Remove > and >= methods on Py to reduce invalidations - #829
Closed
MilesCranmerBot wants to merge 1 commit into
Closed
MilesCranmerBot wants to merge 1 commit into
MilesCranmerBot wants to merge 1 commit into
Conversation
Base only defines > and >= as generic fallbacks (y < x, y <= x), so adding these methods invalidated thousands of method instances elsewhere (6830 -> 2532 unique instances when loading PythonCall + SymbolicRegression on Julia 1.12). Dropping them routes x > y through Base's fallback to the existing < and <= methods, so results and the Py return type are unchanged. The only behavior change is that Python now tries y.__lt__ before x.__gt__. pygt and pyge remain exported and are still used by the @py macro. Refs JuliaPy#828 Co-authored-by: Miles Cranmer <miles.cranmer@gmail.com>
Contributor
Author
|
Closing in favor of #830, which is the same change from the properly named branch (and supersedes this early duplicate). |
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Removes the
>and>=methods onPy(forPy,Py,Py,Number, andNumber,Py), as proposed in #828 (comment).Why. Base defines
>and>=only as generic fallbacks (y < x,y <= x), so calls likex > 1with an abstractly typedxget inferred against PythonCall's methods, and defining them invalidates thousands of method instances elsewhere. Measured with SnoopCompile on Julia 1.12.7 (loadingPythonCall+SymbolicRegression): 6830 → 2532 unique invalidated method instances, ~63% fewer from PythonCall.Behavior.
x > ynow evaluates as Pythony < x, i.e. Python triesy.__lt__beforex.__gt__— only observable for types where__gt__and__lt__disagree. Results and thePyreturn type are otherwise unchanged; verified locally on Julia 1.10 with the fallbacks resolving to the existing</<=methods (Py(3) > 2,2 > Py(3),Py(3) >= 3,4 >= Py(3),Py(3) > Py(1),Py(3) > 2.5, …).pygt/pygeremain exported and are still used by the@pymacro.!=is intentionally untouched on 0.9: its Base fallback!(x == y)requires!(::Py), which does not exist (details in #828). The full fix remains thev1branch (#702, #709).Refs #828