[10:14:51] <tgr|away>	 Krinkle: https://gerrit.wikimedia.org/r/c/mediawiki/core/+/1060787 was a followup to the failures in that task (I just didn't get around to it before the hackathon). There were no isolated errors AFAIK, but back when we had those errors people said the test shouldn't fail due to network issues.
[10:15:27] <tgr|away>	 I don't have strong feelings about it either way, can revert if you want.
[13:45:59] <Krinkle>	 tgr|away: ah this is for local running? I can see how local running indeed, assuming composer deps are up to date, would not need the network otherwise. Especially when running tests for the first time. After first time, it stays in a local dot directory indefinitely, so shouldn't fail on re-run. But during hackathons or other network reliability scenarios, anything you can cut can help yeah. Maybe not in CI though where this should only 
[13:45:59] <Krinkle>	 contact network now if you make a patch that changes libs. So maybe we can limit it to non-Quibble (ie local phpunit).
[15:11:22] <tgr|away>	 Krinkle: I think the idea was that since the test is not about testing a network capability, it should not be sensitive to network issues. It was about CI, not local development. Maybe MatmaRex remembers (when he is here), I can't recall where the discussion happened. It was somewhere in the orbit of T362425.
[15:11:23] <stashbot>	 T362425: ForeignResourceStructureTest flaky in CI due to "Failed to download resource at https://codeload.github.com" - https://phabricator.wikimedia.org/T362425
[23:25:54] <Krinkle>	 xSavitar: fyi, regarding stats lib - T368786
[23:25:55] <stashbot>	 T368786: StatsLib feature to assist users in situations of recursion - https://phabricator.wikimedia.org/T368786