Fixing yet another flaky build failure 
Build is green
Patch application report for D8196 (id=29574)
Could not rebase; Attempt merge onto 4825f40a4f...
Updating 4825f40a..00215c0a Fast-forward swh/storage/cassandra/storage.py | 2 +- swh/storage/tests/storage_tests.py | 24 +++++++++++++++++++++--- 2 files changed, 22 insertions(+), 4 deletions(-)
Changes applied before test
commit 00215c0a7b9882364179e1dce0da0e649ad1ecd6 Author: Valentin Lorentz <firstname.lastname@example.org> Date: Fri Aug 5 11:27:08 2022 +0200 Fix crash of test_*_arbitrary when given objects with the same id commit 9b9eb282e83eb7ef8c11b7e6ca8d0164bf0dd141 Author: Valentin Lorentz <email@example.com> Date: Fri Aug 5 11:20:33 2022 +0200 cassandra: Make origin_visit_status_get_random's interval consistent with postgresql The postgresql implementation uses '3 months', which is closer to 13 weeks than to 12 weeks. commit 56f69e56b81c5abc28cb6e217e298ab95fb501cb Author: Valentin Lorentz <firstname.lastname@example.org> Date: Fri Aug 5 11:18:46 2022 +0200 Fix flakiness of test_origin_visit_status_get_random_nothing_found start is increased from 13 to 14, because 13 weeks is 91 days, ie. 30+31+30; so it is sometimes smaller than 3 months. This was only hit rarely because the number of visits was small, so this commit also increases the number of visits to make the test more likely to fail if it should actually fail.
See https://jenkins.softwareheritage.org/job/DSTO/job/tests-on-diff/1645/ for more details.