Actually with the current iq scheme, that's one thing that should be
clarified:

Can the invisible iq be sent before initial presence?

It seems that would need to be supported in order to not leak your presence
when you first log on, otherwise a contact may see you come online
momentarily and then go offline.


-K


On Wed, Jul 16, 2014 at 9:47 AM, Stefan Karlsson <s...@synergysky.com> wrote:

> Silly question:
>
> Why not just have invisible as a presence mode, and remove the silly
> enforced empty <presence/> at initialization?
>
> /stefan
>
> Peter Saint-Andre skrev 16/07/14 17:11:
>
>  On 6/19/14, 9:30 PM, Lance Stout wrote:
>>
>>> 1. Is this specification needed to fill gaps in the XMPP protocol stack
>>>> or to clarify an existing protocol?
>>>>
>>>
>>> This is a feature that has received a lot of end-user requests, and we
>>> have no other good way to do it, so yes.
>>>
>>> If anyone is going to ever implement this feature, let's have a thought
>>> out approach for them instead of horrible hacks.
>>>
>>>
>>>  2. Does the specification solve the problem stated in the introduction
>>>> and requirements?
>>>>
>>>
>>> Yes, it does.
>>>
>>>
>>>  3. Do you plan to implement this specification in your code? If not,
>>>> why not?
>>>>
>>>
>>> I've implemented this twice already on the client side - in SleekXMPP
>>> and stanza.io.
>>>
>>> However, I'm not aware of any server-side implementation to use those
>>> with.
>>>
>>
>> I've been talking about adding it to Prosody. :-)
>>
>>  4. Do you have any security concerns related to this specification?
>>>>
>>>
>>> As mentioned in the XEP, it's still very easy to expose the fact that
>>> you're online, but any method of accomplishing presence invisibility will
>>> have that issue.
>>>
>>
>> Yes, and this is one reason I don't like the entire concept of
>> invisibility.
>>
>> However, as noted, if we're going to do invisibility (and users want it
>> so clients will be written to support it), then let's at least have a
>> reasonable protocol for it.
>>
>>  One thing I notice not mentioned in the XEP is client handling of
>>> bookmarks set to auto join.
>>>
>>
>> Good point. I'll add a note about that.
>>
>> Peter
>>
>>
>>
>

Reply via email to