patrick.sirref.org
PhD Student @ Uni of Cam
https://patrick.sirref.org
https://github.com/patricoferris
32 posts
58 followers
50 following
Regular Contributor
Conversation Starter
comment in response to
post
It was pretty frustrating come up mostly short on a solution for io-uring that was reliable and that would remain stable across kernel versions. Any pointers there would be appreciated :)
comment in response to
post
This has additional support for tracking subprocesses and io-uring which bypasses the common syscall tracepoints.
comment in response to
post
(( seems a bit broken on safari... ))
comment in response to
post
There's a little post about it here patrick.sirref.org/try-oxcaml
comment in response to
post
Tough times for ppxlib... ^^"
comment in response to
post
> Oops!
> Unable to resolve handle
comment in response to
post
Hey @david.sancho.dev!
Ppxlib needs a target parsetree. A parsetree that Ast_builder functions produce values for. The migrations allow ppxes to move to this target AST. If we don't bump that AST then ppx authors cannot target new features of the language (e.g. Pexp_effect), does that make sense?
comment in response to
post
@kirancodes.me + @anil.recoil.org mostly manual, although we have some tools to help (e.g. a revdeps CI tool that was very useful for building a stable opam-repository overlay).
I just started patching based on the CI though so maybe that should have been more directed... no llamas tho :S
comment in response to
post
All of this was so that VPNKit can become OCaml 5 ready!
comment in response to
post
Furthermore, if we want to retain the ability to extend the functionality of the interfaces (e.g. providing a reset function for a clock for testing) we will need to make upstream changes to the libraries provide by mirage (making them unwrapped) and/or make changes to dune.
comment in response to
post
The TL;DR is that using dune virtual libraries will probably make most applications simpler but there is a cost to porting any code that wanted to have multiple implementations of a given interface inside one application.
comment in response to
post
I suppose this could be an OCaml community Outreachy project for the right mentor too :))
www.outreachy.org/communities/...
comment in response to
post
That's awesome, congratulations!
I saw last week that we just got an FStar on opam that will work with OCaml 5+ github.com/ocaml/opam-r... :))
comment in response to
post
Or see the great progress being made on ocaml-api-watch, our current internship project!
github.com/ocaml-semver...
comment in response to
post
For ideas and inspiration, take a look at past projects on the OCaml website ocaml.org/outreachy
comment in response to
post
As a data point, you could have a look at icfp23.sigplan.org/details/ocam... in particular the "3.2 Node annotations" section (I think the "families" of expressions being distinguished via the type system using object types is pretty nifty)
comment in response to
post
I don't think there would be any real reason but you can request it for V3! bsky.app/profile/anil...
comment in response to
post
As you can see, haven't quite got rich text facets working ^^"