Info

Login

Channels

APIs

Credits

  • cheeser
  • ernimril
  • joed
  • kinabalu
  • lunk
  • ojacobson
  • r0bby
  • ThaDon
  • ricky_clarkson
  • topriddy

« 2019-07-10

2019-07-11

2019-07-12 »

Nick Message Date
tmielke [tmielke!~tmielke@cable-158-181-75-18.cust.telecolumbus.net] has joined #apache-activemq [03:02]
bambitroll [bambitroll!~jmr@193.160.28.38] has joined #apache-activemq [03:03]
calohmn [calohmn!~cal@bln-1-02.inubit.com] has joined #apache-activemq [03:04]
wallbroken [wallbroken!~Thunderbi@unaffiliated/wallbroken] has joined #apache-activemq [03:22]
dejanb [dejanb!~dejanb@dynamic-109-121-74-4.adsl.eunet.rs] has joined #apache-activemq [03:53]
wallbroken julemand101: you never geave me an answer [06:48]
jbertram [jbertram!~jbertram@redhat/jboss/jbertram] has joined #apache-activemq [08:27]
tmielke parted the channel: [08:57]
rkieley [rkieley!~rkieley@24.222.88.212] has joined #apache-activemq [09:19]
tbish [tbish!~timothy@c-69-243-80-114.hsd1.va.comcast.net] has joined #apache-activemq [09:22]
hacst jbertram: Are NullPointerExceptions something to worry about with artemis? Triggered one with my AMQP fumbling: https://pastebin.com/fg6VT4D4 . This was with artemis 2.6.3 I think. [10:15]
jbertram hacst: looks like a bug to me...can you reproduce this in 2.9.0 (i.e. the latest release)? [10:17]
tmielke [tmielke!~tmielke@2a01:598:a80a:2f9a:58ee:8cf2:e6b7:7690] has joined #apache-activemq [10:18]
jbertram hacst: looking at the code it appears that NPE should be resolved in 2.9.0 [10:19]
hacst I see. I was using a docker container for the experiments. Probably not the newest one [10:24]
tmielke [tmielke!~tmielke@2a01:598:a80a:2f9a:5839:91f4:2891:9172] has joined #apache-activemq [10:27]
hacst jbertram: Nope. Updated. Same exception in 2.9.0: https://pastebin.com/9AvX9wPJ [10:35]
hacst (well...not the same but still NPE) [10:35]
jbertram hacst: yeah, it's a different exception [10:36]
jbertram hacst: looks like the address in the source is null...what is your client doing to trigger this? [10:38]
hacst jbertram: Probably sending an empty source. I'm playing around with the mostly undocumented Microsoft.Azure.Amqp library trying to send a message to a queue on artemis. I'm quite convinced it's me not being compliant with what I'm sending it's just that I was wondering whether NPE is a valid error handling strategy for artemis or whether this is something to report. [10:41]
jbertram hacst: IMO any NPE is a bug [10:41]
jbertram hacst: so I'd like to know how to trigger it so I can write a test for it and fix it [10:41]
jbertram hacst: feel free to open a JIRA with a full explanation - https://issues.apache.org/jira/browse/ARTEMIS [10:42]
hacst jbertram: This is the C# program https://pastebin.com/PNBXwWAY . Runs with dotnet on linux. NPE happens when the link is opened. I initialize Source with a default constructed object without passing an address as is done in https://github.com/Azure/azure-amqp/blob/master/test/TestAmqpClient/Sender.cs#L43 . [10:46]
hacst hacst's title: "azure-amqp/Sender.cs at master Azure/azure-amqp GitHub" [10:46]
jbertram gemmellr: any thoughts on this? [10:57]
gemmellr jbertram: hacst: the program is generating an NPE in the brokers consumer handling code, but it indicates its trying to create a sender...except it isnt, hence the source and target defined make no sense [11:06]
gemmellr I'd guess the role=true definition in the link settings is configuring it to be a receiver even though the sender object is being created [11:07]
jbertram gemmellr: so what the client is doing doesn't make sense? [11:07]
jbertram gemmellr: clearly the broker shouldn't be throwing an NPE...just trying to determine if this use-case is actually valid [11:08]
gemmellr jbertram: correct...its asking to consume from nowhere [11:08]
jbertram gemmellr: ok...I suspected something like that looking at the broker's code [11:08]
