[doc-han]

Short Description

When running findAllDanglingIdentifiers in the add-imports transform, we want to ignore top-level object declarations.

Fixes #1029

Implementation Details

A more detailed breakdown of the changes, including motivations (if not provided in the issue).

QA Notes

List any considerations/cases/advice for testing/QA here.

AI Usage

Please disclose how you've used AI in this work (it's cool, we just want to know!):

• Code generation (copilot but not intellisense)
• Learning or fact checking
• Strategy / design
• Optimisation / refactoring
• Translation / spellchecking / doc gen
• Other
• I have not used AI

You can read more details in our Responsible AI Policy

Release branch checklist

Delete this section if this is not a release PR.

If this IS a release branch:

• Run pnpm changeset version from root to bump versions
• Run pnpm install
• Commit the new version numbers
• Run pnpm changeset tag to generate tags
• Push tags git push --tags

Tags may need updating if commits come in after the tags are first generated.

[josephjclark]

Perfect, that's exactly what I'm expecting 👍

[josephjclark]

Dang, something feels off here.

Running this against my (refactored) 1.6mb source file, I'm still seeing 230mb of memory get used up. I can see that the lazy state vistor is exiting early - but there's still too much tree traversal going on.

The changes here look good and I'm happy, but we'll need to investigate a bit more before we can declare victory on this particular problem :(

[josephjclark]

Here's a much better trace output for parsing with the new changes:

[pre acorn parse] Used heap at 58.79mb
[post acorn parse] Used heap at 157.66mb

[transformer lazy-state] Used heap at 158.27mb
transformer lazy-state: 1.743ms

[transformer promises] Used heap at 202.67mb
transformer promises: 82.818ms

[transformer ensure-exports] Used heap at 202.74mb
transformer ensure-exports: 0.662ms

[transformer top-level-operations] Used heap at 249.66mb
transformer top-level-operations: 68.538ms

[transformer add-imports] Used heap at 249.69mb
transformer add-imports: 0.334ms

[finished] Used heap at 249.70mb

For the record, here's what it's like without the optimisation:

[pre acorn parse] Used heap at 60.44mb
[post acorn parse] Used heap at 154.56mb

[transformer lazy-state] Used heap at 201.88mb
transformer lazy-state: 91.469ms

[transformer promises] Used heap at 234.78mb
transformer promises: 99.671ms

[transformer ensure-exports] Used heap at 234.86mb
transformer ensure-exports: 0.67ms

[transformer top-level-operations] Used heap at 234.95mb
transformer top-level-operations: 0.486ms

[transformer add-imports] Used heap at 234.97mb
transformer add-imports: 0.196ms

[finished] Used heap at 234.97mb

I can see that lazy state and promises have improved - but promises and top level ops are still kinda expensive. You were talking about top level ops @doc-han maybe we do need to look if there's a quick win there

[josephjclark]

@doc-han I think I can see two cheap fixes here:

1. Flip top-level-operations to start from Program and do its own traversal - see compiler: optimise top level operations transformer #1055
2. Do the same with promises as we've done for imports: visit call expressions and abort if they're at the top level