Hood Chatham added the comment:
Okay I got it working. I was making several mistakes that each caused the same
symptoms so it was hard to track them all down.
Thanks again for your help!
--
status: open -> closed
___
Python tracker
<
Hood Chatham added the comment:
There is still a pretty good change I am making some dumb mistake.
--
___
Python tracker
<https://bugs.python.org/issue47
Hood Chatham added the comment:
config.log is apparently 1.7 MB and when I try to upload I get "Error 413:
Entity Too Large". I've attached the Makefile.
--
Added file: https://bugs.python.org/file50723/Makefile
___
Python
Hood Chatham added the comment:
I take it back, I am still having this problem.
--
status: closed -> open
___
Python tracker
<https://bugs.python.org/issu
Hood Chatham added the comment:
Okay I found the problem. Elsewhere in my Makefiles `PYTHON_FOR_BUILD` was
getting overwritten.
--
stage: -> resolved
status: open -> closed
___
Python tracker
<https://bugs.python.org/i
New submission from Hood Chatham :
I am trying to build tot Python for Emscripten to test recent changes and
having trouble. The problem is caused by recent changes since 3.11.0a6 to
`sre`. I build a build Python 3.11 from tot, and provided it to `./configure
--with-build-python
Hood Chatham added the comment:
Ideally the library should be able to define a different macro, e.g.,
CTYPES_CUSTOM_MAX_ARGCOUNT. This way the libffi port can define
CTYPES_CUSTOM_MAX_ARGCOUNT and it will be compatible with older Python versions
without causing Warning: "CTYPES_MAX_ARG
Change by Hood Chatham :
--
nosy: +amaury.forgeotdarc, belopolsky, christian.heimes, meador.inge
___
Python tracker
<https://bugs.python.org/issue47208>
___
___
New submission from Hood Chatham :
ctypes defines `CTYPES_MAX_ARGCOUNT` to be 1024:
https://github.com/python/cpython/blob/6db2db91b96aaa1270c200ec931a2250fe2799c7/Modules/_ctypes/ctypes.h#L21
If a function call is attempted with more than 1024 arguments, it will fail.
The libffi emscripten
Change by Hood Chatham :
--
nosy: +amaury.forgeotdarc, belopolsky, meador.inge
___
Python tracker
<https://bugs.python.org/issue47197>
___
___
Python-bugs-list m
Change by Hood Chatham :
--
pull_requests: +30341
pull_request: https://github.com/python/cpython/pull/32280
___
Python tracker
<https://bugs.python.org/issue47
Hood Chatham added the comment:
Actually, I think the _PyImport_InitFunc trampoline call is only there as a
work around for the problematic `test_imp` and `test_import` Python tests. I
don't know of any third party code with a problem
Change by Hood Chatham :
--
keywords: +patch
nosy: +hoodmane
nosy_count: 2.0 -> 3.0
pull_requests: +30317
stage: -> patch review
pull_request: https://github.com/python/cpython/pull/32246
___
Python tracker
<https://bugs.python.org/i
New submission from Hood Chatham :
On wasm targets, `test_code` fails. It's because of a bad function pointer
cast. The problematic code is as follows:
```py
import ctypes
py = ctypes.pythonapi
freefunc = ctypes.CFUNCTYPE(None, ctypes.c_voidp)
RequestCodeExtraIndex
Hood Chatham added the comment:
Every other `PyInit` function in that module also needs the spec argument to be
removed in order for test_importlib to pass.
--
___
Python tracker
<https://bugs.python.org/issue47
Change by Hood Chatham :
--
keywords: +patch
nosy: +hoodmane
nosy_count: 2.0 -> 3.0
pull_requests: +30315
stage: -> patch review
pull_request: https://github.com/python/cpython/pull/32244
___
Python tracker
<https://bugs.python.org/i
New submission from Hood Chatham :
The function PyInit_imp_dummy is declared as void f(PyObject* spec) and ignores
the argument. In the test, it is called via imp.load_dynamic as void f(void).
On wasm targets without the call trampolines added in
https://bugs.python.org/issue47162
this
Hood Chatham added the comment:
As an update, building the chrome devtools with the wasm limit set to 12Mb
following that guide worked fantastic. Highly recommend it. (Though there are a
few paths that are out of date, I had to consult google's devtools docs.)
The problem is
Hood Chatham added the comment:
I'm having trouble pinpointing the bad function pointer because chrome cuts off
the end of the wasm file with:
```
;; text is truncated due to size
```
Maybe I should follow the directions here, since this is a consistent nuisance.
https://www.diver
Hood Chatham added the comment:
Note that there are still Python test suite failures in emscripten without
these trampolines enabled, for instance test_imp fails. The only trampoline
needed to pass the core test suite is _PyImport_InitFunc_TrampolineCall.
--
nosy: +hoodchatham
Change by Hood Chatham :
--
versions: +Python 3.11
___
Python tracker
<https://bugs.python.org/issue47176>
___
___
Python-bugs-list mailing list
Unsubscribe:
Change by Hood Chatham :
--
versions: -Python 3.10
___
Python tracker
<https://bugs.python.org/issue47176>
___
___
Python-bugs-list mailing list
Unsubscribe:
Change by Hood Chatham :
--
keywords: +patch
pull_requests: +30284
stage: -> patch review
pull_request: https://github.com/python/cpython/pull/32209
___
Python tracker
<https://bugs.python.org/issu
New submission from Hood Chatham :
In Pyodide, we need to patch the interpreter to allow keyboard interrupts. We
build Python without pthreads support because Emscripten doesn't currently
support using pthreads and dynamic linking in the same build. It is still
possible to handle UI a
24 matches
Mail list logo
