How come we already have 2.14.x docs? We haven't c...
# development
h
How come we already have 2.14.x docs? We haven't cut an rc for 2.14 yet
h
I forked it yesterday because Tom was creating a new Kotlin page. I figured now that we have in-repo docs, there's no downside to "prematurely forking" like we used to have -- we land changes on
main
, then cherry-pick to 2.13. Whereas before, we changed 2.13 then forward-ported to main Does that make sense?
h
It's visible to end users at the moment
Which is just confusing
Isn't the kotlin stuff going into 2.13?
And it's just another thing to keep in sync, the new docs process is already somewhat laborious
So I think there is a downside
For now I've made it non-public so at least users don't see it
h
And it's just another thing to keep in sync, the new docs process is already somewhat laborious
How so? The new docs process is that we already must be keeping things in sync -- when you make a PR that impacts bheavior, you are expected to fix docs or open a ticket to track it
Which is just confusing
Imo not that confusing -- we have the 2.14 dev releases already out. But fine with me to hide it
Isn't the kotlin stuff going into 2.13?
Yes, but we also need it to show up w/ all future docs
h
None of what you're saying requires the 2.14 docs to exist on readme.com though
By "keep in sync" I mean actually syncing to the live docs
It's one more version to remember to sync
h
okay, do you want us to delete the 2.14 fork then? fine with me
a
@happy-kitchen-89482 We require a development branch to track
main
— current agreed practice is to create a
beta
version that tracks the future release
Readme requires that there be a version of shape or form for every forked version of the docs
h
I don't understand - why would readme care what we have in
main
if we're not syncing it up?
a
I think we should be syncing development version docs as a matter of course (most open source projects do this): 1. We don’t have a way of verifying whether they render correctly on Readme without syncing them to readme, and it’s better to catch errors as they’re introduced rather than during the audit cycle 2. It’s easier to switch between docs versions on Readme, so if features have changed significantly between stable and
main
, it’d be good to find all of the docs in the same place (it makes no sense to go looking for dev docs in-repo and stable docs at Readme)
👍 1
h
We can have a
main
version of the docs, with the implication that they are unstable etc, but it's not awesome for users to see 2.12, 2.13 beta and 2.14 beta as options
(beta meaning tagged as beta on readme.com)
a
Readme requires a numbered version
h
GODDAMNIT README
a
correct
I think it’s possible to have “2.14-development” as a branch name
h
it is, but then when we rename to
2.14
, it will break all prior links. Maybe fine? I guess we don't link a ton to docs for
main
a
the other alternative is to have
main
be labelled as
2.999-development
(beta)
all the time
and then
2.14
would be forked when the rc is cut
h
Ech
Better to have
2.14
than those alternatives
☝️ 1
h
I don't think it's that big of a deal to have 2.13 and 2.14 both be in beta? That's how our release process works - no need to hide it It should be clear 2.12 is what's stable
h
I don't like it, but whatever
a
neither. The joy of inadequate tools 🤷
h
so then are we fine keeping up the 2.14 docs I created? I agree w/ Chris I think it's helpful to have
main
style docs for early adopters
h
"fine" is overstating it... but yeah
I'm grudgingly acknowledging that it is better than the alternatives
👍 1
h
heh I'm very on board w/ moving to Sphinx, when we have the itme