-
Notifications
You must be signed in to change notification settings - Fork 114
Added dynamic routing and redesigned API #46
Conversation
Upgraded .NET Framework to 4.6 Updated example to use web sockets on Kestrel
Added RequestAborted cancellation token support Added basic error handling for connect Replaced CloseAsync with CloseOutputAsync
Renamed iterator
Renamed test method Code formatting Added constants
Improved variable naming
… into dynamic-routing
@mkosieradzki, |
Redesign in-progress. |
Expose utility methods for advanced scenarios using ProxyUtils
@@ -0,0 +1,34 @@ | |||
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Need a copyright and license header here
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks good overall. After we get Travis passing this should be good to merge.
@mikaelm12 Thanks, any idea why is Travis failing? Current source code is failing as well. This specific PR is not introducing any new failed test. Shouldn't we fix travis in a separate PR, because this one is huge enough ;). It looks like a problem with ClientWebSocket on non-windows platform. I was trying to check other WebSocket-related repos but they have even AppVeyor failing. |
Any ideas @davidfowl @anurse? |
@mikaelm12 Just letting you know this is ready to merge. |
@mkosieradzki I haven't forgotten about this 😄 |
@mikaelm12 But this one has nothing to do with the WebSocket test, WebSockets were in the previous commit (but the test didn't fail because travis was down). However I can try to work on the websockets issue. |
Any progress on dynamic routing? |
I have updated this PR and fixed one more issue (cookie support - 1 liner). @mikaelm12 - please merge if possible (again this PR has nothing to do with the flaky WebSockets test). BTW. I can try to investigate this flaky test (but it is unrelated to this PR). |
Why is there a public ProxyService class? |
@davidfowl |
@Tratcher can you clarify? @mkosieradzki It doesn't look as though it's required at all. Also, you didn't rename SharedProxyOptions to ProxyOptions |
@davidfowl Please see (Feb 28):
|
I don't recall the reasoning for using a service. The only discussion I can find for it is here: #46 (comment) You should be able to remove the service, inject the options into the middleware, and flow them from there. |
Added dynamic routing and redesigned API
Fixed cookie support
This PR addresses:
#37, #35, #34 ?, #27, #10, #43, #47
NOTES:
If you are not convinced about merging QueryString parameters - this is an extremely useful feature when forwarding to the Service Fabric Reverse Proxy. You are often adding query string parameters like PartitionKey=.