Release?

[jnothman]

It's been over a year since the last release. I would like to use recent changes in scikit-learn, and it looks like numpy too is being built on the latest release, not on master. Could we more generally plan half-yearly releases or something, even if they're quite small?

[rgommers]

Yes, good idea. I thought we had released 0.7.0 a few months ago, but apparently not. IIRC we need to resolve the duplicate sections issue still.

I can make some time to do this release.

[jnothman]

Great, thanks!

…

On 13 March 2017 at 19:08, Ralf Gommers ***@***.***> wrote: Yes, good idea. I thought we had released 0.7.0 a few months ago, but apparently not. IIRC we need to resolve the duplicate sections issue still. I can make some time to do this release. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#89 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAEz67PLXLTfuEh5xeHEMdg-P2taos1Iks5rlPmCgaJpZM4MauJI> .

[pv]

I guess scipy docs don't build with the current master --- would need to check if the issue is here or elsewhere.

[jschueller]

@pv hi, I think I found a possible regression (#93), maybe it's that problem preventing you from compiling scipy ?

[jschueller]

Also, about the release, latest numpydoc release 0.6 cannot be used with sphinx git anymore as they removed some compatibility module in master last week, so I guess the next sphinx release will need a numpydoc release too.

[amueller]

Lol I was just about to open it, @jnothman beat me to it, of course.
What's the status? Is #93 blocking? Scipy build blocking? Would be cool to have a list.

[amueller]

Also: the current github master identifies itself as 0.6.0, which is not ideal.

[jschueller]

@amueller I incremented the version number to 0.7.0 in #94

[rgommers]

What's the status? Is #93 blocking? Scipy build blocking? Would be cool to have a list.

Last time I checked the Scipy docs built fine - will check again. AFAICS the one thing that needs resolving is gh-67.

[rgommers]

Current numpydoc master + scipy master builds fine.

[amueller]

I have some weird issues in scikit-learn that I don't understand. Any help with that would be much appreciated: sphinx-doc/sphinx#3768 (comment)
It only happens with the numpydoc update, so maybe it's actually a numpydoc error. I'll open an issue here - I have no idea what's going on.

[rgommers]

I have some weird issues in scikit-learn that I don't understand. Any help with that would be much appreciated: sphinx-doc/sphinx#3768 (comment)

For future reference: this was fixed by gh-93.

[amueller]

Also, now I'd be very happy with a release of the current state, because scikit-learn builds fine! Which doesn't sound like a big thing, but making numpydoc master work for us actually was a bit tricky ;)

[rgommers]

Also, now I'd be very happy with a release of the current state,

I think I'll work on that right now, before life interferes again:) gh-67 has a merge conflict and an un-addressed review comment, so we can always cut another release later (not too much work).

[rgommers]

Okay, done: https://pypi.python.org/pypi/numpydoc/0.7.0

Somehow the readme went missing, looked into that but no luck - opened pypi/legacy#653