| Commit message (Collapse) | Author | Age | Files | Lines | |
|---|---|---|---|---|---|
| * | Fix setup.py encoding; bump github codeql from v2 to v2 | Jason Madden | 2023-01-26 | 1 | -1/+2 |
| | | |||||
| * | Adjusting compiler argument settings and detection. | Jason Madden | 2023-01-26 | 1 | -9/+32 |
| | | | | | Fixes #300 (thanks to https://github.com/archaon616) | ||||
| * | Stop trying to use faulthandler because all releases have been yanked from PyPI. | Jason Madden | 2023-01-26 | 1 | -1/+2 |
| | | |||||
| * | Python 3.11: Fix a memory leak switching to greenlets. | Jason Madden | 2022-11-07 | 1 | -1/+2 |
| | | | | | | | This leak was present in the original implementation of Python 3.11 support (#306) Fixes #328 and fixes https://github.com/gevent/gevent/issues/1924 ; I have run gevent's test suite locally on 3.11 with this fix without seeing any regressions. | ||||
| * | setup.py: Add Python 3.11 to the PyPI classifier metadata. | Jason Madden | 2022-11-03 | 1 | -1/+2 |
| | | |||||
| * | Win32: Using /EHr, annotate inner_bootstrap again as noexcept to try to ↵ | Jason Madden | 2022-10-30 | 1 | -1/+1 |
| | | | | | force the termination that we otheriwse can't seem to get. | ||||
| * | win32 compiler: don't assume C functions cannot throw. | Jason Madden | 2022-10-30 | 1 | -3/+6 |
| | | |||||
| * | Why didn't the WIN32 path get taken? | Jason Madden | 2022-10-29 | 1 | -4/+11 |
| | | |||||
| * | setup: Add MSVC options only when python is compiled with MSVC | Biswapriyo Nath | 2022-02-18 | 1 | -1/+1 |
| | | | | | | | | This prevents MSVC options with mingw toolchain and fixes the build error in mingw toolchain, like these: gcc.exe: error: /EHsc: linker input file not found: No such file or directory gcc.exe: error: /GT: linker input file not found: No such file or directory | ||||
| * | Start work on making PyGreenlet opaque. First part of the conversion: the ↵ | Jason Madden | 2021-11-03 | 1 | -1/+3 |
| | | | | | | | exception state. Also fix PyGreenlet_New to use green_new and green_init; before, it was bypassing a lot of initialization code, and in this case that made things crash. | ||||
| * | Merge branch 'object_ownership' | Jason Madden | 2021-11-02 | 1 | -16/+51 |
| |\ | |||||
| | * | Always test C++. | Jason Madden | 2021-11-02 | 1 | -17/+16 |
| | | | | | | | | | we require a c++ compiler anyway. | ||||
| | * | Neither disabling faulthandler nor enabling optimizations changed the SEH depth. | Jason Madden | 2021-10-29 | 1 | -2/+1 |
| | | | | | | | | | Try running the full suite again. | ||||
| | * | Nope, both are needed. So what about if faulthandler is off and we optimize ↵ | Jason Madden | 2021-10-29 | 1 | -4/+6 |
| | | | | | | | | | (inline things)? | ||||
| | * | Can't use /MDd because we would need the Python debug libs. | Jason Madden | 2021-10-28 | 1 | -5/+6 |
| | | | |||||
| | * | Enable some debug options on MSVC. | Jason Madden | 2021-10-28 | 1 | -1/+16 |
| | | | |||||
| | * | One more try with /EHa. | Jason Madden | 2021-10-28 | 1 | -1/+1 |
| | | | |||||
| | * | Sprinkle in some more noexcept. | Jason Madden | 2021-10-28 | 1 | -1/+1 |
| | | | |||||
| | * | Take the c off the /EHsc. | Jason Madden | 2021-10-28 | 1 | -1/+1 |
| | | | | | | | | | We have several extern C functions that call back into C++ and they could theoretically throw exceptions. | ||||
| | * | The SAFESEH stuff looks like a red herring. | Jason Madden | 2021-10-28 | 1 | -11/+0 |
| | | | |||||
| | * | What if we require SAFESEH? | Jason Madden | 2021-10-28 | 1 | -2/+4 |
| | | | |||||
| | * | /EHa did not fix the 32-bit Windows problem. Maybe /SAFESEH:NO? | Jason Madden | 2021-10-28 | 1 | -3/+7 |
| | | | |||||
| | * | The 32-bit windows error seems to have something to do with Microsoft's SEH. | Jason Madden | 2021-10-28 | 1 | -4/+14 |
| | | | |||||
| | * | Windows: 32 bit builds are crashing somewhere, unclear why. Try to see any ↵ | Jason Madden | 2021-10-28 | 1 | -2/+11 |
| | | | | | | | | | compiler warnings. | ||||
| | * | Use faulthandler on Windows as well. | Jason Madden | 2021-10-28 | 1 | -0/+1 |
| | | | |||||
| | * | Add version pin for docutils on Python 2.7. | Jason Madden | 2021-10-28 | 1 | -0/+3 |
| | | | | | | | | | docutils 0.18 breaks old sphinx with 'TypeError: argument to reversed() must be a sequence' because node.traverse returns a generator. | ||||
| | * | Add leakchecks to the actual test cases themselves. | Jason Madden | 2021-10-19 | 1 | -0/+1 |
| | | | | | | | | | | | | | Instead of relying on zope.testrunner, which is meant to only work in a Py_DEBUG build. This denotes several test cases that are actually leaking by these measures. | ||||
| * | | add support for win/arm64 platform | Niyas Sait | 2021-10-21 | 1 | -2/+3 |
| |/ | |||||
| * | Enable exception handling for old MSVC | Jason Madden | 2021-10-12 | 1 | -0/+5 |
| | | |||||
