roadmap to 1.0

[chadwhitacre]

Target: September 19, 2016 👍

1.0:

• port to Windows (implement Windows CI #443)
• rationalize dispatch (Testtable #366)
• switch back to stdlib logging (We should go back to using logging #157)
• port to aspen.py (port to aspen #548)
• port to Python 3 (Support Python 3.x #493)
• simplify configuration (do away with configure-aspen.py #373, stop taking configuration from argv #374, Website should receive instead of subclassing Configuration #444, make configuration immutable #473)
• fix some bugs/issues
  • return empty list for request.body.all('not-there') #429
  • Configuration is still taken from the environment aspen.py#24
  • The names of the codec settings are confusing aspen.py#12
  • README isn't rendered correctly on PyPI aspen.py#22
• mothball the other language versions—roadmap to 1.0 #357 (comment)
• stub out RtD sites—roadmap to 1.0 #357 (comment)
  • Aspen
  • django-aspen
  • Flask-Aspen
  • Pando
• publish docs on RtD
  • Aspen
  • django-aspen
  • Flask-Aspen
  • Pando (move docs to RtD #339)
• release 1.0 versions
  • Aspen
  • django-aspen
  • Flask-Aspen
  • Pando
• dust off editor plugins
  • simplate.vim
  • simplates.emacs
• sort out Aspen plugins
• publish the new http://aspen.io/ (Build out http://aspen.io/ for relaunch aspen.io#2)
• redirect http://simplates.org/ to http://aspen.io/

Post-1.0:

• flatten algorithm further (unroll deep calls) (rationalize and document algorithm functions #423)
• calve off simplates (factor out simplates #341), this would take renderers with it

Also relevant:

• find or make a library library (find or make a library library gratipay/inside.gratipay.com#309)

[chadwhitacre]

These are the things we want before declaring 1.0 and caring about backwards compatibility.

[chadwhitacre]

Yay @pjz for rationalizing dispatch in #366! 💃

[chadwhitacre]

Yay @pjz for porting to Windows in #443! 💃

[chadwhitacre]

Yay @pjz for refactoring simplates into a submodule in #481, paving the way for #341! 💃

[chadwhitacre]

Decided on #537 (oooh! 537 get! 💃) to aim for a 1.0 release on September 19, since that is 10 years to the day from the first commit on the Aspen repo.

[chadwhitacre]

In light of recent events, I've subbed the following in the description:

• drop Request/Response objects (join 'em #526)

for

• port to aspen-core.py

[chadwhitacre]

AspenWeb/aspen.io#1 (comment)

How about a 0.90 release for Aspen in the next few weeks, and 1.0 in September?

My tendency would be to stick with our current 0.n+1 versioning scheme until 1.0.0.

[chadwhitacre]

Yes, python3 is the next item in my TODO. Don't know when I'll get around to it, but it should be doable before 1.0.

Awesome! Besides that, the other major piece of coding to be completed is #548, which you've got underway in #556.

As discussed, I'll commit to AspenWeb/aspen.io#1 (I've added it to the roadmap). I'm also happy to take #339, which could surface some API cleanups along the way.

[chadwhitacre]

@Changaco Should we drop #473 from 1.0?

[Changaco]

I don't know, but I don't see #473 as a blocker for 1.0 if that's what you're asking.

[chadwhitacre]

Yeah, that's what I'm asking.

[Changaco]

I've just noticed that there was an old 1.0 milestone, so I've closed it.

[Changaco]

And the Simplify Simplates milestone too.

[chadwhitacre]

I've closed it.

Eep! Does that mean we've reached 1.0?! 😳

[Changaco]

Nope, especially since it still had an open issue. I closed the milestone because it was out-of-date since you've been using this issue for the 1.0 roadmap instead.

[Changaco]

I'm ready. ;-) I also have some contract work, but the job is almost finished, and it didn't prevent me from working on Aspen in August anyway.

[chadwhitacre]

Eeeeeeeee! We could get pando.wf. 😁

[chadwhitacre]

I've staked out Pando on PyPI and added you as an owner, @Changaco.

[chadwhitacre]

Ditto django-aspen and Flask-Aspen (though I don't expect you to care much about those ;).

[chadwhitacre]

@Changaco I've updated the task list in the ticket description here.

[Changaco]

@whit537 Have you looked at the open aspen.py issues? Is there any you'd like to see closed before 1.0?

[chadwhitacre]

@Changaco Looking at the four issues, it seems that AspenWeb/aspen.py#24 and AspenWeb/aspen.py#12 would be good to get done before 1.0, to avoid having to maintain backwards-compatibility. AspenWeb/aspen.py#23 shouldn't be a problem to add later. AspenWeb/aspen.py#22 should get done, but I expect that to come out in the wash when I start working in earnest on Aspen's docs.

