1
0
mirror of https://github.com/danog/amp.git synced 2024-12-12 09:29:45 +01:00
Commit Graph

27 Commits

Author SHA1 Message Date
Niklas Keller
c1a4d9346d Add warning about locally closed resources being undefined behavior
Also makes Loop docs consistent with Driver. Resolves #106.
2016-10-25 17:35:21 +02:00
Niklas Keller
27e7f2a05b Define error handler behavior
Resolves #95.
2016-10-24 20:43:21 +02:00
Niklas Keller
77764d5b89 Format docs, change LogicException to Exception 2016-09-25 13:22:01 +02:00
Bob Weinand
74ff6a4be1 Clarify that (half-)closed streams also trigger the respective read and write watchers 2016-09-23 16:02:23 +02:00
Niklas Keller
3791a7536a Add missing documentation 2016-09-22 09:15:45 +02:00
Bob Weinand
b3fdfa4ede Fixed #90 - allow disabling of invalid watchers 2016-08-30 22:04:21 +02:00
Niklas Keller
74e2577b25 Rename fetch/storeState to get/setState 2016-08-06 14:49:08 +02:00
Niklas Keller
8ece6be6cf Make final PSR-2 compatible 2016-08-05 14:11:59 +02:00
Niklas Keller
39b7ddd366 Make abstract PSR-2 compatible 2016-08-05 13:56:26 +02:00
Niklas Keller
6cde9f5495 Merge Registry into Driver 2016-08-04 21:08:52 +02:00
Bob Weinand
9f3f6b9523 Add a warning/implementation note about installing handlers for a same signal on different loops
This needs to be undefined behavior as we do not have the possibilities in PHP (i.e. the language and functions don't give us the power) to distinguish properly whether a signal is already installed or not
Also, when we try to set a handler for a signal we've already installed a signal elsewhere, the second onSignal() call will lead implementations to override the original handler (on the system level, with signal()/sigaction())
Ultimately some extensions just flat out emit a warning in that case (e.g. libev).

Thus, we cannot control what happens and hence this needs to be specified as undefined behavior.
2016-07-19 18:10:51 +02:00
Niklas Keller
8f38808ddb Fix line lengths, resolves #76 (#81) 2016-07-15 22:17:24 +02:00
Aaron Piotrowski
8c9efe42aa Undefined callback order (#80) 2016-06-26 16:15:01 +02:00
Bob Weinand
7b30e788bb Specify when exactly a watchers callbacks may start to be called
Delaying activation of watchers to the next tick guarantees consistent behavior between all implementations.

Additionally it eliminates the following problem:
- Two defer()'s are set up with the first one immediately disabled
- Second defer()'s callback is called and enables the first defer()'s watcher
- Executes defer() in same tick or next tick?
If the answer is same tick, what happens in the following scenario:
- An immediately disabled defer() and a delay($msDelay = 0) are set up
- The delay()'s callback enables the defer()'s watcher
- Executes defer() callback in same tick or next tick?
Doing it in same tick violates the requirement that defers are called before any other watchers in the same tick.

We get similar problems with e.g. delay() and repeat() (if their timer all expired now, but we require order to be preserved).

Thus we need to pretty much delay the activation of enable() to the next tick to avoid potential ambiguity or different behaviors depending on when exactly the enable() or watcher registration happened.
2016-06-07 21:17:27 +02:00
Bob Weinand
1e584697b3 Be a little more specific about enable/definition
Enable is when things get put back to the queue, hence that order shall be respected. Otherwise we get undefined behaviors when disabling deferreds which have yet to be executed and re-enabling them after the point they should have been executed.
2016-06-07 01:44:28 +02:00
Andrew Carter
ccb2180184 Clarified that cancel() can be used to prevent a memory leak (#75) 2016-06-06 22:37:17 +02:00
Aaron Piotrowski
289b0fe603 Merge pull request #74 from async-interop/calltime_guarantees
Specify exact guarantees when a callback has to be called
2016-06-05 22:45:14 -05:00
Aaron Piotrowski
c05ba0daa9 Merge pull request #72 from async-interop/invalid-watcher-handling
(Re-)Introduce InvalidWatcherException
2016-06-05 22:44:42 -05:00
Bob Weinand
cf012501a2 Specify exact guarantees when a callback has to be called
This prevents unexpected failures due to different execution orders in different implementations.
2016-06-04 20:28:25 +02:00
Bob Weinand
f6c6f41adf (Re-)Introduce InvalidWatcherException
Fixes #58
2016-06-02 21:56:36 +02:00
Bob Weinand
7cd03f3927 The watcher id must be invalidated before defer and delay callback calls
Fixes #70.
2016-06-02 19:15:52 +02:00
Aaron Piotrowski
4908813e08 Refactor Registry for Drivers 2016-05-27 10:45:45 -05:00
Bob Weinand
104b761e59 Revert "Merge pull request #63 from async-interop/throw-on-invalid-watcher"
This reverts commit 5bd7794dfa, reversing
changes made to 64494951ff.

Revert too quick merge (sorry) as per #58
2016-05-27 14:57:11 +02:00
Aaron Piotrowski
b8f0d521ba Add note that disable must not invalidate 2016-05-26 12:36:56 -05:00
Aaron Piotrowski
b7ed0afd02 Add "or cancelled" to throw condition 2016-05-26 11:10:44 -05:00
Aaron Piotrowski
7c34146ac8 Add InvalidWatcherException 2016-05-26 10:43:03 -05:00
Aaron Piotrowski
7a2c5cdc88 Cleanup namespace 2016-05-26 10:05:17 -05:00