e5452b7fa7
Because the server has a finite amount of memory available to service incoming requests we can't handle smart HTTP requests when they are immediately received. Instead park them using the Jetty continuation features and resume processing only when we have reserved space in the SSH command queue. By reusing the same queue as the SSH daemon we ensure that the two daemons will share the same resources and avoid doubling our memory footprint. Change-Id: I4734b1e36315b82157609e93345e301eda9b8a8f Signed-off-by: Shawn O. Pearce <sop@google.com> |
||
|---|---|---|
| .. | ||
| .settings | ||
| src | ||
| .gitignore | ||
| pom.xml | ||