amqp — archive

[Alan Conway]

Here's a draft that Clemens and I have been playing ping-pong with for everybody's information. We're both pretty much agreed at this point, but need to do some clean-up before making it a WD. ---------- Forwarded message --------- From: Alan Conway < aconway@redhat.com > Date: Thu, Jan 24, 2019 at 11:48 AM Subject: Re: cmts-aconway-amqp-addressing-v1.0-wd07.docx To: Clemens Vasters < clemensv@microsoft.com > Unbelievable. Here you go. On Thu, Jan 24, 2019 at 6:59 AM Clemens Vasters < clemensv@microsoft.com > wrote: Reattach your edits please? From: Alan Conway < aconway@redhat.com > Sent: Tuesday, January 22, 2019 2:13 AM To: Clemens Vasters < clemensv@microsoft.com > Subject: Re: cmts-aconway-amqp-addressing-v1.0-wd07.docx HI Clemens, Mostly agreed, I took out the uncontroversial comments and change markers. Some fixes to address your comments are left as changes so you can review. Main change is to the URI section: I propose allowing URI references (not just URIs) as AMQP addresses. This helps backwards compat and migration - most strings that don't look like a URIs will remain a valid AMQP address with a sensible interpretation as relative URI refs under the spec. New apps can use explicit URI syntax in the spec without conflict. Some existing apps will have minor trouble e.g. "topic:foo" is technically not a valid URI reference. However such patterns should be obvious enough to allow existing impls to support them as special cases while still adopting the spec. URI refs also cover the transport-neutral address case without needing a new scope:// scheme. I redid the URI section to explain this, take a look. On Mon, Jan 21, 2019 at 7:33 AM Clemens Vasters < clemensv@microsoft.com > wrote: I mostly like your rewrite of the Authority section. I ve commented on some of your comments. Attachment: aconway2-amqp-addressing-v1.0-wd07.docx Description: application/vnd.openxmlformats-officedocument.wordprocessingml.document