<@U06A03HV1> <@UB2J9BQA0> is starting to bump into...
# development
a
@witty-crayon-22786 @hundreds-father-404 is starting to bump into v2 target API problems, so we should probably sync up on those
h
Agreed. I’m meeting with Stu today 11-11:30 to talk about V2
a
Awesome :)
h
a
Oh wow… No, that’s pretty old
I suspect there isn’t a huge amount of useful documentation anywhere except the code today…
h
that issue is how I found the doc
a
Aha
Yeah, just the first issue description has been updated to have some information; but I suspect because this is all in flux it’s mostly just in mine and Stu’s and Danny’s heads
h
Which is one of the reasons I’m really excited to be working on V2. All these things are new for me, and I’m hoping to be able to translate what I learn into some type of formal documentation
a
😄 That would be amazing!
(Though part of why it’s not currently documented is because of how in flux it is)
h
Which I was thinking about too. But I think that us writing out documentation will help us to think through our design and decide if we should change it. Also maybe easier to update out-of-date docs than to write them from scratch?
a
Yeah, writing docs up front can definitely help frame design 🙂
👀 1
w
some of that doc might still be useful, some of the comments on 4535 might still be useful
i have some more recent notes, but i ended up getting a bit lost in them this weekend, and needed to just dive in to make an incremental improvement
a big potential trap is just making sure we're ready for codegen. we proved that the engine makes it possible to do that in a very clean way, but it's very tied up in the target API
but i mentioned to eric this morning that i think that to a large degree, we should be trying to just work around the target api for the length of the python test remoting project
(making improvements where obvious/helpful, but mostly doing a lot of set/get attr)
we proved that the engine makes it possible to do that in a very clean way
"we" being @enough-analyst-54434 and i, during the initial engine experiments
h
Agreed. I’m going to be on the lookout for easy wins, but will try to refrain from blocking myself too much from major changes like populating None attributes Once remoting is complete, I will have substantially more experience with V2 and then would love to work on direct V2 stuff though!
w
👍
a
i've been making quite a lot of github issues regarding the v2 engine, that seems to be a lower barrier to entry than making a whole doc unless we're pretty clear on the doc contents
👍 1
w
we're not clear on the doc contents. but that's sometimes sortof-semi a reason to create a doc
when you have a large(r) problem that needs a cohesive solution
unknown.
(the fields externalization doc was a case of not iterating back and forth between implementation and design, i think... it shouldn't have ended up frozen in time like that. mea culpa)