Mark Thomas

Results 274 comments of Mark Thomas

Define that user provided encoders and decoders override the platform provided ones (e.g. for Integer-> String conversions)

> Otherwise, what would be the point ? Indeed. I think this is sufficiently obvious that we can clarify it in 2.0. I'll work on some updates.

Define that user provided encoders and decoders override the platform provided ones (e.g. for Integer-> String conversions)

Ah. As I look into this it gets more complex. I'm using `String` throughout here but the issue also includes `Reader`, `ByteBuffer`, `InputStream` and any other type explicitly mentioned when...

Define that user provided encoders and decoders override the platform provided ones (e.g. for Integer-> String conversions)

I like this idea. It is simple and clear.

Implement RFC6570 (URI Template) Level 2

This change would need to wait until Jakarta EE 10. Level 2 introduces reserved string expansion and fragment expansion. We can ignore fragment expansion as clients (should) never send the...

RemoteEndpoint.Async methods could provide better connection back to the Session

Exposing the session via the `SendResult` looks possible. The comments re `SessionException` appear to be implementation specific. Tomcat, for example, never uses it. Neither does the WebSocket API. I wonder...

RemoteEndpoint.Async methods could provide better connection back to the Session

@joakime What did you have in mind over and above adding `Session` to `SendResult` and returning `Future` rather than `Future`? Note: In the `Future` case, any errors will be exposed...

RemoteEndpoint.Async methods could provide better connection back to the Session

> So you see a use that might look like this? Yes, something along those lines.

RemoteEndpoint.Async methods could provide better connection back to the Session

I've been looking at this some more. `Future` -> `Future` isn't binary compatible so I don't see an easy way to do that. I think that is OK though as...

Need some resolution or clarification for how and when HttpSession last access time gets updated

I've starting a thread in the servlet project to discuss adding an `access()` method to `HttpServlet`.

Need some resolution or clarification for how and when HttpSession last access time gets updated

I tried to explain the point of this here: https://www.eclipse.org/lists/servlet-dev/msg00570.html Good point about HTTP/2 and HTTP/3 - I need to look into how authenticated WebSockets work in those scenarios. Any...