[chadwhitacre]

Picking up from AspenWeb/aspen.io#1 (comment) ...

I suggest doing nothing other than changing the repo descriptions.

I renamed the repos:

https://github.com/AspenWeb/experimental-go-version
https://github.com/AspenWeb/experimental-javascript-version
https://github.com/AspenWeb/experimental-ruby-version

And I modified their description:

Experimental {} version of Aspen/Pando. Want it? :) http://aspen.io/

[chadwhitacre]

I'm working on configuring RtD for our several projects.

@Changaco Should we use subdomains of readthedocs.io or of aspen.io for project docs?

http://pando.aspen.io/
http://flask.aspen.io/
http://django.aspen.io/
http://core.aspen.io/
http://algorithm.aspen.io/
http://filesystem_tree.aspen.io/
http://dependency_injection.aspen.io/

... or ...

http://pando.readthedocs.io/
http://flask-aspen.readthedocs.io/
http://django-aspen.readthedocs.io/
http://aspen.readthedocs.io/
http://algorithm.readthedocs.io/ (already taken)
http://filesystem-tree.readthedocs.io/
http://dependency-injection.readthedocs.io/

For the last three we currently use:

http://algorithm-py.readthedocs.io/
http://filesystem-tree-py.readthedocs.io/
http://dependency-injection-py.readthedocs.io/

This suggests either that the others would also have -py in the domain, or that we would use aspen.io to avoid namespace collisions. Or we could leave those three as-is since they're lower-level libraries that don't really have anything to do with Aspen per se, and use either of these sets for the other four:

http://pando.aspen.io/
http://flask.aspen.io/
http://django.aspen.io/
http://core.aspen.io/

or under RtD, with or without the -py:

http://pando.readthedocs.io/
http://flask-aspen.readthedocs.io/
http://django-aspen.readthedocs.io/
http://aspen.readthedocs.io/

Do you care?

[chadwhitacre]

I'm leaning towards leaving the three lower-level libraries as-is, and using {}.aspen.io for aspen.py and above:

http://pando.aspen.io/
http://flask.aspen.io/
http://django.aspen.io/
http://core.aspen.io/

[Changaco]

I don't really have a preference.

[chadwhitacre]

Okay, proceeding with subdomains of aspen.io for the four new doc sites, leaving the other three alone.

[chadwhitacre]

Okay! I've stubbed out four new RtD projects:

http://core.aspen.io/
http://django.aspen.io/
http://flask.aspen.io/
http://pando.aspen.io/

[Changaco]

If we have enough time we should try to take care of #222 before 1.0.

[chadwhitacre]

@Changaco FYI I'm working top down from AspenWeb/aspen.io#1. Once the homepage looks good enough, I'll start working through the links away from it:

• framework libraries
• aspen plugins
• editor plugins

Making each of those linked pages say what we want them to say will fan out into code changes. Basically, however far we get in the next week is what I expect we'll declare 1.0.

[Changaco]

I'm going to release Aspen 1.0rc2 and Pando 0.44, then see if Liberapay can run on top of those in production.

[Changaco]

/o\ I forgot to adapt Pando's code to AspenWeb/aspen.py#31: https://travis-ci.org/AspenWeb/pando.py/builds/158823441

[Changaco]

Okay, second Liberapay PR is liberapay/liberapay.com#408. I'm going to sleep now, I'll probably deploy tomorrow.

[Changaco]

https://liberapay.com/ is now running on Pando 0.44 and Aspen 1.0rc1. \o/

[chadwhitacre]

Alright, my moderated goal at this point is to release Aspen 1.0 next Monday, with Pando 1.0 to follow sometime after that. Flask-Aspen and django-aspen will also have to wait, along with the new aspen.io website. The old aspen.io site still points to Aspen 0.42, and I think we can leave it that way until all the pieces are in place. I think that's a more realistic goal, given what I can dedicate to this this week.

Can you live with that, @Changaco?

[Changaco]

I think we really need a new aspen.io, the old one is obsolete. I don't think the flask and django plugins have to be ready for production in order to push the new aspen.io, we can refer to them as WIP and put pando on top for now.

It seems like a big share of the remaining work is documentation, and so far you've been doing it alone, but if you delegate some of it to me then we should be able to do more.

[chadwhitacre]

How about doing AspenWeb/aspen.py#32 for Pando? Right now http://pando.aspen.io/ is empty. Surfacing all of the existing module reference documentation seems like the first step in writing the documentation for Pando.

[Changaco]

Okay, I'm on it.

[chadwhitacre]

Sorry we didn't make the September 19 target, @Changaco. 😞

[Changaco]

That's fine, we've still made good progress. :-)