-
-
Notifications
You must be signed in to change notification settings - Fork 34.5k
gh-131798: Narrow the return type of _FORMAT_SIMPLE and _FORMAT_WITH_SPEC to str
#146639
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
base: main
Are you sure you want to change the base?
Changes from 2 commits
c1b9a83
ed1204e
d09a7cc
3dea0df
253f55b
0efc0b5
File filter
Filter by extension
Conversations
Jump to
Diff view
Diff view
There are no files selected for viewing
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,2 @@ | ||
| Allow the JIT to remove unicode guards after ``_FORMAT_SIMPLE`` and | ||
| ``_FORMAT_WITH_SPEC`` when the input type is a known built-in type. |
Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.
| Original file line number | Diff line number | Diff line change |
|---|---|---|
|
|
@@ -264,6 +264,25 @@ _Py_uop_sym_get_const_as_stackref(JitOptContext *ctx, JitOptRef sym) | |
| return PyStackRef_FromPyObjectBorrow(const_val); | ||
| } | ||
|
|
||
| /* | ||
| Indicates whether the type is a known built-in type | ||
| that is safe to narrow. | ||
| */ | ||
| bool | ||
| _Py_uop_sym_is_safe_type(JitOptRef sym) | ||
| { | ||
| PyTypeObject *typ = _Py_uop_sym_get_type(sym); | ||
| if (typ == NULL) { | ||
| return false; | ||
| } | ||
| return (typ == &PyLong_Type) || | ||
| (typ == &PyUnicode_Type) || | ||
| (typ == &PyFloat_Type) || | ||
| (typ == &_PyNone_Type) || | ||
| (typ == &PyBool_Type) || | ||
| (typ == &PyFrozenDict_Type); | ||
|
Member
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. Oops, I meant to factor this out into a common function, sorry!
Contributor
Author
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. You mean factoring safe types into a static helper? No problem. |
||
| } | ||
|
|
||
| /* | ||
| Indicates whether the constant is safe to constant evaluate | ||
| (without side effects). | ||
|
|
||
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This can return a subclass, e.g.
Does the
sym_new_type(ctx, &PyUnicode_Type);not guarantee we have an exact type?There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Good catch! I oversimplified this. Narrowing type for
formatisn't as straightforward as I thought.That said, I'm curious why
FORMAT_SIMPLEneeds to preserve str subclasses. Had a look through the git history it's been this way since 2015, and seems not a deliberate type-preservation contract.Current behaviour is already inconsistent:
Multi-piece concatenation already strips it (both
_PyUnicodeWriterinstr.formatandBUILD_STRINGin f-strings create exactstr).If
FORMAT_SIMPLEnormalised to exactstr, this inconsistency goes away and type narrowing becomes unconditionally correct. Would that be a reasonable change, or is there a reason to preserve str subclass identity through formatting that I'm missing?There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Huh that's kinda scary. The semantics look a little strange, so maybe we should leave this alone for now?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Fair enough. I just noticed the inconsistency and thought it was worth raising.
For the narrowing itself: we could do conditional narrowing for input types where we can prove
__format__returns exactstr(e.g.int,float). But honestly it feels not pythonic to hardcode a list of "safe" types for a uop that's less then 0.1%. I'd prefer just closing this PR unless you think it's worth keeping with that approach?There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Actually we already have something like that called sym_is_safe_type, you can use that
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I couldn't find
sym_is_safe_typeanywhere. The closest thing I can see issym_is_safe_const, but that checks for known constant values rather than types. Have I overlooked something?There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Oh yeah sorry that's the one! Just factor out the type checks in
_Py_uop_sym_is_safe_constinto another function and use that in this case, I think that's fine!There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Done. How does it look now?