Petr Viktorin added the comment:
I haven't forgotten this issue, but when I get to it it always leads to a
rabbit hole. Sometimes just chasing refleaks, but there are deeper issues as
well.
AFAICS, there's no way to call metatype.tp_new for such a class. I guess the
safest op
Petr Viktorin added the comment:
Interned strings were broken in GH-20058, see bpo-46006. Maybe that's also the
issue here?
--
nosy: +vstinner
___
Python tracker
<https://bugs.python.org/is
Petr Viktorin added the comment:
That was a fix for GH-17350, which might need to be reverted as well.
Victor, could you take another look at GH-17350? I must admit I (still) don't
understand this change; what would break if it was reverted (along with the
fixup from bpo-
Petr Viktorin added the comment:
I don't see how instantiating a metaclass with non-default tp_new would work if
you don't know some details about the specific metaclass. So IMO we can we
limit ourselves to scenarios where either:
1) the metaclass uses default tp_new, or
2) the
Petr Viktorin added the comment:
Nice! It's starting to look reasonable, I'll try an implementation when I get
some focus time. (Sadly I can't promise it'll be this year.)
Just one detail:
> A flag probably can do the same. I have no preference, `ApplySpec` seems
&
Petr Viktorin added the comment:
The docs for PyFrame_GetCode say it's returning an "int".
https://docs.python.org/3/c-api/reflection.html#c.PyFrame_GetCode
Same for PyFrame_GetBack. Are there more?
"Get the frame code." is not very clear. Could this link to
https:
Petr Viktorin added the comment:
> Personally, I'd rather we not revert the original change.
Even in 3.10? Your PR looks pretty heavy for a bugfix release, and won't apply
cleanly to 3.10.
> Moving the data to _PyRuntimeState would save me some effort with related
> w
Petr Viktorin added the comment:
> It would perhaps be useful if the os.path.expanduser call was added in the
> cache_from_source() function in importlib?
No: that would mean Python would be doing work that the shell should do.
Possibly it would be duplicating the work. Possibly it wo
Change by Petr Viktorin :
--
keywords: +patch
nosy: +petr.viktorin
nosy_count: 2.0 -> 3.0
pull_requests: +28660
stage: -> patch review
pull_request: https://github.com/python/cpython/pull/30456
___
Python tracker
<https://bugs.p
Petr Viktorin added the comment:
Here's a patch, but testing it would require putting a file in the root
directory. I don't dare do that from the test suite. Is one-time manual testing
OK here?
--
___
Python tracker
<https://bu
Petr Viktorin added the comment:
> The problem is that this change fixed another bug, well, see: bpo-44050.
> While a revert should fix win_py399_crash_reproducer.py, it will reintroduce
> bpo-44050 bug.
bpo-44050 is an attempt to fix a regression introduced in bpo-38858, perh
Petr Viktorin added the comment:
Another proposed alternative is to make these immortal, and share them.
There were some discussions about this, but no PEP to capture the pros/cons of
the alternatives and the decision.
--
nosy: +petr.viktorin
Petr Viktorin added the comment:
This cannot be changed in 3.x, since the PyObject_DelAttr macro calls
PyObject_SetAttr(..., NULL), and the macro is expanded in all extensions that
use the stable ABI.
(Technically, it would be possible to add a PyObject_SetAttr *macro* that would
warn/fail
Petr Viktorin added the comment:
If we have static types, that means there is a mechanism to share some objects
across interpreters.
And if that's the case, why can't small ints (like
sys.float_info.n_unnamed_fields) be static & shared as well?
--
nosy:
New submission from Petr Viktorin :
I want to test a web application by crawling every reachable page. If an error
occurs, I need to keep track of the page the error occured at (and additional
info like what links were followed to get to the page, so a `__note__` string
isn't enough).
Petr Viktorin added the comment:
Changing the existing functions is a no-go for backwards compatibility reasons.
I think the best way forward would be to add a new function, and then possibly
deprecate the old one if it's deemed dangerous.
