Fix imports being unsilenced when checking stale SCCs #2037
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This commit fixes a bug where silenced imports were being unsilenced after incremental mode + silent_imports was run several times after forcing an SCC to be rechecked.
|
Whee! Don't worry about writing that test for now. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This pull request fixes an edge case where previously silenced imports were being un-silenced when using silent imports with incremental mode.
Unfortunately, I wasn't able to provide a test case because this problem seems to manifest mainly in larger code bases (with import cycles?) that are hard to simplify and after running incremental mode at least 3 times -- our test suite can currently handle only three. See #2036 for more details about the bug.
The problem ended up being with the order in which the
parse_file(...)andload_graph(...)functions are called -- the former function will mark imports with the# type: ignoredirectives as suppressed and the latter function corrects the dependencies computed byparse_fileand will mark silenced modules as suppressed.The problem is that when running mypy with a cold cache, the
parse_filefunction will be executed at the start of theload_graphfunction, but when checking a stale SCC, the logic will end up makingparse_filebe executed after theload_graphfunction.This is problematic since
parse_filewill potentially clobber over the (often correct) information regarding suppressed and unsuppressed dependencies that was loaded from the old cache files.This fix adds an extra pass over the dependencies after calling
parse_filewhen checking stale SCCs to correct the clobbering.