[mypyc] Record MRO ancestor modules as deps of subclass-defining modules#21743
Open
geooo109 wants to merge 4 commits into
Open
[mypyc] Record MRO ancestor modules as deps of subclass-defining modules#21743geooo109 wants to merge 4 commits into
geooo109 wants to merge 4 commits into
Conversation
Fixes python#21742 Mypy's incremental system only recompiles a module when a recorded dependency changes, and a dependency edge to an ancestor's module is created in cases such as an explicit import or a checked expression whose type references that ancestor. A bare subclass definition produces neither, so the transitive ancestors never enter the module's dependency graph. The goal of this PR is to make the dependency graph reflect what the generated C actually depends on. When compiling with mypyc, a module that defines a class now records the modules defining every ancestor in that class's MRO as indirect dependencies, so an ancestor's interface change reaches the subclass-defining module directly.
geooo109
force-pushed
the
mypyc-class-ancestor-deps
branch
from
July 17, 2026 12:09
5f4468f to
c3b4e51
Compare
This comment has been minimized.
This comment has been minimized.
p-sawicki
reviewed
Jul 20, 2026
Comment on lines
+3538
to
+3557
| def compiled_class_ancestor_refs(self) -> set[str]: | ||
| """Modules defining MRO ancestors of classes defined in this module. | ||
|
|
||
| Only used for mypyc-compiled modules: the generated C for a class | ||
| (vtable arrays, getter/setter tables, object struct) references | ||
| every inherited method/attribute of every ancestor, including | ||
| ancestors defined in modules this module does not import directly. | ||
| Recording them as indirect dependencies makes an ancestor's | ||
| interface change re-trigger type checking (and hence C regeneration) | ||
| of this module. Only top-level classes are scanned: mypyc rejects | ||
| nested class definitions. | ||
| """ | ||
| assert self.tree is not None | ||
| mods: set[str] = set() | ||
| for sym in self.tree.names.values(): | ||
| node = sym.node | ||
| if isinstance(node, TypeInfo) and node.module_name == self.id: | ||
| for ancestor in node.mro[1:]: | ||
| mods.add(ancestor.module_name) | ||
| return mods |
Collaborator
There was a problem hiding this comment.
could this be moved to get_additional_deps in mypyc/codegen/emitmodule.py? if the dependencies are only needed by mypyc then it would be better to keep this out of the top-level mypy logic, since mypyc functions as a plugin.
Author
There was a problem hiding this comment.
I did a refactor. I think we can't add this inside the get_additional_deps because we want to apply it after the analysis. So, I created a new method for this. Does this approach look good, or is there a simpler solution?
geooo109
commented
Jul 20, 2026
| """ | ||
| return [] | ||
|
|
||
| def get_additional_indirect_deps(self, file: MypyFile) -> set[str]: |
Author
There was a problem hiding this comment.
Is this naming ok? or We should make it more precise on which phase the method is being called.
Contributor
|
According to mypy_primer, this change doesn't affect type check results on a corpus of open source code. ✅ |
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.
Fixes #21742
Mypy's incremental system only recompiles a module when a recorded dependency changes, and a dependency edge to an ancestor's module is created in cases such as an explicit import or a checked expression whose type references that ancestor. A bare subclass definition produces neither, so the transitive ancestors never enter the module's dependency graph.
The goal of this PR is to make the dependency graph reflect what the generated C actually depends on. When compiling with mypyc, a module that defines a class now records the modules defining every ancestor in that class's MRO as indirect dependencies, so an ancestor's interface change reaches the subclass-defining module directly.