If you want to push this forward, coul
Petr Viktorin added the comment:
Thanks for looking into it!
> If you don't define derive the superclass constructor is used, which means
> you get something of type ExceptionGroup, not your subclass.
That might be fine in my case (for a MVP at least). Is there any other dan
Petr Viktorin added the comment:
> can you just assign the task to a field on the exception other than __note__?
That might work, but I'm afraid of touching namespaces I don't own. If the
subclass is feasible, I'd r
Petr Viktorin added the comment:
I'd post it to capi-sig, or to the existing thread on python-dev. But here's a
good place too, especially if you want feedback from a smaller group of people
first.
--
___
Python tracker
<https://bu
New submission from Petr Viktorin :
I'm looking at the _PyType_GetModuleByDef optimization in
https://github.com/python/cpython/pull/25504/files -- previously I assumed it's
OK since it passed review.
But it doesn't look correct:
- in the `_PyType_HasFeature` assert, we shoul
Change by Petr Viktorin :
--
keywords: +patch
pull_requests: +28894
stage: -> patch review
pull_request: https://github.com/python/cpython/pull/30696
___
Python tracker
<https://bugs.python.org/issu
Petr Viktorin added the comment:
> Sharing objects between interpreters is bad
That's your opinion, I don't necessarily share it.
> and is causing complex bugs.
But converting static things (types, small ints) to heap is also causing bugs :(
Anyway, for this issue: i
Petr Viktorin added the comment:
I guess at least there should be a warning about this in dataclasses docs?
The reproducer with dataclasses (which exhibits the same error on 3.10 and
3.11):
import dataclasses
@dataclasses.dataclass(slots=True)
class A:
pass
@dataclasses.dataclass
Petr Viktorin added the comment:
> Is there a reason to not add these fields to PyTypeObject?
Can't say I haven't thought of that, but AFAIK it would mean breaking the C API
substantially. Even if not it'd be a PEP-sized change, IMO.
--
no
Petr Viktorin added the comment:
Yeah, that looks like it's for some long-forgotten compiler that didn't
implement `signed char` at all. 1994 was a fun time, apparently.
--
nosy: -miss-islington
___
Python tracker
<https://bu
Change by Petr Viktorin :
--
pull_requests: +29101
pull_request: https://github.com/python/cpython/pull/30922
___
Python tracker
<https://bugs.python.org/issue45
Petr Viktorin added the comment:
Embarassingly, it seems it bust needs regen-importlib, at least for 3.10
--
___
Python tracker
<https://bugs.python.org/issue45
Petr Viktorin added the comment:
> The first and last test should use assertAlmostEqual with a suitable
> tolerance (the default tolerance is probably fine).
The merged PR only added tolerance to the last test. On some architectures, the
first test still fails.
(No one is to blame -
Petr Viktorin added the comment:
New changeset aa78287bc6d1c4fc07ee134642eb72db67b771a0 by Karolina Surma in
branch 'main':
bpo-44734: Fix floating point precision in test_turtle (GH-30910)
https://github.com/python/cpython/commit/aa78287bc6d1c4fc07ee134642eb72
Change by Petr Viktorin :
--
stage: patch review -> resolved
status: open -> closed
___
Python tracker
<https://bugs.python.org/issue44734>
___
___
Pyth
Change by Petr Viktorin :
--
nosy: +petr.viktorin
nosy_count: 2.0 -> 3.0
pull_requests: +29215
pull_request: https://github.com/python/cpython/pull/31032
___
Python tracker
<https://bugs.python.org/issu
Petr Viktorin added the comment:
> I would hope that competent C programmers would know not to read or write to
> undocumented fields. But if they come from a Python background, that might
> not be obvious.
