qpid-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Andrew Stitcher" <astitc...@apache.org>
Subject Re: Review Request: proton-c: support for AMQP 1.0 connection idle timeout negotiation and keepalives
Date Tue, 13 Nov 2012 15:11:57 GMT


> On Nov. 12, 2012, 10:32 p.m., Andrew Stitcher wrote:
> > /proton/trunk/proton-c/src/driver.c, line 843
> > <https://reviews.apache.org/r/8021/diff/1/?file=188831#file188831line843>
> >
> >     gettimeofday() is obsolete (read man 2 gettimeofday) use clock_gettime in preference
(although this then requires -lrt, but that's not a bad thing in itself).
> 
> Kenneth Giusti wrote:
>     Ok - the original code used "gettimeofday" in a couple of places - specifically in
messenger.  Seemed like that functionality was better abstracted to the driver.
>     
>     I'll try clock_gettime()...

As the both gettimeofday() and clock_gettime() are not ANSI and are POSIX platform dependent
they should definitely be put in the driver not in messenger, which should ideally be ANSI
as there's no real reason for it to contain platform dependent code itself.


- Andrew


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/8021/#review13372
-----------------------------------------------------------


On Nov. 13, 2012, 3:08 p.m., Kenneth Giusti wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/8021/
> -----------------------------------------------------------
> 
> (Updated Nov. 13, 2012, 3:08 p.m.)
> 
> 
> Review request for qpid, Andrew Stitcher and Rafael Schloming.
> 
> 
> Description
> -------
> 
> Attached is an attempt at adding Connection idle-time-out support to Proton-c, as defined
by the AMQP 1.0 spec. 
> 
> Notes:
> 
> 1) Allows configuration of the AMQP Connection idle timeout locally, and support for
receiving the remote's configured idle timeout.
> 2) Will send periodic keepalive (empty) frames to satisfy remote's idle requirements,
if needed.
> 3) Will close the connection if remote violates the locally configured idle deadline
> 
> 
> This addresses bug proton-111.
>     https://issues.apache.org/jira/browse/proton-111
> 
> 
> Diffs
> -----
> 
>   /proton/trunk/proton-c/CMakeLists.txt 1408412 
>   /proton/trunk/proton-c/bindings/python/proton.py 1408412 
>   /proton/trunk/proton-c/include/proton/driver.h 1408412 
>   /proton/trunk/proton-c/include/proton/engine.h 1408412 
>   /proton/trunk/proton-c/include/proton/types.h 1408412 
>   /proton/trunk/proton-c/src/dispatcher/dispatcher.h 1408412 
>   /proton/trunk/proton-c/src/dispatcher/dispatcher.c 1408412 
>   /proton/trunk/proton-c/src/driver.c 1408412 
>   /proton/trunk/proton-c/src/engine/engine-internal.h 1408412 
>   /proton/trunk/proton-c/src/engine/engine.c 1408412 
>   /proton/trunk/proton-c/src/messenger.c 1408412 
>   /proton/trunk/proton-c/src/util.h 1408412 
>   /proton/trunk/proton-c/src/util.c 1408412 
>   /proton/trunk/proton-j/src/main/scripts/proton.py 1408412 
>   /proton/trunk/tests/proton_tests/engine.py 1408412 
> 
> Diff: https://reviews.apache.org/r/8021/diff/
> 
> 
> Testing
> -------
> 
> Added an engine test case to verify that the timers are tracked, and the action when
timers fire, but I need to add a test at the driver level, too.
> 
> 
> Thanks,
> 
> Kenneth Giusti
> 
>


Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message