Sorry, something went wrong.
There was a problem hiding this comment.
This is IMO a nice improvement. Thanks!
You might want to add Modules/_testcapi/testdcapimodule_parts.h to MODULE__TESTCAPI_DEPS in Makefile.pre.in.
Sorry, something went wrong.
|
I suggest shorter file names:
Would it be possible to move Modules/_testcapimodule.c to Modules/_testcapi/module.c? |
Sorry, something went wrong.
|
@vstinner, @markshannon: Do you think this is a good idea? Yes, I like the idea of splitting the long _testcapimodule.c into multiple shorter C files. |
Sorry, something went wrong.
|
Would it be possible to move Modules/_testcapimodule.c to Modules/_testcapi/module.c? In a different PR, maybe. There are a few more files to go on that ride, and I'd like to limit this PR to one change. |
Sorry, something went wrong.
There was a problem hiding this comment.
LGTM.
While _testcapimodule.c is built as a single C extension, I'm not convinced by the purpose of _testcapi/parts.h which only declares a single function. But it's more explicit for you, go for it.
Sorry, something went wrong.
|
Thanks! |
Sorry, something went wrong.
|
_testcapi/parts.h should list the init functions from all the parts. There'll be more :) |
Sorry, something went wrong.
|
Status check is done, and it's a success ✅ . |
Sorry, something went wrong.
|
The PR broke WASM build builts: https://buildbot.python.org/all/#/builders/1050/builds/86 error: unable to open output file 'Modules/_testcapi/vectorcall.o': 'No such file or directory'
1 error generated.
emcc: error: 'ccache /opt/emsdk/upstream/bin/clang -target wasm32-unknown-emscripten -D__EMSCRIPTEN_SHARED_MEMORY__=1 -DEMSCRIPTEN -D__EMSCRIPTEN_major__=3 -D__EMSCRIPTEN_minor__=1 -D__EMSCRIPTEN_tiny__=15 -fignore-exceptions -mllvm -combiner-global-alias-analysis=false -mllvm -enable-emscripten-sjlj -mllvm -disable-lsr -Werror=implicit-function-declaration -Xclang -iwithsysroot/include/SDL --sysroot=/opt/buildbot/.emscripten_cache/sysroot -Xclang -iwithsysroot/include/compat -Wsign-compare -Wunreachable-code -DNDEBUG -g3 -fwrapv -O3 -Wall -pthread -std=c11 -Wextra -Wno-unused-parameter -Wno-missing-field-initializers -Werror=implicit-function-declaration -fvisibility=hidden -I../../Include/internal -IObjects -IInclude -IPython -I. -I../../Include -DPy_BUILD_CORE_BUILTIN -c -matomics -mbulk-memory ../../Modules/_testcapi/vectorcall.c -o Modules/_testcapi/vectorcall.o' failed (returned 1)
make: *** [Makefile:2800: Modules/_testcapi/vectorcall.o] Error 1
make: *** Waiting for unfinished jobs....
emmake: error: 'make -j2 all' failed (returned 2)
gh-94695 should fix them |
Sorry, something went wrong.
The _testcapimodule.c file is getting too large to work with effectively.
This PR lays out a general structure of how tests can be split up, with more splitting to come later if the structure is OK.
Vectorcall tests aren't the biggest issue -- it's just an area I want to work on next, so I'm starting here.
An issue specific to vectorcall tests is that it wasn't clear that e.g. MethodDescriptor2 is related to testing vectorcall: the /* Test PEP 590 */ section had an ambiguous end. Separate file should make things like this much clearer.
OTOH, for some pieces it might not be clear where they should be -- I left meth_fastcall with tests of the other calling conventions. IMO, even with the ambiguity it's still worth it to split the huge file up.
I'm not sure about the buildsystem changes, hopefully CI will tell me what's wrong.
@vstinner, @markshannon: Do you think this is a good idea?
Automerge-Triggered-By: GH:encukou