Skip to content

dev meeting 20200806

Jump to bottom
Nathan Rebours edited this page Sep 17, 2020 · 1 revision

Present at the meeting:

  • Sonja Heinze (@pitag-ha)
  • Carl Eastlund (@ceastlund)
  • Jérémie Dimino (@jeremiedimino)
  • Nathan Rebours (@NathanReb)

Current plan

  • Carl finishes the omp2 compatible ppxlib
  • Sonja adds Location.init to ppxlib
  • We release omp.2.0.0 and the corresponding ppxlib
  • Nathan adds the required features for bisect_ppx bucklescript compatible driver

Meeting notes

Nathan haven't done much on ppxlib since the last meeting, he merged the documentation PR and will address the external reviews separately. He hasn't received an answer from Anton regarding bisect_ppx and bucklescript so no news on this front.

Sonja addressed the comments on the tyxml ppxlib port PR so this should be ready to merge anytime now. She added the support for instrumentors in ppxlib that was required to finish up the bisect_ppx port properly. She also released ppxlib.0.15.0 which doesn't depend on base anymore. A few packages silently depended on base transitively but the right upper bounds have been added so everything's good on that front. Finally while working on tyxml she realized that ppxlib's Location module lacked the init function required to initialize a lexing buffer so she'll work on adding that.

Carl made good progress on making ppxlib compatible with omp2. Everything that we needed has been vendored into ppxlib and he is now working on polishing it but it should be ready by next week so we can move on and release both omp.2.0.0 and a new ppxlib that's compatible!

As Jeremie will be off for a little while we discussed the following steps once we have reached that first milestone and released the new omp and ppxlib:

  • We'll work on Astlib, starting on the local copy that will leave in ppxlib before upstreaming it to the compiler. It will basically contain everything that's currently in omp2 + everything we need to implement ppxlib on top of it. We should aim at keeping its API minimal to make sure it can stand the test of time. Sonja will take care of this part.
  • We'll need to propose a solid and simple workflow for updating Astlib with new versions of the Parsetree types so that its maintenance doesn't become a burden for the compiler team. Carl will look into this.
Clone this wiki locally