[Standards] Delayed Delivery for CSI and possibly SM

Daniel Gultsch daniel at gultsch.de
Wed May 31 09:46:15 UTC 2017


2017-05-31 11:35 GMT+02:00 Dave Cridland <dave at cridland.net>:
> On 31 May 2017 at 10:25, Daniel Gultsch <daniel at gultsch.de> wrote:
>> 2017-05-30 23:48 GMT+02:00 Dave Cridland <dave at cridland.net>:
>>> On 30 May 2017 at 22:33, Daniel Gultsch <daniel at gultsch.de> wrote:
>>>> That subelement could be in the namespace of the XEP that adds it.
>>>> I'm fairly certain that this shouldn't cause problems for any half way
>>>> decent implementation.
>>>
>>> It's the less than halfway decent implementations I'm more concerned about.
>>>
>>> It might be worth giving it a go to see what happens.
>>
>> It's worth pointing out that this will primarily affect
>> implementations that implement SM or CSI. I think it's fair to assume
>> that all implementations that implement those are above the 'halfway
>> decent' threshold.
>>
>> I'm willing to create PRs for the following three changes if we hereby
>> agree that we want to handle it this way.
>>
>> 1) Add wording to XEP-0203 that the delay element MAY contain either
>> character data that provide a natural language description OR an
>> element of a different namespace describing the reason for the delay.
>
> Only the latter, please. I could be persuaded into having a <text/>
> element or some such if people *really* want that, but I wouldn't want
> text-or-element anywhere if I can avoid it.

Well that's kinda the problem here. The delayed delivery XEP already
has the MAY add a natural language description wording.
And implementations seem to use that as well.

>
>> 2) Add a section to XEP-0353 that the server SHOULD add a delay
>> element to delayed message and presence stanzas that MUST contain a
>> <csi/> element
>> 3) Add a section to XEP-0198 that the server SHOULD add a delay
>> element to delayed messages and presences which MUST contain a <sm/>
>> element
>>
>> (I think SHOULD wording in doesn't require a namespace bump. But I'm
>> willing to go with MAY)
>
> I'd go MAY/SHOULD to avoid a namespace bump.
>
> Is there any reason a client shouldn't add these delay stamps too?
> Seems sensible particularly for messages.

Well some clients already do. For example when resending messages
after a SM timeout.
But that's a different story isn't it? Or would you like me to add
wording to XEP-0198 that clients MAY add stamps when the resend
something?

cheers
Daniel


More information about the Standards mailing list