| 9 Jan 2025 |
Tristan Ross | Rip, openexr failed tests due to 1 test timing out | 19:45:13 |
Alex | In reply to @rosscomputerguy:matrix.org Rip, openexr failed tests due to 1 test timing out This unfortunately happens for some packages. Commonly available RISC-V cores can be far slower than the heavily optimised and mature x86 platforms that often set the timings. You could report the issue upstream, but for now disabling checks is an easy solution. It could even save hours of building if you do it pervasively. | 21:20:48 |
Tristan Ross | In reply to @alex:tunstall.xyz
This unfortunately happens for some packages. Commonly available RISC-V cores can be far slower than the heavily optimised and mature x86 platforms that often set the timings. You could report the issue upstream, but for now disabling checks is an easy solution. It could even save hours of building if you do it pervasively. Yeah... I'm kinda aware of that. Hopefully someday we get more performant systems. | 21:22:08 |
Alex | Emulated builds (e.g. via binfmt) also struggle with the problem. Access to clocks and timings interferes with reproducibility, but they are very difficult system calls to take away. | 21:24:54 |
Tristan Ross | Unfortunately, I cannot do emulated builds because those don't work with a page size larger than 4k. | 23:13:05 |
Tristan Ross | I've gotten things like bus errors with just running ls | 23:13:20 |
| 10 Jan 2025 |
| 12b joined the room. | 16:23:35 |
fgaz | In reply to @rosscomputerguy:matrix.org Rip, openexr failed tests due to 1 test timing out I think I increased the timeout of that package already | 22:38:49 |
fgaz | like months ago | 22:38:52 |
fgaz | try building with --max-jobs 1 --cores 1 | 22:39:10 |
Tristan Ross | Yeah lol, it's timing out by like half a second | 22:39:32 |
Tristan Ross | But I'm past that now | 22:39:46 |
Tristan Ross | Python stuff is building | 22:39:55 |
Tristan Ross | 459 derivations left | 22:40:11 |
fgaz | https://github.com/NixOS/nixpkgs/blob/61b77b92ab78c6c4eed66823ef619a1cbe0337c4/pkgs/development/libraries/openexr/default.nix#L54-L55 | 22:40:13 |
fgaz | In reply to @rosscomputerguy:matrix.org But I'm past that now Nice! | 22:40:25 |
| 11 Jan 2025 |
Tristan Ross | error: builder for '/nix/store/sq3vjc0dn0rsfhjvdgphgbyb2yipc77h-python3.12-hypothesis-6.112.2.drv' failed with exit code 1;
last 25 log lines:
> You can add @seed(166370714267267105264357939584607094565) to this test or run pytest with --hypothesis-seed=166370714267267105264357939584607094565 to reproduce thi.
> ============================= slowest 20 durations =============================
> 300.00s call hypothesis-python/tests/cover/test_slices.py::test_start_will_equal_0
> 215.85s call hypothesis-python/tests/cover/test_stateful.py::TestLotsOfEntropyPerStepMachine::runTest
> 138.40s call hypothesis-python/tests/cover/test_stateful.py::test_removes_needless_steps
> 125.83s call hypothesis-python/tests/cover/test_stateful.py::test_saves_failing_example_in_database
> 121.27s call hypothesis-python/tests/cover/test_flakiness.py::test_failure_sequence_inducing
> 110.93s call hypothesis-python/tests/cover/test_database_backend.py::TestGADReads::runTest
> 105.06s call hypothesis-python/tests/cover/test_slices.py::test_step_will_be_negative
> 91.98s call hypothesis-python/tests/cover/test_statistical_events.py::test_stateful_with_one_of_bundles_states_are_deduped
> 88.29s call hypothesis-python/tests/cover/test_observability.py::test_observability
> 81.35s call hypothesis-python/tests/cover/test_stateful.py::TestMachineUsingMultiple::runTest
> 67.26s call hypothesis-python/tests/cover/test_statistical_events.py::test_stateful_states_are_deduped
> 66.79s call hypothesis-python/tests/cover/test_sampled_from.py::test_efficient_lists_of_tuples_first_element_sampled_from
> 65.39s call hypothesis-python/tests/cover/test_regex.py::test_fullmatch_generates_example[[ab]*-aBb]
> 55.93s call hypothesis-python/tests/cover/test_testdecorators.py::test_float_addition_is_associative
> 54.70s call hypothesis-python/tests/cover/test_type_lookup.py::test_can_generate_from_all_registered_types[BaseExceptionGroup]
> 53.84s call hypothesis-python/tests/cover/test_composite.py::test_can_shrink_matrices_with_length_param
> 51.04s call hypothesis-python/tests/cover/test_regex.py::test_matching[True-True-\\s-None]
> 48.16s call hypothesis-python/tests/cover/test_lookup.py::test_specialised_mapping_types[typing.OrderedDict[tests.cover.test_lookup.Elem, tests.cover.test_lookup]
> 47.87s call hypothesis-python/tests/cover/test_regex.py::test_matching[True-True-\\w-is_word]
> 47.57s call hypothesis-python/tests/cover/test_lookup.py::test_resolves_forwardrefs_to_builtin_types[object]
> =========================== short test summary info ============================
> FAILED tests/cover/test_stateful.py::test_can_run_with_no_db - hypothesis.errors.FailedHealthCheck: Data generation is extremely slow: Onl...
> ====== 1 failed, 3211 passed, 5 skipped, 1 xfailed in 2774.21s (0:46:14) =======
For full logs, run 'nix log /nix/store/sq3vjc0dn0rsfhjvdgphgbyb2yipc77h-python3.12-hypothesis-6.112.2.drv'.
Fun
| 00:30:20 |
Tristan Ross | I've enabled --keep-going so these one off errors don't kill when I still have like 440 | 00:35:45 |
hexa | run on a potato | 01:51:44 |
hexa | * apparently run on a potato | 01:51:49 |
Tristan Ross | It's a VisionFive 2 lol | 01:54:27 |
Tristan Ross | And the remote builder is a HiFive Premier P550 that Matthew Croughan is running | 01:54:51 |
hexa | ouch. | 01:55:06 |
hexa | I still remember the HiFive Unmatched | 01:55:19 |
hexa | and it was indeed unmatched 😕 | 01:55:23 |
Tristan Ross | lol | 01:55:40 |
Tristan Ross | All the RISC-V hardware with great performance is probably under an NDA and behind a ridiculous amount of money. | 01:56:08 |
hexa | 🤷♂️ | 01:56:55 |
| @orzklv:matrix.org left the room. | 04:07:38 |
Tristan Ross | 368 derivations to go | 06:52:01 |