Jetty 12 idletimeout #9905
Jetty 12 idletimeout #9905
Conversation
+ pass TimeoutException through all APIs + HttpConnection now passes on TimeoutException to HttpChannel.onFailure
Signed-off-by: Simone Bordet <[email protected]>
* Improved reset of the earliest timeout before iteration. * Removed check for getExpireNanoTime() == -1, since it's a valid value. * When onExpired(Expirable) returns false, the Expirable should arrange to move its timeout in the future. Signed-off-by: Simone Bordet <[email protected]>
Signed-off-by: Ludovic Orban <[email protected]>
….failed() in all cases to complete the wrapped callback. Signed-off-by: Simone Bordet <[email protected]>
…ll super.failed() in all cases to complete the wrapped callback." This reverts commit 5ac57c1.
|
@lorban @lachlan-roberts The key changes in this PR are:
|
# Conflicts: # jetty-ee10/jetty-ee10-servlet/src/main/java/org/eclipse/jetty/ee10/servlet/ServletChannel.java # jetty-ee9/jetty-ee9-nested/src/main/java/org/eclipse/jetty/ee9/nested/HttpChannel.java
| @@ -1202,6 +1203,23 @@ public void call(Invocable.Callable callable, Request request) throws Exception | |||
| } | |||
| } | |||
|
|
|||
| public <T> boolean test(Predicate<T> predicate, T t, Request request) | |||
There was a problem hiding this comment.
This looks unused. Should we keep it nevertheless?
There was a problem hiding this comment.
@lorban It has been used before, but now that we don't have listening predicates..... oh but wait yes we do. This should be used by the context addIdleTimeoutListener wrapper that is not implemented!!!! stand by.....
There was a problem hiding this comment.
@lorban I've added the context wrapper so idle timeouts are callback in the context. I've also added a test to check that. Please re-review... at least that last commit.
|
|
||
| // TODO what do we do about the failing write? | ||
| // should timeout errors be non persistent? | ||
| Callback ocb = callback; |
There was a problem hiding this comment.
Should we keep this wrapping?
There was a problem hiding this comment.
long term no, but whilst we still have work to do....
| } | ||
|
|
||
| @Test | ||
| public void testIdleTimeoutNoListenerHttpConfigurationOnly() throws Exception |
There was a problem hiding this comment.
I fail to see what this test asserts that the above one (testIdleTimeoutNoListener()) does not.
There was a problem hiding this comment.
This one has the idle timeout configured in HttpConfiguration only, so it is testing that it is correctly applied by HttpChannel. I.e the idle timeout between requests is 10x the idle timeout during requests.
There was a problem hiding this comment.
(probably could have been parameterised... will do when I next touch this test.)
...t/src/main/java/org/eclipse/jetty/fcgi/client/transport/internal/HttpConnectionOverFCGI.java
Outdated
Show resolved
Hide resolved
jetty-core/jetty-io/src/test/java/org/eclipse/jetty/io/ByteArrayEndPointTest.java
Show resolved
Hide resolved
jetty-core/jetty-server/src/main/java/org/eclipse/jetty/server/internal/HttpChannelState.java
Show resolved
Hide resolved
jetty-core/jetty-server/src/main/java/org/eclipse/jetty/server/internal/HttpChannelState.java
Show resolved
Hide resolved
Rework idle timeout mechanism, undisable tests