> On 2 Dec 2016, at 16:03, forenjunkie <[email protected]> wrote:
> 
> Why would you querry the whole archive?

Because the question isn't "do I have unread messages for contact X?" It's "in 
what chats do I have unread messages and how many?" on login. I don't see a way 
to solve this using the building blocks we have without exhaustive querying of 
the archive. The case of fetching context when the user later opens a chat is 
the easy part. 

This is why I asked with a specific example, for which I don't think we 
currently have protocol. Can you show me what the protocol would look like for 
the case I mentioned earlier?

/K

> 
> if you open a chat with a contact you querry "Give me the last X messages for 
> contact A"
> 
> And if you open the chat window with contact B you do the same for contact B.
> 
> I never did write a implementation of MAM myself, but if im reading the XEP, 
> you can filter with various attributes, you dont have to step through EVERY 
> message.
> 
>> Am 02.12.2016 um 16:57 schrieb Kevin Smith:
>>> On 2 Dec 2016, at 15:49, forenjunkie <[email protected]> wrote:
>>> Its not written down somewhere that its up to the client, but it makes no 
>>> sense to put a selflimiting hard rule on a archiving XEP like MAM and 
>>> exclude certain messages per XEP rule.
>> 
>> 
>>> We have store hints, the most prominent servers respect these. And it would 
>>> make no sense to not respect it if a client explicitly wishes for a message 
>>> to be stored.
>> Oh, it’d make lots of sense. Server admins can very reasonably want to 
>> choose how their archive is populated, rather than having remote clients do 
>> so.
>> 
>>> So to make it more clear, if i want to save a message on a current prosody 
>>> or ejabbered MAM implemenation, i can do this as a client with a store hint.
>> While that’s nice for you, Prosody and ejabberd are certainly not the whole 
>> world ;)
>> 
>>> you dont have to querry the whole archive, you just querry until you get a 
>>> read marker, then you know everything that comes before that was already 
>>> read so i dont have to query it.
>> This is untrue, though. If I have two contacts, it’s quite easy for me to 
>> have unread messages for contact A that are hundreds or thousands of 
>> messages older than my messages from contact B, all of which are read. If I 
>> merely read the archive backwards until I found a read marker, I wouldn’t 
>> find the unread messages for contact A.
>> 
>> /K
>> 
>>> 
>>>> Am 02.12.2016 um 16:34 schrieb Kevin Smith:
>>>>> On 2 Dec 2016, at 15:26, forenjunkie <[email protected]> wrote:
>>>>> in a single chat conversation, it makes no sense to query unread messages.
>>>> I’m not sure that’s true at all, but putting that to the side for the 
>>>> minute...
>>>> 
>>>>> you could just query the last 10 MAM Messages, if no readmarker comes 
>>>>> with it, query another 10. such a implementation of MAM would be very 
>>>>> weird to me, but you could do it and get only unread messages with it. So 
>>>>> this already works without problem with the MAM and 0333 XEP.
>>>>> And in single chat its not possible to "miss" a message in between, which 
>>>>> you would want to query.
>>>> I think you’re missing the details of what I asked - how do you achieve 
>>>> this where there are a sufficient number of messages that just keeping 
>>>> querying until you have every message in your local archive isn’t viable?
>>>> 
>>>>> And this is not a theory, XEP-0333 are stored by the server if the client 
>>>>> wishes that, and working implementations of MAM and 0333 together you can 
>>>>> witness for example in Conversations.
>>>> It’s not up to the client whether 333 is stored by the server or not.
>>>> 
>>>>> All i see in this proposal is adding complexity to the whole process in 
>>>>> introducing another thing the server has to support.
>>>> Referring back to my previous question, can you provide an example of how 
>>>> to achieve this case with just 313 and 333 (in protocol)?
>>>> 
>>>> /K
>>>> _______________________________________________
>>>> Standards mailing list
>>>> Info: https://mail.jabber.org/mailman/listinfo/standards
>>>> Unsubscribe: [email protected]
>>>> _______________________________________________
>>> _______________________________________________
>>> Standards mailing list
>>> Info: https://mail.jabber.org/mailman/listinfo/standards
>>> Unsubscribe: [email protected]
>>> _______________________________________________
>> _______________________________________________
>> Standards mailing list
>> Info: https://mail.jabber.org/mailman/listinfo/standards
>> Unsubscribe: [email protected]
>> _______________________________________________
> 
> _______________________________________________
> Standards mailing list
> Info: https://mail.jabber.org/mailman/listinfo/standards
> Unsubscribe: [email protected]
> _______________________________________________
_______________________________________________
Standards mailing list
Info: https://mail.jabber.org/mailman/listinfo/standards
Unsubscribe: [email protected]
_______________________________________________

Reply via email to