jbertram gemmellr: I doubt we have any tests that do that :) [11:09]
jbertram gemmellr: thx for the input [11:09]
hacst Yeah. role should've been false. [11:13]
hacst jbertram: Next NPE: https://pastebin.com/21cgXnSh . That's me setting the delivery tag field null ;) Though this time I guess it's wrapped in an proper error code. [11:18]
hacst I assume this wouldn't be something report worthy? [11:18]
jbertram hacst: well, that's coming from ProtonJ itself, not Artemis [11:19]
jbertram gemmellr: thoughts? [11:19]
jbertram hacst: FWIW, lots of people use Artemis with AMQP and there are lots of AMQP tests in the test-suite [11:21]
hacst I mean it doesn't bring down the broker or anything. [11:22]
gemmellr jbertram: I think the NPE is descriptive enough personally [11:29]
jbertram gemmellr: it's a bit ambiguous as to what the problem actually is...is it a bug in ProtonJ or did the client send something it shouldn't have? [11:30]
gemmellr jbertram: its a bug in the client [11:30]
jbertram hacst: ^ [11:31]
gemmellr jbertram: it omitted one of the most basic things needed when sending a message, the delivery tag...which hacst already noted [11:31]
wallbroken somebody use activemq 5 ? [11:38]
jbertram wallbroken: yes, I believe some people do use ActiveMQ 5 [12:05]
jbertram wallbroken: you still working on the same problem we discussed yesterday? [12:08]
tmielke [tmielke!~tmielke@cable-158-181-75-18.cust.telecolumbus.net] has joined #apache-activemq [12:12]
wallbroken yes [12:18]
jbertram wallbroken: did you send a note on the mailing list? [12:19]
wallbroken no [12:19]
jbertram wallbroken: how come? [12:19]
wallbroken so slow way [12:20]
jbertram wallbroken: what do you mean? [12:20]
wallbroken i don t want to use mailing list [12:20]
jbertram wallbroken: how come? [12:20]
wallbroken slow interacting [12:21]
jbertram wallbroken: well, the mailing list reaches a lot more folks who could potentially help you [12:22]
jbertram wallbroken: as you've noticed, there's not much activity in here [12:22]
jbertram wallbroken: just about the same 20 or so folks in here almost every day [12:23]
hacst jbertram: Trying the divert suggestion from yesterday. Doesn't works yet. Looking at a message in a classic DLQ the gui shows me a JMS_AMQP_MA__AMQ_ORIG_QUEUE property instead of an _AMQ_ORIG_QUEUE one. Is this just a display thing or do I have to divert on that? [12:38]
jbertram hacst: so if you're using AMQP it looks like it's adding the JMS_AMQP_MA_ prefix [01:43]
jbertram hacst: so not a gui thing [01:44]
hacst jbertram: Interesting. I thought JMS as a core concept wasn't a thing anymore with artemis [01:44]
jbertram hacst: it's not [01:45]
hacst https://activemq.apache.org/components/artemis/documentation/ also mentions the unprefixed props [01:45]
hacst eh... https://activemq.apache.org/components/artemis/documentation/latest/undelivered-messages.html .. iframes... [01:46]
hacst hacst's title: "Message Redelivery and Undelivered Messages ActiveMQ Artemis Documentation" [01:46]
hacst Maybe there's an additional flag/config value to set to get the pure _AMQ_ sutff? [01:48]
jbertram hacst: I'm not 100% sure why the property is named that way [01:49]
jbertram hacst: you're still using the C# AMQP client? [01:49]
hacst Yes [01:49]
hacst This header is only added after the message was DLQd (well expired actually) [01:50]
jbertram hacst: I don't work much on the AMQP stuff so it's hard to say why the header is named that way without more research [01:52]
pioto [pioto!~pioto@unaffiliated/pioto] has joined #apache-activemq [01:55]
cek [cek!uid23454@gateway/web/irccloud.com/x-bsprsaqitgdqerot] has joined #apache-activemq [03:46]
hacst Is there a way to define actual top-level names for queues instead of accessing them fully qualified in artemis? [08:23]
artnaseef [artnaseef!4167c21e@65.103.194.30] has joined #apache-activemq [08:41]