[new release] Five Ortac packages 0.4.0

[n-osborne]

This release brings a number of new features and improvements:

• New features:
  • #247: Generated tests cover now functions with multiple SUT arguments
  • #253: Generated tests cover now functions returning SUT values
  • #259: Generated tests can be run in a separated process with a timeout
• Improvements:
  • #245: Fix the analysis of function signature to explicitly not support SUTs inside another type
  • #251: Fix the display of the runnable scenario for value returned by a function that could have raised an exception

[mseri]

Does
ortac-qcheck-stm.0.2.0 (failed: This expression has type ('a, STM.res) Either.t) require an upper bound on ortac 0.4.0?

[jmid]

Indeed. A few of the with-test dependencies need to be version-constrained too, to avoid such cross-version combinations.

[n-osborne]

True. Thanks @jmid for the updates 😃

[jmid]

Yikes, this is still failing, e.g. with

ortac-core.0.4.0
  revdeps
    ortac-qcheck-stm.0.2.0 (failed: This expression has type ('a, STM.res) Either.t)

I think we need to separately constrain 0.2 from installing with 0.4.
In ortac-qcheck-stm/ortac-qcheck-stm.0.2.0/opam for example, I can see

depends: [
..
  "ortac-core" {>= "0.2.0"}
  "ortac-runtime" {with-test}
  "ortac-runtime-qcheck-stm" {with-test}
  "odoc" {with-doc}
]

which could benefit from = version constraints too!
This is probably best done in a separate PR though...

[n-osborne]

I was thinking the same and in the process of adding a commit to this PR with stricter dependencies constraints (next release of Gospel is also expected to bring breaking changes).
But I'm happy to put these changes in another PR.

[n-osborne]

#26697 should improve the situation

[jmid]

#26697 and #26702 have now been merged. Can you push a rebased version? I'm keen to see if the fixes have solved the red lights... 😃

[jmid]

All green and good to go!

[mseri]

Thanks!