Tread carefully when changing decades-old API, documented or not. The docs
Change by Petr Viktorin :
--
pull_requests: +29260
pull_request: https://github.com/python/cpython/pull/31076
___
Python tracker
<https://bugs.python.org/issue45
Change by Petr Viktorin :
--
resolution: -> fixed
stage: patch review -> resolved
status: open -> closed
___
Python tracker
<https://bugs.python.or
New submission from Petr Viktorin :
_PyType_GetModuleByDef (added in bpo-42100)allows module state access
from slot methods (like tp_init or nb_add),
the main thing missing from PEP 573 (Module State Access from C
Extension Methods).
It's time to make it public.
The function itself c
Change by Petr Viktorin :
--
keywords: +patch
pull_requests: +29264
stage: -> patch review
pull_request: https://github.com/python/cpython/pull/31081
___
Python tracker
<https://bugs.python.org/issu
Petr Viktorin added the comment:
Since this is a user-visible change in 3.11, could you add a What's New entry?
--
nosy: +petr.viktorin
___
Python tracker
<https://bugs.python.org/is
Petr Viktorin added the comment:
Almost. It's a bugfix so it needs backports to 3.10 & 3.9.
Thanks for the reminder!
I should get to them next week.
--
status: pending -> open
___
Python tracker
<https://bugs.python
Petr Viktorin added the comment:
Adding new C99 features needs a change in PEP 7
(https://www.python.org/dev/peps/pep-0007/#c-dialect)
--
nosy: +petr.viktorin
status: closed -> open
___
Python tracker
<https://bugs.python.org/issu
Petr Viktorin added the comment:
> IMO this PEP is outdated for a long time.
It is not. Even if it is, it should be marked as such, and that is not a
decision that should be done in this issue.
Please, don't break the rules because you think they're outdated.
> Well, if you
Petr Viktorin added the comment:
> I think we want the scheme to be static and accessible on all platforms, like
> the others. So it probably should be 'nt_venv' and 'posix_venv' with
> additional/improved logic to help apps determine when they need each.
Why
Change by Petr Viktorin :
--
pull_requests: +29429
stage: backport needed -> patch review
pull_request: https://github.com/python/cpython/pull/31262
___
Python tracker
<https://bugs.python.org/issu
Petr Viktorin added the comment:
Just 3.10, after all. 3.9 doesn't have the function yet.
I did the backport, but I'd welcome a review by a fresh set of eyes!
--
versions: -Python 3.9
___
Python tracker
<https://bugs.python.o
Petr Viktorin added the comment:
New changeset 8b8673fe940c4ebc4512bff5af180b66def3d1ae by Petr Viktorin in
branch '3.10':
[3.10] bpo-46433: _PyType_GetModuleByDef: handle static types in MRO (GH-30696)
(GH-31262)
https://github.com/python/cpyt
Change by Petr Viktorin :
--
resolution: -> fixed
stage: patch review -> resolved
status: open -> closed
___
Python tracker
<https://bugs.python.or
Petr Viktorin added the comment:
New changeset 204946986feee7bc80b233350377d24d20fcb1b8 by Petr Viktorin in
branch 'main':
bpo-46613: Add PyType_GetModuleByDef to the public API (GH-31081)
https://github.com/python/cpython/commit/204946986feee7bc80b233350377d2
Petr Viktorin added the comment:
It's in public API, adding it to limited is still pending.
--
___
Python tracker
<https://bugs.python.org/issue46613>
___
___
New submission from Petr Viktorin :
In main, cpython/pystate.h newly includes stdbool.h, providing a definition for
`bool` that might be incompatible with other software.
See here: https://github.com/cmusphinx/sphinxbase/pull/90
Eric, is this necessary? Would an old-school `int` do?
Or
Petr Viktorin added the comment:
Yes, stdbool.h is essentially a backport of C++'s bool type to C.
--
___
Python tracker
<https://bugs.python.org/is
Petr Viktorin added the comment:
It is in C99, but in an optional header.
IMO, including the header in the internals is perfectly OK, but opting everyone
else in isn't very nice.
i.e. it would probably be OK to use `_Bool`, the actual C99 keyword that `bool`
is defined as. But that&#
Petr Viktorin added the comment:
I assume a PR review should be enough.
