Time Nick Message 04:06 MTDiscord Ah, wonderful, there's an unhandled exception in minetestserver that I somehow exposed by running the unittests in a docker container. 04:13 MTDiscord I would like feedback on https://github.com/minetest/minetest/pull/13573. 11:58 sfan5 when was vector.new added again? 5.0 probably right? 11:58 sfan5 if we had better documentation it would contain this information 12:09 sfan5 0.4.8 actually lmao 12:32 rubenwardy agreed, version info like that would be nice 12:32 rubenwardy wasn't that one of the goals of minetest_docs? 17:19 MTDiscord We have an ipv6 unit test that can only run out of a docker container on Linux hosts after the experimental ipv6 support has been manually enabled. 17:20 MTDiscord Maybe it would be better classified as an integration test. I'm wondering whether it would be better to disable the test on docker, or try to move it to some kind of integration suite. 17:39 sfan5 which unit test would that be? 18:18 MTDiscord TtstIPv6Socket 18:18 MTDiscord TestIPv6Socket* 18:21 sfan5 it already has a graceful fallback 18:22 MTDiscord By graceful fallback do you mean throwing an unhandled exception in the server and resulting in a test run failure? 18:24 sfan5 no 18:26 * schwarzwald[m] posted a file: (13KiB) < https://libera.ems.host/_matrix/media/v3/download/matrix.org/MnOCjnqMzPtJzsaTBhpYIOcC/log.txt > 18:26 MTDiscord That's the log. I'll look into it and see whether I can add a regression test. 18:27 sfan5 seems weird that it can open and bind the socket and everything but fails to send the packet 18:30 sfan5 needs more investigation 18:34 MTDiscord I ran fresh copy of the same build configuration on the host Manjaro VM and the test passed there. It does seem to be related to docker's ipv6 restrictions. The log shows the related ipv4 test passing. 19:00 MTDiscord The error is "Address not available", which doesn't seem to be documented in the man page for sendto. 19:06 MTDiscord The error code is probably from bind indicating that the requested address is not available from the local machine. Interesting. 19:07 MTDiscord I've been on the computer since 5 am (it's now 2 pm) so I should take a break from this. At least there's some progress. 19:39 MTDiscord rubenwardy: no, documenting decade-old features was never a goal; we're currently targeting latest master, and even that we can not keep up with... 19:39 MTDiscord documenting the age of these features was never a goal* 19:41 MTDiscord Though it isn't a bad one to add to a future list 20:09 sfan5 it's not about documenting that the feature is a decade old 20:09 sfan5 vector.new is a bad example 20:10 sfan5 but for other functions is very valid to wonder "was this added in 5.5" when you're e.g. targeting your game or mod at two minor versions back 20:16 MTDiscord indeed