Chore: Lin DSL test renaming

[jmid]

This PR is the sister PR to #396, eliminating historical occurrences of DSL and dsl in file and test names.
Together, the effect of the two PRs is the following renaming:

lin_tests.ml      --> lin_internal_tests.ml      # for Lin_internal tests
lin_tests_dsl.ml  --> lin_tests.ml               # for Lin tests

The DSL name is historic and was eliminated before our initial opam package release.
As Lin is the primary user-facing library for sequential consistency testing, I think this is preferable.
It should be more intuitive to locate the right test with these names.

If done right, the PR should have no effect on test suite functionality... (I'm already seeing lots of weird red CI lights though! 😬 )

While at it, I also spotted a few tests that were missing from the overview in src/README.md.

[jmid]

I just restarted a bunch of jobs as roughly 1/2 of the GitHub actions workflows failed after seconds, due to some internal glitch I suppose. The restarted runs seem to progress fine though.

[jmid]

CI summary:

• Linux 5.1 debug triggered an assertion failure in threadomain [ocaml5-issue] Assertion failure on threadomain under debug runtime #404
• Linux trunk debug triggered an assertion failure in thread_joingraph [ocaml5-issue] Abort / crash on thread_joingraph and thread_createtree under debug runtime #353
• MinGW 5.1 timeout in threadomain [ocaml5-issue] Windows failures on threadomain #203
• MinGW bytecode 5.1 timeout in threadomain [ocaml5-issue] Windows failures on threadomain #203
• Cygwin trunk / part1 timeout during 213.3s negative Lin Dynlink test with Domain shrink after multiple >1000s runs [ocaml5-issue] General Cygwin slowness #384 [ocaml5-issue] Deadlock in Dynlink test on Cygwin+MinGW+MSVC #307
  • 1268.8s Domain.spawn/join - atomic,
  • 3103.6s domain_spawntree - with Atomic
  • 1855.2s Lin Atomic test with Domain
  • 4690.8s Lin Bigarray.Array1 (of ints) test with Domain
  • 1930.3s Lin Bytes test with Domain

Out of 63 workflows (incl. 10 opam install ones) 5 failed, of which 4 were genuine and 1 was a borderline case

[jmid]

CI summary for merge to main:

• MinGW bytecode trunk threadomain crashed [ocaml5-issue] Windows failures on threadomain #203

Out of 38 workflows, 1 failed - with a genuine failure