As mentioned in the other email I was not saying this change was a risk (I was actually the one who noticed this particular problem whilst discussing a different issue with Andrew earlier). I was just saying that in general I wouldn't be messing with accept modes at this point in the release process normally, however taking this corrective action to fix the inadvertent changes already introduced is a no-brainer.
Robbie. -----Original Message----- From: Rajith Attapattu [mailto:[email protected]] Sent: 08 March 2011 20:42 To: [email protected] Cc: Justin Ross Subject: Re: Request for the following commits to be included in 0.10 I disagree with item for #4 being a risk. Andrew's commit was to restore a feature that I dropped when I committed rev http://svn.apache.org/viewvc?rev=1076800&view=rev Chances are that there might be apps out there that rely on NO_ACKNOWLEDGE mode (even though it's not a standard JMS option). All what Andrew has done is to use that ack mode to set a consumer level default and if somebody has a destination level setting that will be overridden. Rajith On Tue, Mar 8, 2011 at 3:33 PM, Justin Ross <[email protected]> wrote: > Hi, Robbie. I appreciate your caution regarding item 4, and I'm > prepared to revert that change if it's not truly low risk. > > Anyone else have an opinion on this item? > > > On Tue, 8 Mar 2011, Robbie Gemmell wrote: > > 4. This specific change looks good to restore the prior behaviour, > though >> in >> general I wouldn't consider work around change of accept mode to be >> very low risk. I would suggest there should have been tests added to >> verify the behaviour of the original change that caused the issue >> though (same goes for all really). >> >> Robbie >> >> -----Original Message----- >> From: Justin Ross [mailto:[email protected]] >> Sent: 08 March 2011 18:08 >> To: [email protected] >> Subject: Re: Request for the following commits to be included in 0.10 >> >> Hi, Rajith. I approved item 4. >> >> Robbie, would you indicate briefly up or down for 0.10 in the >> comments of >> qpid-3109 and -2732? >> >> Justin >> >> On Tue, 8 Mar 2011, Rajith Attapattu wrote: >> >> Hi Justin, >>> >>> I'd like the following commits to be included in the 0.10 release branch. >>> >>> 1. http://svn.apache.org/viewvc?rev=1078961&view=rev - QPID-3109 - >>> Fairly low risk. >>> This is a simple bug fix. >>> >>> 2. http://svn.apache.org/viewvc?rev=1078971&view=rev - QPID-2732 - >>> Low risk. >>> This is an addition to the initial commit made under this JIRA. >>> It adds default reliability modes and throws an exception for an >>> unsupported combination. >>> >>> 3. http://svn.apache.org/viewvc?rev=1079408&view=rev - QPID-2732 - >>> Low >>> >> Risk >> >>> The reliability mode is used to mark a transfer unreliable and >>> that flag is used to determine if a transfer should be stored for >>> replay or >>> >> not. >> >>> Again a fairly straightforward and low risk commit. However I have >>> asked Rafi to have a quick look at it as well. >>> >>> 4. http://svn.apache.org/viewvc?rev=1079402&view=rev - QPID-3127 - >>> Very low risk. >>> Fixes a bug in the initial commit for QPID-2732. >>> It just sets accept mode to NONE for NO_ACKNOWLEDGE. >>> >>> Regards, >>> >>> Rajith >>> >>> >> --------------------------------------------------------------------- >> Apache Qpid - AMQP Messaging Implementation >> Project: http://qpid.apache.org >> Use/Interact: mailto:[email protected] >> >> >> >> --------------------------------------------------------------------- >> Apache Qpid - AMQP Messaging Implementation >> Project: http://qpid.apache.org >> Use/Interact: mailto:[email protected] >> >> >> > --------------------------------------------------------------------- > Apache Qpid - AMQP Messaging Implementation > Project: http://qpid.apache.org > Use/Interact: mailto:[email protected] > > --------------------------------------------------------------------- Apache Qpid - AMQP Messaging Implementation Project: http://qpid.apache.org Use/Interact: mailto:[email protected]