--
components: +Library (Lib) -Interpreter Core
___
Python tracker
<https://bugs.python.org/issue46
Petr Viktorin added the comment:
So, this will break Cython and gevent, but (unlike the optimization work that
broke f_code/f_frame) it won't provide any value to users?
I don't see how that's a good idea.
--
___
Python
Change by Petr Viktorin :
--
keywords: +patch
pull_requests: +29674
stage: -> patch review
pull_request: https://github.com/python/cpython/pull/31553
___
Python tracker
<https://bugs.python.org/issu
Petr Viktorin added the comment:
OK, looking at it more carefully, it makes sense to do the change.
--
___
Python tracker
<https://bugs.python.org/issue46
Petr Viktorin added the comment:
New changeset 2c228a7b8f89e9ed8d390370abd771d4993b79d8 by Petr Viktorin in
branch 'main':
bpo-46748: Don't import in public headers (GH-31553)
https://github.com/python/cpython/commit/2c228a7b8f89e9ed8d390370a
Petr Viktorin added the comment:
Thank you, Kumar & Victor, for fixing up the issue!
I meant to check the buildbots before closing the issue, but got side-tracked.
--
stage: patch review -> resolved
status: pending -> closed
___
Pyth
Petr Viktorin added the comment:
The current fix skips all of the tests, rather than just the parts that fail.
--
nosy: +petr.viktorin
resolution: fixed ->
status: closed -> open
___
Python tracker
<https://bugs.python.org/i
Petr Viktorin added the comment:
Steve, could you take a look at the PR discussion and chime in about how this
should be done?
--
___
Python tracker
<https://bugs.python.org/issue45
New submission from Petr Viktorin :
In the yield statement documentation
(http://docs.python.org/3.0/reference/simple_stmts.html#the-yield-statement),
the old way of calling generators is used:
"The body of the generator function is executed by calling the
generator’s next() method repea
Petr Viktorin added the comment:
New changeset 591f6754b56cb7f6c31fce8c22528bdf0a99556c by Taneli Hukkinen in
branch 'main':
bpo-40059: Add tomllib (PEP-680) (GH-31498)
https://github.com/python/cpython/commit/591f6754b56cb7f6c31fce8c22528bdf0a99556c
--
nosy: +pet
Petr Viktorin added the comment:
The PR is merged and buildbots are green. Thank you to everyone who helped!
Now would be a good time to bikeshed wording in the documentation.
>From the PR:
> Would it be good to mention in the docs why load() takes only binary files?
> The
Change by Petr Viktorin :
--
assignee: -> docs@python
components: +Documentation
nosy: +docs@python
___
Python tracker
<https://bugs.python.org/issu
Petr Viktorin added the comment:
New changeset 48d926269963cfe7a49c0a4f34af4fe9b832399b by Miro Hrončok in
branch 'main':
bpo-45413: Define "posix_venv", "nt_venv" and "venv" sysconfig installation
schemes (GH-31034)
Petr Viktorin added the comment:
Thanks for the report! You're right that this is misleading. I'll clarify the
docs for this and other structs.
- struct PyTypeObject is part if the limited API.
- its fields and size are not part of the API or
New submission from Petr Viktorin :
importlib.resources.files is documented [0] as returning
importlib.resources.abc.Traversable, which is an undocumented class (dead link)
now. It should be documented.
