Put the licence in the README. |
|
More...
|
over 12 years ago
|
RoutingTable.Copy: needed for UDP transport refactoring. |
|
More...
|
over 13 years ago
|
Updated documentation. |
|
More...
|
over 13 years ago
|
TPortNum |
|
More...
|
over 13 years ago
|
Still trying to chase down the timing difference between my machine and Kelvin. |
|
More...
|
over 13 years ago
|
ThreadEvent is far too overused. Use a separate event so we can more accurately track things. |
|
More...
|
over 13 years ago
|
* Fold Bind() into Listen(). * Disallow writing to a listening connection. |
|
More...
|
over 13 years ago
|
Simplified PartialSend. |
|
More...
|
over 13 years ago
|
Interim commit. I THINK all the tests pass? |
|
More...
|
over 13 years ago
|
Ignore. Whitespace change. |
|
More...
|
over 13 years ago
|
And then actually commit the actual code. |
|
More...
|
over 13 years ago
|
TIdSipLocations gets a set difference function |
|
More...
|
over 13 years ago
|
I'm sure I already committed this file...? |
|
More...
|
over 13 years ago
|
Part #2 of connection/server interface simplification. |
|
More...
|
over 13 years ago
|
Uniform error message for rejected message waits. |
|
More...
|
over 13 years ago
|
Part 1 of simplifying connection/server comms: OnAccept, OnConnect, OnDisconnect, OnMessageSent no longer use Notifications to tell the server to schedule a Wait, but simply schedule a Wait directly. (This has the upside of being simpler, and the downside of being less flexible: Connections now MUST know a TimerQueue's reference, and will schedule Waits on that TQ directly.) |
|
More...
|
over 13 years ago
|
When a WaitFor fails, it's nice to know why. |
|
More...
|
over 13 years ago
|
Oops. You need MessageQueues referenced in the dpr, not just added to the repo! |
|
More...
|
over 13 years ago
|
Async connect means you can only set the Via's sent-by/port when the socket's actually connected. We set the "advertised port" - what goes in the topmost Via's port field - so that the ClientConnection rewrites the necessary headers just before swizzling the messages down the wire. |
|
More...
|
over 13 years ago
|
PR 854: Messages sent from non-standard setups can have wrong Via headers. |
|
More...
|
over 13 years ago
|
PR 853: Threaded TimerQueue Terminated -> access violation. |
|
More...
|
over 13 years ago
|
Ensure that the port also changes; this test passed _accidentally_. |
|
More...
|
over 13 years ago
|
Simpler test case that also stops a stupid error in the test: the client connection closes before the transport has a chance to read the message! |
|
More...
|
over 13 years ago
|
Nearly finished! Still broken: because Connect() is asynchronous we must somehow defer RewriteLocationHeaders() to after the socket's actually connected... and still maintain the socket/message association. |
|
More...
|
over 13 years ago
|
Comment moving: now in a saner spot! |
|
More...
|
over 13 years ago
|
And even handier is committing the actual implementation. |
|
More...
|
over 13 years ago
|
Handy for matching a Location against a socket description. |
|
More...
|
over 13 years ago
|
Interim commit. Basic Connection replacements. TLS broken until further notice. |
|
More...
|
over 13 years ago
|
Lots of tests need to check whether something's bound to a particular ip:port/transport. IdNetworking's the right place to put this. |
|
More...
|
over 13 years ago
|
TIdTimerQueue.TerminateAndWaitFor() - handy for when you have to wait for a TQ to actually die. |
|
More...
|
over 13 years ago
|