2012-05-08 16:17:56 |
cyli |
bug |
|
|
added bug |
2012-05-08 16:17:56 |
cyli |
attachment added |
|
txamqp_protocol_clock.patch https://bugs.launchpad.net/bugs/996614/+attachment/3137078/+files/txamqp_protocol_clock.patch |
|
2012-05-08 17:24:42 |
cyli |
description |
I'm not entirely sure if this is a bug or if this was intended, but it seems like the clock that is passed in and used as the clock to check the heartbeat should be the same clock used to make the heartbeat.
clock.callLater is used, but LoopingCall is also used, without setting the clock of the looping call.
Also, time.time() is used for the last heartbeat received and last heartbeat sent - if the actual seconds since the epoch don't need to be used, reactor/clock.seconds() can be used instead of time() (currently, reactor.seconds() just returns time.time(), but this may change in the future to return the time since the reactor has been started). |
I'm not entirely sure if this is a bug or if this was intended, but it seems like the clock that is passed in and used as the clock to check the heartbeat should be the same clock used to make the heartbeat.
clock.callLater is used, but LoopingCall is also used, without setting the clock of the looping call.
Also, this is not in the patch, but time.time() is used for the last heartbeat received and last heartbeat sent - if the actual seconds since the epoch don't need to be used, reactor/clock.seconds() can be used instead of time() (currently, reactor.seconds() just returns time.time(), but this may change in the future to return the time since the reactor has been started). |
|
2012-05-08 17:24:49 |
cyli |
attachment removed |
txamqp_protocol_clock.patch https://bugs.launchpad.net/txamqp/+bug/996614/+attachment/3137078/+files/txamqp_protocol_clock.patch |
|
|
2012-05-08 17:25:37 |
cyli |
attachment added |
|
txamqp_protocol_clock.patch https://bugs.launchpad.net/txamqp/+bug/996614/+attachment/3137149/+files/txamqp_protocol_clock.patch |
|
2012-05-08 17:26:23 |
cyli |
description |
I'm not entirely sure if this is a bug or if this was intended, but it seems like the clock that is passed in and used as the clock to check the heartbeat should be the same clock used to make the heartbeat.
clock.callLater is used, but LoopingCall is also used, without setting the clock of the looping call.
Also, this is not in the patch, but time.time() is used for the last heartbeat received and last heartbeat sent - if the actual seconds since the epoch don't need to be used, reactor/clock.seconds() can be used instead of time() (currently, reactor.seconds() just returns time.time(), but this may change in the future to return the time since the reactor has been started). |
I'm not entirely sure if this is a bug or if this was intended, but it seems like the clock that is passed in and used as the clock to check the heartbeat should be the same clock used to make the heartbeat.
clock.callLater is used, but LoopingCall is also used, without setting the clock of the looping call.
Also, this is not in the patch, but time.time() is used for the last heartbeat received and last heartbeat sent - if the actual seconds since the epoch don't need to be used, reactor/clock.seconds() can be used instead of time(). (For those who want to use the clock for timing). |
|
2013-09-29 09:39:20 |
Esteve Fernandez |
txamqp: status |
New |
Fix Released |
|