On Mon, Jan 30 2023, Stuart Henderson <stu@spacehopper.org> wrote:
> ok to import these which were split off from py-aiohttp?
> The descriptions aren't brilliant but I didn't find anything
> better to crib from upstream.
Good enough IMO, ok jca@ to import
> Update diff for py-aiohttp below if anyone would like to give that
> a spin too. (Not tested with LLVM 15 but I doubt it will hurt).
It builds and packages using clang 15. There's one problem...
[...]
> Index: Makefile
> ===================================================================
> RCS file: /cvs/ports/www/py-aiohttp/Makefile,v
> retrieving revision 1.21
> diff -u -p -r1.21 Makefile
> --- Makefile 27 Nov 2022 11:54:23 -0000 1.21
> +++ Makefile 30 Jan 2023 16:28:31 -0000
> @@ -1,10 +1,8 @@
> COMMENT = http client/server for asyncio
>
> -# 3.8.0 adds new deps on frozenlist, aiosignal
> -MODPY_EGG_VERSION = 3.7.4
> +MODPY_EGG_VERSION = 3.8.3
> DISTNAME = aiohttp-${MODPY_EGG_VERSION}
> PKGNAME = py-${DISTNAME}
> -REVISION = 3
>
> CATEGORIES = www
>
> @@ -19,23 +17,36 @@ FLAVORS = python3
>
> MODPY_PI = Yes
> MODPY_PYBUILD = setuptools
> -# XXX some tests fail as we don't have ports for re_assert and async_generator
> -
> -RUN_DEPENDS = devel/py-async-timeout${MODPY_FLAVOR} \
> - devel/py-attrs${MODPY_FLAVOR}>=17.3.0 \
> - devel/py-typing-extensions${MODPY_FLAVOR}>=3.6.5 \
> - textproc/py-chardet${MODPY_FLAVOR}>=2.0 \
> - www/py-multidict${MODPY_FLAVOR}>=4.5 \
> - www/py-yarl${MODPY_FLAVOR}
> +# disable tests using "proxy" and "re_assert" modules which aren't in ports
> +MODPY_PYTEST_ARGS = --ignore tests/test_proxy_functional.py \
> + --ignore tests/test_client_session.py \
> + --ignore tests/test_streams.py \
> + --ignore tests/test_urldispatch.py \
> + --ignore tests/test_web_response.py
> +
> +RUN_DEPENDS = devel/py-attrs${MODPY_FLAVOR}>=17.3.0 \
> + textproc/py-charset-normalizer${MODPY_FLAVOR} \
> + www/py-multidict${MODPY_FLAVOR}<7 \
> + devel/py-async-timeout${MODPY_FLAVOR}<5 \
> + www/py-yarl${MODPY_FLAVOR}<2 \
> + devel/py-frozenlist${MODPY_FLAVOR} \
> + devel/py-aiosignal${MODPY_FLAVOR}
> +# optional, but let's avoid different behaviour depending on whether
> +# they're installed
here you're resetting RUN_DEPENDS.
I got very confused when debugging this because the missing deps seen at
the start of ''make test'' *should* have come from the TEST_DEPENDS on
www/py-gunicorn, which itself has a RUN_DEPENDS on www/py-aiohttp.
But I guess that works...
After updating devel/py-async-timeout to 4.0.2 I get tests that perform
better than in -current (10 failures):
[...]
FAILED tests/test_client_request.py::test_default_loop[pyloop]
FAILED tests/test_http_parser.py::test_c_parser_loaded - AssertionError: asse...
FAILED tests/test_http_parser.py::test_url_connect[py-parser-pyloop] - TypeEr...
FAILED tests/test_http_parser.py::test_headers_connect[py-parser-pyloop] - Ty...
FAILED tests/test_web_app.py::test_set_loop_default_loop
FAILED tests/test_web_server.py::test_unsupported_upgrade[pyloop]
====== 6 failed, 2036 passed, 20 skipped, 6 xfailed in 243.53s (0:04:03) =======
ok jca@, nitpick below
> +RUN_DEPENDS = archivers/py-brotli${MODPY_FLAVOR} \
> + net/py-aiodns${MODPY_FLAVOR}
>
> TEST_DEPENDS = devel/py-freezegun${MODPY_FLAVOR} \
> devel/py-test${MODPY_FLAVOR} \
> devel/py-test-cov${MODPY_FLAVOR} \
> + devel/py-test-mock${MODPY_FLAVOR} \
> devel/py-test-timeout${MODPY_FLAVOR} \
> devel/py-test-xdist${MODPY_FLAVOR} \
> - www/py-gunicorn${MODPY_FLAVOR}
> + www/py-gunicorn${MODPY_FLAVOR} \
> + security/py-trustme${MODPY_FLAVOR}
Could you please keep the TEST_DEPENDS list sorted?
--
jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF DDCC 0DFA 74AE 1524 E7EE
No comments:
Post a Comment