Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

SuggestionBuilder needs to send ascribedType of constructor parameters #6655

Conversation

4e6
Copy link
Contributor

@4e6 4e6 commented May 12, 2023

Pull Request Description

close #6611

Changelog:

  • update: run compiler passes on the ascribedType field of the constructor arguments
  • update: suggestion builder uses the type information attached to ascribedType
  • feat: resolve qualified names in type signatures

Important Notes

Checklist

Please ensure that the following checklist has been satisfied before submitting the PR:

  • The documentation has been updated, if necessary.
  • All code follows the
    Scala,
    Java,
    and
    Rust
    style guides. In case you are using a language not listed above, follow the Rust style guide.
  • All code has been tested:
    • Unit tests have been written where possible.

@4e6 4e6 added the CI: No changelog needed Do not require a changelog entry for this PR. label May 12, 2023
@4e6 4e6 self-assigned this May 12, 2023
Copy link
Member

@JaroslavTulach JaroslavTulach left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Much more Scala-like changes! I assume you will need the same changes as I did overall the code base. Plus you need support for type arguments in constructor - I solved it by adding resolveType additional parameter with type parameter names.

@@ -4723,6 +4722,7 @@ object IR {
def mapExpressions(fn: Expression => Expression): Specified = {
copy(
name = name.mapExpressions(fn),
ascribedType = ascribedType.map(fn),
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I guess this change of IR.DefinitionArgument.Specified.mapExpressions is answering my question:

I was hoping the d.mapExpressions(resolveExpression ...) will iterate over IR.DefinitionArgument, but it doesn't.

Overall the changes in this PR look more Scala-like than my #6584

@4e6 4e6 marked this pull request as ready for review May 12, 2023 14:20
Copy link
Collaborator

@hubertp hubertp left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm missing some tests in TypeSignaturesTest since this change touches that pass. But overall, looks sane.

private def buildArgument(arg: IR.DefinitionArgument): Suggestion.Argument = {
val signatureOpt = arg.name.getMetadata(TypeSignatures).map { meta =>
buildTypeSignature(meta.signature) match {
case Vector(targ) => buildTypeArgumentName(targ)
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This seems to assume that you don't deal with Function, which can return vector of size > 1?

private def resolveDefinitionData(
data: IR.Module.Scope.Definition.Data
): IR.Module.Scope.Definition.Data = {
data.copy(
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We should eventually start using a smarter copy that only copies if data.arguments.map(resolvedArgument) != data.arguments. But that optimization can be done in multiple places so can be done later.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Am I right that transformExpressions does such kind of optimization a bit? My understanding is that transformExpression callback may specify partial match on just a few IR.Expression.SubTypes and only those get duplicated - the rest of the IR tree remains unchanged.

Shall we prefer the use of transformExpressions over copy? Are there any places to perform such a replacement in current code?

start using a smarter copy

Is there such a "smarter copy" already or do we have to write one? Can we create such "no-copy-if-unmodified" copy in a generic way or do we have to manually override each copy implementation or wrap each copy invocation?

Copy link
Member

@radeusgd radeusgd left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Tests look good. I guess a negative test with unknown type could be nice, but if it's already handled by @JaroslavTulach 's PR I guess it's also fine to skip it.

@4e6
Copy link
Contributor Author

4e6 commented May 12, 2023

PR enables the resolution of qualified names in type signatures. At this moment stdlib compilation fails with two errors:

In module Standard.Table.Internal.Excel_Reader:
Compiler encountered errors:
Excel_Reader.enso[33:42-33:68]: The name `ExcelHeaders.HeaderBehavior` could not be found.
In module Standard.Table.Data.Type.Value_Type:
Compiler encountered errors:
Value_Type.enso[135:77-135:84]: The name `SQL_Type` could not be found.
Aborting due to 2 errors and 0 warnings.

@JaroslavTulach
Copy link
Member

JaroslavTulach commented May 13, 2023

Excel_Reader.enso[33:42-33:68]: The name ExcelHeaders.HeaderBehavior could not be found.

An attempt to import Java inner class! Mitigated in 74c6f99 and reported #6678

Value_Type.enso[135:77-135:84]: The name SQL_Type could not be found.

Same problem as in my PR - addressed by e56bb3a

Let's see how far the CI can get...

@JaroslavTulach JaroslavTulach force-pushed the wip/db/6611-suggestionbuilder-needs-to-send-ascribedtype-of-constructor-parameters branch from 11bf895 to 74c6f99 Compare May 13, 2023 05:55
@JaroslavTulach
Copy link
Member

3d73eb7 addresses this failure.

