Remove legacy daml script

[samuel-williams-da]

Resolves #20639

• 8bafdb8:

  • Remove daml and scala code for daml script "v1"
  • Renames daml3-script to daml-script
  • Update expandSdkPackages to accept daml3-script and daml-script-lts as simply daml-script, with respective warnings
  • Adds a warning to the runner when attempting to run a legacy daml script (i.e. <=3.2 daml-script), by checking the Module Path of the Script monad (since the package name has been changed)
  • Update release script for daml-libs to only try to include daml-script
  • daml-sdk-head works
  • All tests under compiler/damlc are passing
  • Found a bug, errors from daml3-script submit do not correct refer to the source definition, as such cases matching Script execution failed on commit at <somewhere> have been loosened. To be fixed in a later commit and test changes reverted.
  • Another annoying note, the Submit(.daml) tests use contract keys, so can only run under 2.dev. Should consider splitting them up so we can run a subset of them in 2.1

• b22abff:

  • Selective replacement of daml3-script with daml-script project wide
  • Turns out the issue with source locations was a missing replacement from above, so that is resolved with this commit
  • Noticed that all the expected files have had whitespace added. Going to revert that in next commit to get changed lines down a bit

• 3d32a86:

  • Revert all *.EXPECTED.ledger files to their state on master, to reduce changes in this PR

• 37a616d:

  • Buildifier >:(
  • Tests under daml-script are failing, as still split into daml2-script and daml3-script runners, address in next commit

• b603bb8:

  • Fix tests under daml-script by removing Daml2ScriptRunnerTest, and dropping the 3 from the various Daml3Script tests
  • Also needed to update an error assertion as daml3-script uses engine.free, over its own free implementation

• f135c2c:

  • LSP cancellation test was failing, as once moved over to new script implementation, cancellation on long running pure computations wasn't working, and was delaying the cancellation request until after the computation finished. This was fixed by updating Remy's Free machine to calculate the initial Free monad (before script interpretation) in a Future, and having it check for cancellation there as well.

• 263dc4e:

  • Update Submit.daml test to use ifdef so it can run in 2.1, This required adding if= to diagnostic expectations so they can be gated by defs

• ef1de7a:

  • Update hoogle and haddock templates for daml script

• 58819c3:

  • Move the legacy script check to the runner, so it can use the same error types
  • Add test for this

• d04c037:

  • Test was uploading dars every time, and previous test uploaded the real daml-script. Naturally the upload of the fake daml-script after that fails, as its not a valid upgrade. Updated the test to not upload the fake daml-script, as it's not needed.

TODO:

• Ensure documentation logic works
  • Hard to know until docs consumes it
• Account for canton issues
  • Sounds like a Monday Sam problem

[samuel-williams-da]

We can use an ifdef here and keep it running in 2.1

[samuel-williams-da]

Outdated

[samuel-williams-da]

Same with line below

[dylant-da]

Reviewed on a call - LGTM modulo comments

[dylant-da]

LGTM