| * | Implement cleanup using the thread dictionary on platforms that require it. | Jason Madden | 2021-10-12 | 1 | -3/+6 |
| | | |||||
| * | Queue pending cleanups to reduce the number of Py_AddPendingCall slots we take. | Jason Madden | 2021-10-12 | 1 | -1/+2 |
| | | | | | Also add a testing API to be sure cleanups have finished. | ||||
| * | Try declspec(thread) for old MSC. | Jason Madden | 2021-10-12 | 1 | -10/+22 |
| | | |||||
| * | Drop Python 2.7 on Windows; the compiler is far too old. | Jason Madden | 2021-10-12 | 1 | -0/+3 |
| | | |||||
| * | See if all compilers can work in C++ mode. | Jason Madden | 2021-10-12 | 1 | -1/+2 |
| | | | | | | | | | Thread-local C++ objects should be a portable way to get thread destructors. The slp_* functions need to be compiled as C because we call them from assembly. forward declarations should also be c linkage, otherwise MSVC complains. | ||||
| * | Add extra metadata to setup.py | Jason Madden | 2021-09-29 | 1 | -0/+6 |
| | | |||||
| * | Add support for Python 3.10 | Jason Madden | 2021-05-05 | 1 | -0/+1 |
| | | | | | | | Fixs #236 and fixes #237 Locally, the python.org installer for macOS 3.10.0b1 is crashing on the test suite, but versions I've built myself are fine. I'm still trying to figure out what's going on. | ||||
| * | fix: typo in source code url in setup file | Federico Caselli | 2021-01-18 | 1 | -1/+1 |
| | | |||||
| * | Move definition of __version__ into __init__.py. | Jason Madden | 2020-11-16 | 1 | -1/+12 |
| | | | | | | | | | Instead of getting it from C in the GREENLET_VERSION macro. Also read this value in setup.py. This gets us DRY (don't repeat yourself) and compatibility with tools like zest.releaser. | ||||
| * | Build and test in the manylinux environment for amd64 and aarch64, and macos. | Jason Madden | 2020-11-16 | 1 | -0/+1 |
| | | | | | | | | | All of these environments will upload built wheels on release. AppVeyor now also uploads built wheels on releases. Fixes #175 and fixes #191. | ||||
| * | Minor cleanup after merge of #208 | Jason Madden | 2020-11-12 | 1 | -4/+7 |
| | | |||||
| * | Move to the src/ layout, and make greenlet a package.issue184 | Jason Madden | 2020-11-12 | 1 | -16/+32 |
| | | | | | | | Make 'tests' become 'greenlet.tests', and be compatible with 'python -m unittest discover'. Fixes #184. Fixes #184. Fixes #189. | ||||
| * | Remove custom my_build_ext and stop trying to build extensions at test time. | Jason Madden | 2020-11-11 | 1 | -11/+28 |
| | | | | | | | | | | | | | | | | | | | | Building the extensions and getting them in the right place is the job of tox. This includes a temporary minor regression in that the test extensions are now also packaged in the binary distributions. This will be resolved with #189 and #184. Removes some (maybe all, didnt check yet) uses of distutils so partly addresses #185. Fixes #187 Specify an image for appveyor that works for Python 2.7 (the right visual studio is missing on other images) Add installation of missing python versions. Use PYTHON, not PYTHON_ROOT on appveyor. The former is more common in my experience. Use a newer image for 3.9 on appveyor | ||||
| * | Remove redundant code for Python 2.4-2.6 and 3.0-3.3 | Hugo van Kemenade | 2020-10-02 | 1 | -3/+0 |
| | | |||||
| * | Officially drop support for 2.4/5/6 and 3.0/1/2/3/4.issue186 | Jason Madden | 2020-10-01 | 1 | -6/+2 |
| | | | | | | | | | | | Change the classifiers, update the version number, remove those versions from CI. This does not yet make any code changes. Using 3.4 as the cutoff point because that's where the biggest drop in usage comes; it's also no longer supported by gevent or eventlet. Fixes #186 | ||||
| * | Drop 3.0, 3.1 and 3.2; they're no longer supported by setuptools.publish-changes | Jason Madden | 2020-09-30 | 1 | -5/+4 |
| | | | | | And restore the test_suite argument as that's currently required on Appveyor, even though its deprecated. | ||||
| * | Publish the change log on RTD. | Jason Madden | 2020-09-30 | 1 | -24/+31 |
| | | | | | | | | | | | Fixes #188 Some other minor modernizations as well: - use a RTD config file - Rename NEWS to CHANGES.rst as seems to be th current standard - Likewise, rename doc to docs and *.txt to *.rst (hooray syntax highlighting) - Partially require setuptools; we effectively did anyway. More work required here. | ||||
| * | Prepare to release greenlet 0.4.17release-0.4.17 | Alexey Borzenkov | 2020-09-22 | 1 | -3/+1 |
| | | |||||
| * | Prepare to release greenlet 0.4.16release-0.4.16 | Alexey Borzenkov | 2020-06-05 | 1 | -1/+2 |
| | | |||||
| * | Start NEWS for 0.4.16 | Alexey Borzenkov | 2020-04-22 | 1 | -1/+1 |
| | | |||||
| * | Update setup.py | Bengt Lüers | 2020-02-04 | 1 | -0/+1 |
| | | | | Document Python 3.8 compatibility achieved with #152 to be enable #151. | ||||
 |
index : delta/python-packages/greenlet.git | |
| github.com: python-greenlet/greenlet.git |
| summaryrefslogtreecommitdiff |