Adapt w.r.t. coq/coq#19943.

[ppedrot]

No description provided.

[ppedrot]

This is not super pretty given that sexplib doesn't like GADTs. I had to hack my way around.

[ejgallego]

It is indeed a bit PITA, actually I've been thinking if we want to seralize stuff in kernel, in the past it was used by the ML work of Kaigu etc..., I dunno. Most people is using the AST these days, I'll do some research into that.

There are some researchers in SE which have stated interest in using seralized kernel envs tho, what do you think @palmskog ?

[palmskog]

@ejgallego personally I think it would be nice to investigate kernel-level fuzzing (the original idea), but there are no plans from my side (or any other side I know about) to pursue this right now.

[ejgallego]

@ejgallego personally I think it would be nice to investigate kernel-level fuzzing (the original idea), but there are no plans from my side (or any other side I know about) to pursue this right now.

The original infra is still functional, and the fuzzers are much better these days, it wouldn't be too hard to co-advise an internship on that topic (I can get students / funds in Paris).

[palmskog]

The original infra is still functional, and the fuzzers are much better these days, it wouldn't be too hard to co-advise an internship on that topic (I can get students / funds in Paris).

Then sure, if you find a student (and you have funding for your own time), please let me know.