activemq-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Hiram Chirino <hi...@hiramchirino.com>
Subject Re: Null pointer exception in AnyProtocolCodec using WebSockets
Date Fri, 18 May 2012 15:29:00 GMT
On Thu, May 17, 2012 at 9:30 PM, Will Daniels <mail@willdaniels.co.uk>wrote:

> Hi Hiram,
>
>
> On 14/05/12 10:17, Hiram Chirino wrote:
>
>> This issue was reported in:
>> https://issues.apache.org/**jira/browse/APLO-203<https://issues.apache.org/jira/browse/APLO-203>
>>
>> It's now been fixed in the nightly snapshot.  Could verify it's working on
>> there?
>>
>
> Sorry, I had initially searched for reports of the problem the day before
> before writing to the list and had waited to first check that it worked OK
> in 1.2, so I missed that issue report that was logged in the meantime!
>
>
>  What error are you getting from the source build?
>>
>
> Well, aside from scala wanting to create filenames too long for my
> filesystem[1] (so I'm building in a tmpfs ramdisk instead) there was a
> problem with maven not finding the snapshot for hawtdispatch-scala v1.11 in
> the fusesource repository[2] which I fixed by changing:
>
> -<hawtdispatch-version>1.11</**hawtdispatch-version>
> +<hawtdispatch-version>1.12-**SNAPSHOT</hawtdispatch-**version>
>
>
I just tried a clean build against a clean local maven repo and everything
built ok for me.  Perhaps your local maven repo had a partially
downloaded hawtdispatch 1.11 jar or something.


> ...in the root pom.xml file.
>
> That doesn't seem quite right somehow (I'm completely guessing where it
> comes to maven and these snapshots/artifacts) although the apollo-util
> source then built OK, but it failed next running tests on some other
> component (maybe apollo-leveldb IIRC).
>
> I just ran the build again to confirm which component failed and to copy
> the output, but this time it all built successfully :S
>
> Anyway, thanks for the quick fix, it seems to be working now over
> WebSockets, except that it appears to timeout eventually and stop receiving
> messages, which I hadn't noticed happen with other clients on other
> transports. If you know off-hand that there is a setting for this, or can
> tell me that it shouldn't happen, would be appreciated.
>
>
Make sure your using the latest version of the stomp.js lib.  Older
versions could not deal with multiple STOMP frames being sent over 1
websocket message.


> One other thing, at the risk of going too far OT for dev list, I can't
> seem to get topic retained messages to work (on any transport). On
> subscribing to a topic that should have retained messages, I don't see any
> initial frames at all. But if I subscribe to the topic while the messages
> are sent, I can see that that "retain:true" header is correctly set on them.
>
> My understanding is that upon a new subscription to the topic, the last
> retained message should be sent? Am I misunderstanding what retain:true
> should do here? Is there any caveats I should be aware of about it?
>
>
That sounds about right.  Perhaps we have bug somewhere, I'll take a look.
 But please open issue on the issue tracker at:
https://issues.apache.org/jira/browse/APLO


> That may yet turn out to be a problem with me, not Apollo (I'm still
> investigating) but if anyone knows to confirm that it should be working in
> the current build, that would at least help me to look for the problem in
> the right places :P
>
> Cheers!
> -Will
>
> [1] http://www.lorrin.org/blog/**2011/11/17/error-file-name-**
> too-long-during-scala-compile/<http://www.lorrin.org/blog/2011/11/17/error-file-name-too-long-during-scala-compile/>
> [2] http://repo.fusesource.com/**nexus/content/groups/public-**
> snapshots/org/fusesource/**hawtdispatch/hawtdispatch-**
> scala/1.11-SNAPSHOT/<http://repo.fusesource.com/nexus/content/groups/public-snapshots/org/fusesource/hawtdispatch/hawtdispatch-scala/1.11-SNAPSHOT/>
>
>
>>
>>
>>> Any advice about this issue would be much appreciated!
>>>
>>>
>>>  Just try the nightly development snapshot.  You can find links to it on
>> the
>> download page [1]. It should be much better.
>>
>> [1]: http://activemq.apache.org/**apollo/download.html<http://activemq.apache.org/apollo/download.html>
>>
>> Thanks,
>>
>>> -Will
>>>
>>>
>>> [1] https://github.com/apache/****activemq-apollo/commit/**<https://github.com/apache/**activemq-apollo/commit/**>
>>> 5862484bc44c7f46457cb19cf6638b****fd3584595b<https://github.**
>>> com/apache/activemq-apollo/**commit/**5862484bc44c7f46457cb19cf6638b**
>>> fd3584595b<https://github.com/apache/activemq-apollo/commit/5862484bc44c7f46457cb19cf6638bfd3584595b>
>>> >
>>> [2] https://github.com/chirino/****stomp-websocket/blob/master/**<https://github.com/chirino/**stomp-websocket/blob/master/**>
>>> dist/stomp.js<https://github.**com/chirino/stomp-websocket/**
>>> blob/master/dist/stomp.js<https://github.com/chirino/stomp-websocket/blob/master/dist/stomp.js>
>>> >
>>>
>>>
>>
>>
>>


-- 

**

*Hiram Chirino*

*Software Fellow | FuseSource Corp.*

*chirino@fusesource.com | fusesource.com*

*skype: hiramchirino | twitter: @hiramchirino<http://twitter.com/hiramchirino>
*

*blog: Hiram Chirino's Bit Mojo <http://hiramchirino.com/blog/>*

*
*

*
*

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