Doccords v0 Release

[drbeefsupreme]

There are two main issues with doccords release:

• Memory usage upon upgrade
• Changes to the binary to jet the changes to the parser

Updating hoon.hoon means rebuilding everything, so as an OTA it has a high
memory requirement. Thus we probably ought to wait until @philipcmonk
completes changes to Ford that are designed help with the OTA memory usage
issue.

I don't really understand what would be involved with changes to the binary. I
should meet with @joemfb to help figure this out.

[drbeefsupreme]

#5849 needs to be fixed in a separate OTA that goes out before doccords release.

[zalberico]

Last week:

• The :: syntax is working!

This week:

• Knocking out a bunch of small issues
• Refactoring the parser to clean things up prior to release
• Jon ran into an issue with the Ford rune parsing, would be helpful to look it over with @Fang-

By EOW should be ready for code review of hoon.hoon changes - Jon will ping the engineering chat when that's ready.

[zalberico]

What remains (summary from @drbeefsupreme in team sync):

• Going through each rune one by one and making sure comments attach as they're supposed to
• Get the hoon.hoon PR ready for review
• A little bit of refactoring work to get the parser in a good state
• Then moving to the printing library: sometimes when asking the dojo for a doccords on a type it hangs for a long time, not sure why (may need a timeout here).

[drbeefsupreme]

Current branch for :: syntax changes:

https://github.com/urbit/urbit/tree/jon/doccords-parser-simplify - pardon the unhelpful commit messages, I'm planning on squashing them all into one when merging back into this branch.

It's cooled down from "exploded whitespace parser" back into something that thankfully looks a lot like the parsing for :> and :< syntax with a handful of new whitespace parsers in +norm that fit the pattern of the existing ones.

I'm expecting it will be good for review sometime this week (hopefully tomorrow).

[zalberico]

From talking with Jon in 1:1

• This may require a new runtime version for some jet related stuff (if so we should create an issue with details and link it here)
• +* is overloaded and the second syntax can mostly be replaced by |$ #6052 is separated out because it would require a hoon kelvin version change, but since it doesn't require app developers to change things we should do it and ship it.
• @drbeefsupreme will follow up with @joemfb about doing a synchronous code review on the hoon.hoon PR to get feedback and get it ready to release.

[zalberico]

Only remaining thing here is @drbeefsupreme needs to go over jet signatures with @joemfb, otherwise we think we're ready for release.

@drbeefsupreme is also thinking through other potential ways doccords could cause failures.

[zalberico]

@drbeefsupreme were you able to get the review in a good place? I'd like to have this in the 'ready to merge' category before the end of the year so we can get it into one of the first weekly releases.

[drbeefsupreme]

Sorry @zalberico , I need to mess with my GH notification settings so I don't miss these things.

The printing library and dojo changes haven't been reviewed at all yet. I'd say hoon.hoon is roughly 75% reviewed (whatever that means - its just somewhere more than halfway). Since @joemfb is sick and @Fang- is out for the rest of the year, we might not be able to finish the review before the end of the week. It will just depend on how much time @joemfb ends up having this week.

[drbeefsupreme]

Today I finally got my hands dirty reading/working through the documentation on jets. Since @joemfb is out and we want to have a release ready in a couple weeks and I was going to rely on Joe to handle the release/runtime stuff, I'm going to take a crack at handling all the release stuff myself. I'll definitely still need some help, especially process-wise.

Today was all tutorials, tomorrow I'll try actually fixing the jet registration in the doccords branch. Then I'll move onto figuring out the changes in Vere required for a Hoon kelvin decrement.

One outstanding question: what does the Kelvin shimming stuff #6073 have to do with jets and the runtime? Are we going to leave old jet registrations in the runtime or....?