Stefan Eissing
Stefan Eissing
Please do. I tried to filter a bit between fail/informative/debugging classifications, but you have more experience here.
Replaced by #14826
I like the idea. I am currently adding a `H2EarlyHintLink` directive that gets the complete content of a `Link:` header configured. Reason being that browser keep tweaking this header and...
> > Why not instead simply make `CURLSSLOPT_NATIVE_CA` work in more TLS backends? > > The system-cert-store behavior is indeed gated by `CURLSSLOPT_NATIVE_CA`, but implementing it for Apple platforms required...
> This is adding a new way to disable CA verification and/or to break it in suble ways, effectively disabling security. Well, we are not here to prevent people from...
Another point for having this verify callback in the curl API: As OCSP stapling is going away, browser are making their own mechanisms for checking certificate revocations (see [CRLite](https://blog.mozilla.org/security/2020/01/09/crlite-part-1-all-web-pki-revocations-compressed/) from...
@rcombs nice work, getting an all green in CI!🎉 If you would make a split of this, the `--native-ca` support on Apple OSs would be merged quite fast, I believe....
@vszakats is working on the GHA analysis job. This is appearing for all PRs right now.
Thanks or the report! I propose #17638 to fix this. Could you verify this in your setup? Thanks.
Made #18479 to fix the auto PONG handling. PONG frame is no longer inserted into the middle of a frame currently being assembled. Instead it is set aside to be...