Hello Daniel > > I couldn't find anything in the release notes which look like a > > breaking change[0] > > Lots of people and CI systems run all the tests flawlessly all the time, so > there is reason to suspect that there is something slightly unusual in your > test setup that causes these problems. Meaning that I suspect this is not a > curl problem, this is a curl test problem. > > Are these problems different or the same as the ones already filed at > https://github.com/curl/curl/issues/10682 ?
I think it's a different issue, we have good evidence that those failures are triggered by an IPv6-only host and I acknowledge your solution to document that tests require IPv4 support for now. Salvatore can double check but the build logs indicate that the host used to build lnav had an IPv4 address. I also don't think lnav is running any of the curl's tests, but rather making use of the curl library to run their tests, which leads me to believe that whatever issue it is, it has to be something very specific and uncommon (and not related to curl's tests), otherwise there would be more reports. By the way I should have replied on that issue just in case: but feel free to close it if you'd like to track IPv6-only support somewhere else, or to rename it if you'd like to use it to track support for that. Me and sergiodj are thinking about giving it a try at solving that but we're not sure when (in the packaging, for now, we are ignoring those test results). Me and sergiodj are also currently investigating a test issue related to ppc64el, we have got some good insights already, but would like to fully understand what's going on and have a patch ready before reporting. Also that issue only affects curl's own tests so it can't be related to this. Cheers, -- Samuel Henrique <samueloph>