Viktor Szakats
Viktor Szakats
Found this in every log (not sure how I missed it before!): ``` curl: (69) TFTP: Access Violation ``` Ref: https://github.com/curl/curl/actions/runs/9076301772/job/24938697525#step:3:3674 edit: Seems like a false lead...
@orgads: Yes, the MSYS2 mingw-w64 ones sometimes **hang** in tests. I'm tracking the problem here: https://github.com/curl/curl/pull/13599#issuecomment-2119372376. Could not yet figure out _which_ test is prone to hanging though.
Run into a CRLF vs LF failure in test 1478 with all 3 MSYS jobs (and not with Cygwin). Expected is CRLF, generated is LF.
Runs OK now, has OpenSSL + libpsl (plus a bunch of others it detected automaticall). Tests pass. Runtimes compared to AppVeyor jobs (no SSL) (tested on a weekend): ``` AppVeyor...
Plan to merge this today with the disclaimer that tests may need tweaking if we see flakiness or failures once these jobs get more use.
Last minute I enabled WebSockets in the mingw-w64 job and 2301 and 2305 started failing right away: https://github.com/curl/curl/actions/runs/9133449443/job/25116951463#step:14:14502 https://github.com/curl/curl/actions/runs/9133227792/job/25116241601#step:14:18246 It's also the first time WebSocket is tested on native Windows...
Post merge update: mingw-w64 tests seem to _hang_, which is new compared to AppVeyor CI and wasn't typical while testing this. Is there a way to tell which test is...
Is this option causing an actual problem in the curl codebase?
How and why would these slow down the builds? Also with the explicit inlines just added, these warnings now seem useful to know when those don't actually work, or am...
> > How and why would these slow down the builds? > > As you pointed out in the discussion of the #13355, extra processing in `configure` or CMake slow...