@@ -1,5 +1,6 @@
from Standard.Base import all
import Standard.Base.Errors.Illegal_Argument.Illegal_Argument
import Standard.Database.Data.SQL_Type.SQL_Type
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This makes Standard.Table library dependent on Standard.Database. I'm not sure if we want that because the Database lib already depends on the Table

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

In fact, this caused an issue with IR caching https://github.com/enso-org/enso/actions/runs/4966001323/jobs/8887184345#step:10:3121
What happened is that IR for Standard.Database.Connection.Database module was not cached, and it resulted in compilation error:

INFO ide_ci::program::command: bashℹ️ In module Standard.Database.Connection.Database:
 INFO ide_ci::program::command: bashℹ️ Compiler encountered errors:
 INFO ide_ci::program::command: bashℹ️ Database.enso[23:11-23:28]: The name `Connection_Details` could not be found.
 INFO ide_ci::program::command: bashℹ️ Database.enso[23:33-23:50]: The name `Connection_Options` could not be found.
 INFO ide_ci::program::command: bashℹ️ Database.enso[23:55-23:64]: The name `Connection` could not be found.
 INFO ide_ci::program::command: bashℹ️ Database.enso[23:68-23:76]: The name `SQL_Error` could not be found.
 INFO ide_ci::program::command: bashℹ️ Database.enso[24:25-24:42]: The name `Connection_Options` could not be found.
 INFO ide_ci::program::command: bashℹ️ Database.enso[28:29-28:41]: The name `Single_Choice` could not be found.
 INFO ide_ci::program::command: bashℹ️ Database.enso[30:28-30:33]: The name `Vector` could not be found.
 INFO ide_ci::program::command: bashℹ️ Database.enso[31:63-31:67]: The name `False` could not be found.
 INFO ide_ci::program::command: bashℹ️ Database.enso[35:9-35:14]: The name `Option` could not be found.
 INFO ide_ci::program::command: bashℹ️ Database.enso[36:5-36:17]: The name `Single_Choice` could not be found.
 INFO ide_ci::program::command: bashℹ️ Database.enso[36:27-36:33]: The name `Display` could not be found.
 INFO ide_ci::program::command: bash⚠️ Aborting due to 11 errors and 0 warnings.
 INFO ide_ci::program::command: bashℹ️ Execution finished with an error: Compilation aborted due to errors.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This makes Standard.Table library dependent on Standard.Database

Cyclic dependencies are bad. It'd be good to prevent them with some failure - #6679

In fact, this caused the issue with IR caching

Oh, my dear! I was hunting the problem for almost a week and it is caused by such a little oversight. Thank you very much for working on this PR in parallel with #6584 - it was very refreshing and useful to see different approach to the problem. Without your help I'd still be blindly searching for a bug in BindingsMap.

@4e6 4e6 added the CI: Clean build required CI runners will be cleaned before and after this PR is built. label May 13, 2023
@4e6
Copy link
Contributor Author

4e6 commented May 13, 2023

I can reproduce the issue with IR caches that fails the CI:

  • clean build-distribution and .local/share/enso/cache directories
  • run $ sbt buildEngineDistribution
  • run $ ./built-distribution/enso-engine-0.0.0-dev-linux-amd64/enso-0.0.0-dev/bin/enso --run test/Benchmarks

there's no issues when run with --no-ir-caches flag

@4e6
Copy link
Contributor Author

4e6 commented May 13, 2023

Merging it to unblock @JaroslavTulach

@4e6 4e6 added the CI: Ready to merge This PR is eligible for automatic merge label May 13, 2023
@mergify mergify bot merged commit 7067917 into develop May 13, 2023
@mergify mergify bot deleted the wip/db/6611-suggestionbuilder-needs-to-send-ascribedtype-of-constructor-parameters branch May 13, 2023 18:33
JaroslavTulach added a commit that referenced this pull request May 14, 2023
Procrat added a commit that referenced this pull request May 15, 2023
…z-6260

* develop:
  Build nightlies every day. (#6681)
  Force `newDashboard` default on the CI-built packages. (#6680)
  Verify ascribed types of parameters really exist (#6584)
  SuggestionBuilder needs to send ascribedType of constructor parameters (#6655)
  Improvements to the Table visualization. (#6653)
  Removing flush (#6670)
  Improving widgets for take/drop (#6641)
@hubertp
Copy link
Collaborator

hubertp commented May 15, 2023

I can reproduce the issue with IR caches that fails the CI:

  • clean build-distribution and .local/share/enso/cache directories
  • run $ sbt buildEngineDistribution
  • run $ ./built-distribution/enso-engine-0.0.0-dev-linux-amd64/enso-0.0.0-dev/bin/enso --run test/Benchmarks

there's no issues when run with --no-ir-caches flag

I'm guessing this was the circular dependency thing because I have no issues right. And yes, I'm able to reproduce the problems when I introduce that import.

Procrat added a commit that referenced this pull request May 16, 2023
* develop:
  Build nightlies every day. (#6681)
  Force `newDashboard` default on the CI-built packages. (#6680)
  Verify ascribed types of parameters really exist (#6584)
  SuggestionBuilder needs to send ascribedType of constructor parameters (#6655)
  Improvements to the Table visualization. (#6653)
  Removing flush (#6670)
  Improving widgets for take/drop (#6641)
  disable authentication and newDashboard by default (#6664)
  Add COOP+COEP+CORP headers (#6646)
@enso-bot enso-bot bot mentioned this pull request Jun 1, 2023
2 tasks
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
CI: Clean build required CI runners will be cleaned before and after this PR is built. CI: No changelog needed Do not require a changelog entry for this PR. CI: Ready to merge This PR is eligible for automatic merge
Projects
None yet
Development

Successfully merging this pull request may close these issues.

SuggestionBuilder needs to send ascribedType of constructor parameters
5 participants