The [source] has helpful docstrings, but is missing some details. I'd like to
know
Petr Viktorin added the comment:
So. According to PEP 384 (which added all structs in the stable ABI, except
Py_buffer), some structs are opaque and others have a few members exposed:
https://peps.python.org/pep-0384/#structures
I will split the latter into 1) structs that have a few fields
Change by Petr Viktorin :
--
keywords: +patch
pull_requests: +30272
stage: -> patch review
pull_request: https://github.com/python/cpython/pull/32196
___
Python tracker
<https://bugs.python.org/issu
Change by Petr Viktorin :
--
keywords: +patch
pull_requests: +30274
stage: -> patch review
pull_request: https://github.com/python/cpython/pull/32195
___
Python tracker
<https://bugs.python.org/issu
New submission from Petr Viktorin :
This issue groups together a few changes I'd like to make:
- Rename Doc/data/stable_abi.dat to *.csv, so it gets syntax highlighting. (The
only user of that file should be the Sphinx extension)
- Mark autogenerated files as such in .gitattributes
- Co
New submission from Petr Viktorin :
- PyThread_get_thread_native_id is only available when PY_HAVE_THREAD_NATIVE_ID
is defined. The test currently always expects it to be available.
- PyOS_CheckStack is only available when USE_STACKCHECK is defined (i.e. on
Windows). It should be exported
Petr Viktorin added the comment:
> Ah, sorry, I could have described the issue better. It's not a problem with
> exporting, PyThread_get_thread_native_id() isn't available on Solaris (and
> possibly other platforms) at all.
Jakub, does this mean test_stable_abi_ctypes fa
New submission from Petr Viktorin :
malloc() returns memory that's "suitably aligned for any built-in type".
All of Python's allocation functions should do the same.
In bpo-27987 (PR-12850, PR-13336), the alignment was raised* to 16 bytes and
`long double`. Th
Petr Viktorin added the comment:
New changeset 079143df7e40c4d336cb5c385b166aa91058d050 by Petr Viktorin in
branch 'main':
bpo-47168: Mark files generated by `make regen-limited-abi` as generated
(GH-32195)
https://github.com/python/cpython/commit/079143df7e40c4d336cb5c385b166a
Petr Viktorin added the comment:
- PEP 570 (Python Positional-Only Parameters) is final
- The language spec was updated to mention it
- the FAQ entry was revised to not link to the PEP and be a self-contained
explanation
IMO the only thing left is to make searching for `/` yield the right
Change by Petr Viktorin :
--
pull_requests: +30412
pull_request: https://github.com/python/cpython/pull/32365
___
Python tracker
<https://bugs.python.org/issue47
Petr Viktorin added the comment:
New changeset d79f118d044e9b4244b5dfda35448d39202d7f56 by Petr Viktorin in
branch 'main':
bpo-47115: Document which parts of structs are in limited API/stable ABI
(GH-32196)
https://github.com/python/cpython/commit/d79f118d044e9b4244b5dfda35448d
Change by Petr Viktorin :
--
keywords: +patch
pull_requests: +30414
stage: -> patch review
pull_request: https://github.com/python/cpython/pull/32365
___
Python tracker
<https://bugs.python.org/issu
Petr Viktorin added the comment:
This has been solved by the new theme. The Goolgle report linked above shows
“Passed” and “96” (out of 100).
It does show a few opportunities to improve, but many seem to indirectly
complain that the page is big.
I'm closing the issue.
--
Change by Petr Viktorin :
--
resolution: -> fixed
stage: needs patch -> resolved
status: open -> closed
___
Python tracker
<https://bugs.python.or
Change by Petr Viktorin :
--
pull_requests: +30440
pull_request: https://github.com/python/cpython/pull/32414
___
Python tracker
<https://bugs.python.org/issue47
Change by Petr Viktorin :
--
pull_requests: +30441
pull_request: https://github.com/python/cpython/pull/32415
___
Python tracker
<https://bugs.python.org/issue47
Petr Viktorin added the comment:
Thinking more about Doc/data/stable_abi.dat, I don't think the rename is worth
it. The file is not meant to be used/edited by humans.
If someone needs the data for something other than running the Sphinx
extension, let me know. We should provide a proper
Petr Viktorin added the comment:
I am not convinced.
I'm wary of making error messages depend on the str representation of a
function; that would prevent us from changing it later.
I'm wary of "%S" used in error messages. Those are for the programmer, not the
user, s
Petr Viktorin added the comment:
I don't understand how the function ended up with func_code=NULL. That
shouldn't be a valid function to call, IMO.
