Anton Bachin

For reference, I just parallelized Bisect_ppx's tester, which uses OUnit (https://github.com/aantron/bisect_ppx/commit/2784eef518321aeb6773370285a05030742d4d85). This changed its running time from about 13 seconds to 3 seconds (on my machine). Bisect_ppx's tests are CPU-intensive,...

I guess the issue with using threads is that concurrency libraries usually assume they are running in the main thread. For the fastest testing, the tester would run a number...

Quoting the relevant comment from https://github.com/ocsigen/lwt/pull/712 here. https://github.com/ocsigen/lwt/pull/712 restructured Lwt's custom tester to run tests concurrently instead of sequentially. The result: > This decreases testing time for Lwt_unix from about...

This also changes the nature of the tests one is comfortable writing. I/O tests often introduce artificial delays. When tests are run sequentially, these tests end up wasting developer time...

One potential issue here is that if tests are spread out over multiple files, either (1) the files have to still be referenced from somewhere to be linked, or (2)...

I want to add a vote in favor of textual diffing of S-expressions. For example, odoc's parser is tested almost entirely by [feeding it input strings](https://github.com/ocaml/odoc/blob/d5612dc9ad9a6efbf1d385514f5fba14cfee0eea/test/parser/test.ml#L19-L1075), getting the AST `result`s,...

To add more detail, during odoc parser development, I would make changes to the parser and the AST type that would break a large proportion of the test cases, quickly...

As for `ppx_alcotest`, I think that should be optional. In some cases, it is nice to have the expected output stored in the test source file like with `ppx_expect`. However,...

This should be potentially more feasible with https://github.com/aantron/luv.

Can you give some more detail on what you would want for the second (non-syntax) part of point 3?