conda-forge / osqp-feedstock

A conda-smithy repository for osqp.
BSD 3-Clause "New" or "Revised" License
0 stars 13 forks source link

bump to 0.6.2 #28

Closed h-vetinari closed 3 years ago

h-vetinari commented 3 years ago

Checklist

Since the migrator hasn't come past yet, do it manually

conda-forge-linter commented 3 years ago

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

h-vetinari commented 3 years ago

Aside from the missing aarch-builds for qdldl-python (see https://github.com/conda-forge/qdldl-python-feedstock/pull/7), this should be ready. Can you please have a look @bstellato @gbanjac ?

bstellato commented 3 years ago

It looks great to me! Thanks a lot @h-vetinari for the help with the new release. Just a couple of questions:

h-vetinari commented 3 years ago

Hey @bstellato, thanks for the feedback!

* is Python 3.9 supported? It seems to me the azure builds are up to Python 3.8

24 had failing tests (those happened across all current PRs on 0.6.1 for reasons I could not determine); as soon as this PR is through, I'll rebase #24 and that shouldn't be an issue.

* I am not sure why some of the tests (in particular drone CI) are failing. I am not familiar with it though.

These aren't failing tests, just missing builds because the newly-added dependency qdldl-python hadn't been built for aarch/ppc64le yet (but solved by https://github.com/conda-forge/qdldl-python-feedstock/pull/7, for which master is currently building)

* Regarding Travis CI, is it still used in conda-forge? I have seen many projects are transitioning to github actions.

It's still in use for ppc64le builds. Not sure what agreement was reached between conda(-forge) and travis, but all transitioning projects (that I know of) only do so because the free tier has way too little CI capacity to support development.

h-vetinari commented 3 years ago

If you restart this PR in about an hour (once the builds on for qdldl are through and the builds have reached the server cache), the CI should go green, and the PR can be merged - just in my timezone it's a bit late already to wait for that. ;-)

h-vetinari commented 3 years ago

Unfortunately, we have what looks like real test errors on ppc64le (aarch still didn't pick up the dependency):

=========================== short test summary info ============================
FAILED tests/basic_test.py::basic_tests::test_basic_QP - AssertionError: 
FAILED tests/basic_test.py::basic_tests::test_update_bounds - AssertionError: 
FAILED tests/basic_test.py::basic_tests::test_update_l - AssertionError:
FAILED tests/basic_test.py::basic_tests::test_update_q - AssertionError: 
FAILED tests/basic_test.py::basic_tests::test_update_u - AssertionError: 
FAILED tests/feasibility_test.py::feasibility_tests::test_feasibility_problem
FAILED tests/polishing_test.py::polish_tests::test_polish_random - AssertionE...
FAILED tests/polishing_test.py::polish_tests::test_polish_simple - AssertionE...
FAILED tests/polishing_test.py::polish_tests::test_polish_unconstrained - Typ...
FAILED tests/unconstrained_test.py::unconstrained_tests::test_unconstrained_problem
FAILED tests/update_matrices_test.py::update_matrices_tests::test_solve - Ass...
FAILED tests/update_matrices_test.py::update_matrices_tests::test_update_A - ...
FAILED tests/update_matrices_test.py::update_matrices_tests::test_update_A_allind
FAILED tests/update_matrices_test.py::update_matrices_tests::test_update_P - ...
FAILED tests/update_matrices_test.py::update_matrices_tests::test_update_P_A_allind
FAILED tests/update_matrices_test.py::update_matrices_tests::test_update_P_A_indA
FAILED tests/update_matrices_test.py::update_matrices_tests::test_update_P_A_indP
FAILED tests/update_matrices_test.py::update_matrices_tests::test_update_P_A_indP_indA
FAILED tests/update_matrices_test.py::update_matrices_tests::test_update_P_allind
================== 19 failed, 33 passed, 1 skipped in 25.76s ===================

I'm still going to merge this, and we can figure out aarch support independently of #24 an #26.