Jay Satiro
Jay Satiro
Any opinions on this?
I don't see how this is our problem. It looks to me like the atexit registered OpenSSL_cleanup is called, then some other library's atexit registered function is called, but that...
> and then later on when pkg_shutdown is called (another atexit registered by pkg), you end up calling curl_multi_cleanup: I see, so you are saying that library pkg's atexit function...
I'm still not convinced this is our problem
curl for windows 8.4.0 returns CURLE_BAD_FUNCTION_ARGUMENT (43) and the [daily](https://github.com/curl/curl-for-win/actions/runs/7038183128) build curl-8.5.0-DEV_20231129-win64-mingw returns CURLE_WEIRD_SERVER_REPLY (8) 8.4.0: ~~~ == Info: Trying 162.159.138.85:443... == Info: Closing connection curl: (43) A libcurl function...
> https://github.com/curl/curl-for-win/actions/runs/6632238595 [7fdf30f](https://github.com/curl/curl/commit/7fdf30f25604195245b05c6bf6f3b10bee6af36a) bad argument > https://github.com/curl/curl-for-win/actions/runs/6645832682 [6ec70a9](https://github.com/curl/curl/commit/6ec70a9dd305e17b017f1283a034dfa1c03bbd7c) work Yeah I see the same. I notice those curl commits don't have any bearing on QUIC and that otherwise the manifest...
> Is this still happening? AFAICT no. I tested [curl-8.6.1-DEV_20240212-win64-mingw](https://github.com/curl/curl-for-win/actions/runs/7877559459/artifacts/1239549651) ([backup link for posterity](https://github.com/curl/curl/files/14256564/curl-8.6.1-DEV_20240212-win64-mingw-test.zip)) and I see a bunch of "weird server reply" to all the QUIC but after that...
> AFAIK there was no related (intented) change in curl between these versions, nor any curl-for-win (official Windows build) updates that might introduce such difference. curl-for-win had Unicode enabled only...
> In particular, I've seen one server where it is impossible to upload a file of size 9977696 bytes. The client always runs on windows 10 and I have no...
[1.9 Cache negative name resolves](https://curl.se/docs/todo.html#Cache_negative_name_resolves): > _"A name resolve that has failed is likely to fail when made again within a short period of time. Currently we only cache positive...