tom A's spreadsheet, and only the Anacoda/pip one failed. So I'm ok with releasing after #1804 if we can get a fix for that actually in early enough tomorrow. (Also need a decision on #1808, but that's pretty much just "merge or not").

[astrofrog]

@embray - can we go ahead and close this or are there any remaining things to be done apart from announcing the new release?

[embray]

Assuming the announcement is something you'll just do regardless then yes, this can be closed.