-
-
Notifications
You must be signed in to change notification settings - Fork 654
Open
Description
Currently there are 46 instances of them in every pull request. It makes the GitHub annotations highlight annoying to look at.
One example test-long
test run
2025-02-22T13:29:16.0611644Z Test ran for 30.22s cpu, 30.58s wall
2025-02-22T13:56:25.4496259Z Test ran for 31.11s cpu, 31.48s wall
2025-02-22T13:28:23.1625804Z Test ran for 31.73s cpu, 32.16s wall
2025-02-22T13:57:32.9606147Z Test ran for 32.38s cpu, 32.99s wall
2025-02-22T13:53:15.8347722Z Test ran for 32.69s cpu, 32.72s wall
2025-02-22T13:29:36.9565765Z Test ran for 33.03s cpu, 33.57s wall
2025-02-22T13:35:25.1981149Z Test ran for 33.78s cpu, 33.79s wall
2025-02-22T13:27:25.8378732Z Test ran for 34.25s cpu, 36.98s wall
2025-02-22T13:57:58.6533147Z Test ran for 34.36s cpu, 34.74s wall
2025-02-22T13:54:03.6386977Z Test ran for 34.72s cpu, 35.63s wall
2025-02-22T13:41:49.1500562Z Test ran for 34.78s cpu, 33.49s wall
2025-02-22T13:47:03.5916907Z Test ran for 34.78s cpu, 34.79s wall
2025-02-22T13:57:11.1852268Z Test ran for 34.92s cpu, 34.93s wall
2025-02-22T13:52:42.6317100Z Test ran for 35.26s cpu, 35.28s wall
2025-02-22T13:57:07.2761191Z Test ran for 35.43s cpu, 35.48s wall
2025-02-22T13:47:03.6276858Z Test ran for 35.97s cpu, 35.99s wall
2025-02-22T13:33:59.8155323Z Test ran for 36.08s cpu, 36.53s wall
2025-02-22T13:33:59.8160794Z Test ran for 36.36s cpu, 36.80s wall
2025-02-22T13:58:24.2374105Z Test ran for 37.01s cpu, 37.01s wall
2025-02-22T13:38:20.1084890Z Test ran for 39.34s cpu, 32.77s wall
2025-02-22T13:32:30.2429610Z Test ran for 40.09s cpu, 40.54s wall
2025-02-22T13:30:13.0849221Z Test ran for 41.46s cpu, 41.95s wall
2025-02-22T13:56:24.1576362Z Test ran for 42.12s cpu, 36.86s wall
2025-02-22T13:33:00.8968907Z Test ran for 45.19s cpu, 45.95s wall
2025-02-22T13:47:03.4548029Z Test ran for 47.25s cpu, 47.26s wall
2025-02-22T13:37:55.8668785Z Test ran for 48.04s cpu, 48.94s wall
2025-02-22T13:30:12.6184455Z Test ran for 50.20s cpu, 49.13s wall
2025-02-22T13:34:54.3308336Z Test ran for 52.42s cpu, 53.20s wall
2025-02-22T14:01:16.4019104Z Test ran for 58.09s cpu, 58.15s wall
2025-02-22T13:32:14.9510569Z Test ran for 59.29s cpu, 59.71s wall
2025-02-22T13:33:40.1948935Z Test ran for 59.87s cpu, 60.52s wall
2025-02-22T13:31:46.0065634Z Test ran for 60.58s cpu, 61.24s wall
2025-02-22T13:44:58.3631500Z Test ran for 66.36s cpu, 66.39s wall
2025-02-22T13:44:06.1227057Z Test ran for 66.68s cpu, 67.55s wall
2025-02-22T13:37:06.6633748Z Test ran for 67.79s cpu, 67.80s wall
2025-02-22T13:34:25.2372204Z Test ran for 69.46s cpu, 70.60s wall
2025-02-22T13:36:15.2720051Z Test ran for 81.34s cpu, 82.21s wall
2025-02-22T13:34:50.9208405Z Test ran for 89.42s cpu, 90.39s wall
2025-02-22T13:28:56.5552520Z Test ran for 98.55s cpu, 100.16s wall
2025-02-22T13:37:05.8736350Z Test ran for 100.18s cpu, 100.22s wall
2025-02-22T13:45:23.2359017Z Test ran for 135.93s cpu, 136.09s wall
2025-02-22T13:39:47.2658831Z Test ran for 151.55s cpu, 151.65s wall
2025-02-22T13:39:09.2373108Z Test ran for 170.78s cpu, 170.91s wall
2025-02-22T13:36:21.6029463Z Test ran for 258.19s cpu, 260.96s wall
2025-02-22T13:54:24.6318133Z Test ran for 341.58s cpu, 341.77s wall
2025-02-22T13:46:55.1348524Z Test ran for 427.29s cpu, 427.51s wall
Some of them might be performance regression (note that https://doc.sagemath.org/html/en/developer/coding_basics.html#special-markup-to-influence-doctests requires any doctest to take less than 30 seconds), some are intentional like
sage/groups/perm_gps/permgroup_named.py
3517: ....: for n in [1..64] for k in [1..numgps(n)]) # long time (180s)
I can see a few ways around it.
- just bump the current limit from 30s to 500s. (easiest to implement, if nobody complains I'd pull request with this) (disadvantage: performance regression got swept under the rug)
- implement a feature where a test that is
# long time (
$x$ s)
has limit$x$ seconds instead of$30$ (disadvantage: as above, sort of) - temporarily replace all tests that take too long with
# known bug (see :issue:`39569`)
(disadvantage: their correctness is then no longer tested)
Metadata
Metadata
Assignees
Labels
No labels