Allow globally forcing C file regeneration by setting the env var CYTHON_FORCE_REGEN=1, e.g. from external build systems.

Always regenerate .c/cpp output files when changing the Cython version. This is probably what users expect since silently keeping outdated files is error prone.

By disabling "CYTHON_FAST_THREAD_STATE" and "CYTHON_USE_EXC_INFO_STACK"

I think this still leaves some breakage in Coroutines.c but it's enough of a fix that Cython succeeds in building itself. Therefore I think it's worth doing now even if it doesn't fix everything.

Related to https://github.com/cython/cython/issues/4500

Add Cython.Utils to the list of compiled modules and include a faster @contextmanager for try-finally cases.

This reverts commit aaba1229.

Revert "Check for "Optional[ctype]" earlier because we need to make sure that "Optional[int]" etc. interprets "int" as (valid) Python int type and not (invalid) C int type."

This reverts commit f73108cc.

This reverts commit cd5fb480.

This reverts commit 7914e3b3.

Check for "Optional[ctype]" earlier because we need to make sure that "Optional[int]" etc. interprets "int" as (valid) Python int type and not (invalid) C int type.

See https://github.com/cython/cython/issues/3883

Improve the error message for "Optional[ctype]". Previously, it failed complaining about "or None", which wasn't the actual syntax construct on user side.

See https://github.com/cython/cython/issues/3883

See https://github.com/cython/cython/issues/3883

New decorator/function "@cython.dataclasses.dataclass" and "cython.dataclasses.field()" to mark dataclasses and their fields.

Tries to match the interface provided by a regular dataclass as much as possible.
This means taking the types from the dataclasses module if available (so they match exactly) or a fallback Python version that just implements the core parts (executed with "PyRun_String()" in the C source).

Use of placeholders in generated "__init__" code means the code in the C file isn't hugely readable. Probably not a huge issue, but don't really see a way round that.

As part of this I've also also implemented a Cython version of "typing.ClassVar". Although really designed for use with dataclasses it behaves sensibly when used in types in a normal cdef class. This is worth documenting more thoroughly.

Closes https://github.com/cython/cython/issues/2903

Fix division-by-zero crash in tests when the C ompiler discards the computational body of a Cython coroutine.

Add a new command line option so that
$ cythonize -M foo.pyx
produces a file `foo.c.dep` with the dependencies of foo.pyx, in addition to `foo.c`.
Try to write relative paths as much as possible.

Backport of https://github.com/cython/cython/pull/4563

Make the test files that depend on "cythonarrayutil.pxi" compilable by themselves, without the search directory setup of the test runner.

Use "Py_ssize_t" instead of "int" for a node counter variable to avoid uselessly limiting its value range.

Closes https://github.com/cython/cython/issues/2636

See https://github.com/cython/cython/issues/4594#issuecomment-1022092245

Disable parallel builds in the 0.29.x branch since they currently misbehave and fail in CI, trying to rebuild everything multiple times in parallel.

Move the option to "exclude" children from the tree traversal from the "VisitorTransform" class up to its parent "TreeVisitor" class.

Use "Py_ssize_t" instead of "int" for a node counter variable to avoid uselessly limiting its value range.

Adapt to "exc_info" changes in CPython 3.11a4, where "->exc_type" and "->exc_traceback" are gone and now get inferred from "->exc_value".

Closes https://github.com/cython/cython/issues/4500