Github user andrewcox commented on the pull request:
https://github.com/apache/thrift/pull/807#issuecomment-173754745
One mitigating factor here is our "ThriftServer" class (our most used
server) doesn't use "THeaderTransport" at all, it uses the "THeader" core.
And then some of our servers that *do* use THeaderTransport do it through
TAsyncChannel in a way that it ends up reading a buffer that already contains a
full message, so it would never stall in the middle of a message.
Still though, I believe at least some of our legacy servers still use
THeaderTransport and synchronous I/O, and we've never seen this AFAICT. That's
why I'd like to see the callstack.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
