|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
 |
|
MQSeries Messageid/Correlid genrator change in 5.3 |
« View previous topic :: View next topic » |
Author |
Message
|
markymarku |
Posted: Tue Mar 02, 2004 8:10 pm Post subject: MQSeries Messageid/Correlid genrator change in 5.3 |
|
|
Newbie
Joined: 02 Mar 2004 Posts: 2
|
Hi All.
Has anyone encountered an issue where old applications use string manipulation functions to work with the message id and an upgrade to Websphere MQ 5.3 makes those old applications break as they SHOULD have been treating those fields as a BYTE24 instead of string.
I have found that WEBSPhere MQ 5.3 always places 00's towards the end of a 5.3 messageid. If an old program (written before my time) uses pointers or string functions to copy the messageid to correlation id on the way back, half the message id is chopped off,the server application thinks it didnt get a response and there goes the programs.
So working on a site with a large number of programs written by external contractors, i have to find the code and fix it.
I spoke to IBM and they admitted a change occurred in 5.3 where it will always put those 00's in the messageid, which is new but undocumented. it only did it once in a while in 5.2, 5.1.
So does anyone think IBM should put out a patch to make it act like it used to, or should i fix apps as they break?
Any recommendations.... |
|
Back to top |
|
 |
bower5932 |
Posted: Wed Mar 03, 2004 6:18 am Post subject: |
|
|
 Jedi Knight
Joined: 27 Aug 2001 Posts: 3023 Location: Dallas, TX, USA
|
Quote: |
So does anyone think IBM should put out a patch to make it act like it used to, or should i fix apps as they break?
|
If you're asking for opinions, I'd say that IBM doesn't need to put out a patch. It has always been documented that the fields are bytes and not strings. You just happened to be lucky (or unlucky) depending on your point of view. |
|
Back to top |
|
 |
JasonE |
Posted: Wed Mar 03, 2004 9:44 am Post subject: |
|
|
Grand Master
Joined: 03 Nov 2003 Posts: 1220 Location: Hursley
|
Given the fact I think it was me who answered your original question (PMR), my opinion probably wont help as its very biased.
I would however emphasize the 0x00's could turn up in the middle of the generated msgid/correlids at 5.2 and earlier - I have answered and debugged intermittent problems where people have done the same as your apps, ie rely on the str* C type functions, and they are a real pain to debug. Note this wasnt why the change was made (I dont know *why* it was done, something to do with consistency), just an additional comment. |
|
Back to top |
|
 |
markymarku |
Posted: Thu Mar 04, 2004 8:53 pm Post subject: |
|
|
Newbie
Joined: 02 Mar 2004 Posts: 2
|
Thanks for your response.
My problem is that the legacy applications use pointers to pass data to and from MQ API, and we know what pointers do with nulls, chop chop.
Anyhow ill just accept that the original programs were incorrectly written and will have to fix up the problems as they show up ( third party developers did alot of work and didnt leave much source for me).
My suprise is that this change wasnt documented and IBM are REALLY good with their documents, very detailed.
Oh well, more work for me. |
|
Back to top |
|
 |
|
|
 |
|
Page 1 of 1 |
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
|
|
|