Do you have any info on how the function ended up in that state?
--
___
Python track
Petr Viktorin added the comment:
I'm not sure adding a check would solve this. What should be done when a
function with func_code=NULL is called? "Silently do nothing" is not really an
option; raising an exception wouldn't help much in this case.
I wonder if a function&
Petr Viktorin added the comment:
> What downsides do we see raising an exception?
Yeah, on second thought, that would probably be best. We still want PR #15641
as well, so the exception doesn't actually happen in practice.
Note that _PyEval_EvalCodeWithName (the main use of func
Petr Viktorin added the comment:
The traceback does have some ctypedescr_dealloc and cfield_dealloc from _cffi.
Could you check what those are?
--
___
Python tracker
<https://bugs.python.org/issue38
Change by Petr Viktorin :
--
pull_requests: +15429
pull_request: https://github.com/python/cpython/pull/15776
___
Python tracker
<https://bugs.python.org/issue37
Petr Viktorin added the comment:
New changeset ff023ed36ea260ab64be5895f1f1f087c798987a by Petr Viktorin (Eddie
Elizondo) in branch 'master':
bpo-37879: Suppress subtype_dealloc decref when base type is a C heap type
(GH-15323)
https://github.com/python/cpyt
Petr Viktorin added the comment:
New changeset 7a6873cdb1f496447ac5d57ae457eacbb56b7972 by Petr Viktorin (Jeroen
Demeyer) in branch 'master':
bpo-37151: remove special case for PyCFunction from PyObject_Call (GH-14684)
https://github.com/python/cpyt
Petr Viktorin added the comment:
All linked PRs are merged, so I'm closing this. Feel free to reopen if there's
more!
--
resolution: -> fixed
stage: patch review -> resolved
status: open -> closed
___
Python tracker
<h
New submission from Petr Viktorin :
Hello,
In Fedora, we've long used a Bash script to compile modules to bytecode, as the
compileall module was historically unsuitable.
Recently, Lumír re-wrote our enhancements to compileall, and published the
result as compileall2 on PyPI. We'd l
Change by Petr Viktorin :
--
pull_requests: +15599
pull_request: https://github.com/python/cpython/pull/15966
___
Python tracker
<https://bugs.python.org/issue37
Petr Viktorin added the comment:
I like PR 14890 better. I like the separation of representation for error
messages (where it's clearer that this is a callable) and for __str__.
Also, changing the __str__ of functions would need much wider discussion than
on issues/PRs.
I left
Change by Petr Viktorin :
--
pull_requests: +15629
pull_request: https://github.com/python/cpython/pull/16004
___
Python tracker
<https://bugs.python.org/issue37
Petr Viktorin added the comment:
My bad, I didn't publish the comments. They should be there now.
--
___
Python tracker
<https://bugs.python.org/is
Petr Viktorin added the comment:
New changeset 5e9caeec76119a0d61c25f1466c27b7dbd5115bd by Petr Viktorin in
branch 'master':
bpo-37879: Fix warnings in _testcapimodule (GH-16004)
https://github.com/python/cpython/commit/5e9caeec76119a0d61c25f1466c27b
Petr Viktorin added the comment:
New changeset 3562ae25402aad36583bc27d4d82c67554323d5e by Petr Viktorin in
branch '3.8':
[3.8] bpo-37879: Suppress subtype_dealloc decref when base type is a C heap
type (GH-15323, GH-16004) (GH-15966)
https://github.com/python/cpyt
Change by Petr Viktorin :
--
resolution: -> fixed
stage: patch review -> resolved
status: open -> closed
___
Python tracker
<https://bugs.python.or
Petr Viktorin added the comment:
Fixed here and in https://bugs.python.org/issue37499 (
https://github.com/python/cpython/pull/15776 )
--
resolution: -> fixed
stage: patch review -> resolved
status: open -> closed
___
Python tracke
1 - 100 of 907 matches
Mail list logo
