Conversation
|
@maskit What do you think about placing the SNI check in |
|
Sounds reasonable and I'm fine with it, however, I wonder how other implementations behave. Also, I should have known this but what happens after ATS rejects a connection at |
However, I noticed in the Autest that the curl/proxy verifier client times out. And doing a packet capture on the rejection case, I don't see ATS closes the quic connection(no |
|
Ok, I think this PR is fine. I don't think I implemented a way to initiate connection close on H3 layer. |
* asf/master: (221 commits) LSan: Fix leak of test_Metrics (apache#10179) LSan: Fix memory leak of test_EventSystem (apache#10178) LSan: Fix memory leak of test_X509HostnameValidator (apache#10161) Remove cqtx log field (apache#10111) Require ATS plugins to be compiled with C++17. (apache#10007) Fix conf_remap plugin build on macOS (apache#10177) libswoc: Update to 1.5.4 (apache#10155) Makes cmake build again, on macOS (apache#10172) Check SNI in h3 (apache#10184) Remove autoconf headers during CMake configuration (apache#10173) test_QUICLossDetector.cc: Add back get_hrtime() (apache#10185) ink_ink_get_hrtime -> ink_get_hrtime (apache#10182) mgmt: make libconfigmanager a true static library (apache#10181) Make sure that the thread local time is updated timely (apache#10163) Unrequire remap rules for OCSP (apache#10146) cache_range test performance improvement (apache#10170) Clean up certifier plugin debug messages. (apache#9975) cmake: add check for clock_gettime (apache#10169) Remove Http3NoError allocations (apache#10165) Fix Throttler initialization. (apache#10154) ...
This closes #10089.