Optimize Denotation#current #9578
Conversation
Have a separate stack of type comparers instead.
Split into smaller parts
This was observed in TabCompleterTests. The problem in general is that we might see symbols in runs that have fewer phases than the phase at which the symbol was created. Previously, we updated the symbol anyway. But this is problematic since it means the symbol has a validity period that does not correspond to a phase in the current run. We now treat those symbols as stale without a way to recover instead.
There was a problem hiding this comment.
Hello, and thank you for opening this PR! 🎉
All contributors have signed the CLA, thank you! ❤️
Commit Messages
We want to keep history, but for that to actually be useful we have
some rules on how to format our commit messages (relevant xkcd).
Please stick to these guidelines for commit messages:
- Separate subject from body with a blank line
- When fixing an issue, start your commit message with
Fix #<ISSUE-NBR>:- Limit the subject line to 72 characters
- Capitalize the subject line
- Do not end the subject line with a period
- Use the imperative mood in the subject line ("Add" instead of "Added")
- Wrap the body at 80 characters
- Use the body to explain what and why vs. how
adapted from https://chris.beams.io/posts/git-commit
Have an awesome day! ☀️
|
test performance please |
|
performance test scheduled: 1 job(s) in queue, 1 running. |
|
Performance test finished successfully: Visit http://dotty-bench.epfl.ch/9578/ to see the changes. Benchmarks is based on merging with master (f46c030) |
|
test performance please |
|
performance test scheduled: 1 job(s) in queue, 1 running. |
|
Performance test finished successfully: Visit http://dotty-bench.epfl.ch/9578/ to see the changes. Benchmarks is based on merging with master (64a239f) |
| initial.validFor.firstPhaseId.max(typerPhase.id) | ||
| if firstPhaseId > ctx.lastPhaseId then | ||
| false | ||
| else if (initial ne denot) || ctx.phaseId != firstPhaseId then |
There was a problem hiding this comment.
If the phases have changed (e.g., adding or removing some phases in the middle), then the periods of denotations will be incorrect in the new run. Therefore, do we need to invalidate all non-permanent denotations in such cases, as the same phaseId does not mean the same phase?
There was a problem hiding this comment.
Yes, that looks like a problem. But can phases change like this in the same compiler instance? In that case I don't see how we can re-use any symbols. So it might be better to just start a new compiler instance in that case.
Based on #9405