Please can you check "Possible typo in messenger.c" message?
Is this a real typo?
(this is present in earlier versions also, not just in 0.11)
Also I observe library looping in the following code:
#0 pn_selector_select (selector=0x22bff80, timeout=600000) at
qpid-proton-0.11.0-alpha/proton-c/src/posix/selector.c:156
#1 0x00007f5f6cb4a7bb in pni_wait (messenger=0x22bf610, timeout=600000) at
qpid-proton-0.11.0-alpha/proton-c/src/messenger/messenger.c:1424
#2 0x00007f5f6cb4a8f4 in pn_messenger_tsync (messenger=0x22bf610,
predicate=0x7f5f6cb4c9af <work_pred>, timeout=-1)
at qpid-proton-0.11.0-alpha/proton-c/src/messenger/messenger.c:1465
#3 0x00007f5f6cb4c9f4 in pn_messenger_work (messenger=0x22bf610, timeout=-1)
at qpid-proton-0.11.0-alpha/proton-c/src/messenger/messenger.c:2151
#4 0x00007f5f6cb4ca45 in pni_messenger_work (messenger=0x22bf610) at
qpid-proton-0.11.0-alpha/proton-c/src/messenger/messenger.c:2161
#5 0x00007f5f6cb4d92d in pn_subscription_address (sub=0x22d2f80) at
qpid-proton-0.11.0-alpha/proton-c/src/messenger/subscription.c:130
What happens here, I am testing error handling in my program and I start it
with qpidd broker down. The progam executes pn_messenger() and
pm_messenger_start()
and then tries to subscribe to a queue. It gets no errors from messenger create
and start calls, but when it invokes pn_subscription_address() it starts to loop
inside pn_messenger_tsync(). When I set messenger timeout to some value this at
least allows to return from this loop eventually (and actually to recover from
the qpidd restart) but during messenger timeout intervals it still loops taking
up to 100% cpu.
best regards,
27.10.2015 16:33, Justin Ross пишет:
> I've received precisely zero feedback on the alpha release. Should I
> proceed with the beta and branching as planned, or is there a desire for
> more time to do some testing?
>
> Justin
>
> On Mon, Oct 19, 2015 at 8:20 AM, Justin Ross <justin.r...@gmail.com> wrote:
>
>> Hi, folks. The alpha is now available from the following URL:
>>
>> https://dist.apache.org/repos/dist/dev/qpid/proton/0.11.0-alpha/
>>
>> I also ran automated tests for the related components at Qpid:
>>
>>
>> http://people.apache.org/~jross/qpid-releases/quirk-proton-0.11.0-alpha.log
>>
>> The proton tests themselves are passing. There is one C++ failure in the
>> ssl tests that might be related to trunk proton (or might not, hard to
>> tell). All my testing was on x86-64 Fedora 22.
>>
>> I've updated the schedule on the release page[1]. Beta is set for Monday,
>> the 26th, and RC is set for 2 November. As you can tell, this is a
>> relatively compressed schedule. Please prioritize testing so we can
>> discover and resolve issues for the release.
>>
>> Thanks!
>> Justin
>>
>> ---
>> [1] Proton 0.11.0 release page:
>> https://cwiki.apache.org/confluence/display/qpid/Qpid+Proton+0.11.0
>>
>
--
\ / | |
(OvO) | Михаил Иванов |
(^^^) | Тел.: +7(911) 223-1300 |
\^/ | E-mail: iv...@isle.spb.ru |
^ ^ | |
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscr...@qpid.apache.org
For additional commands, e-mail: users-h...@qpid.apache.org
