Re: [Standards] Byte count of XEP-0084: User Avatar image data

2019-08-29 Thread Dave Cridland
On Wed, 28 Aug 2019 at 16:21, Paul Schaub  wrote:

> Hi Standardists!
>
> I noticed, that XEP-0084 was update in version 1.1.2 to allow image
> "height"s and "width"s up to 65536 (previously 256). While I agree, that
> this is a sensible up to date range for width and height, I think that
> the range of the size attribute (bytes) would need to be updated as well.
>
> As of today, the XML Schema for the metadata namespace states the
> datatype of bytes to be unsigned short, which is at maximum 65536. 65536
> bytes will probably quickly be exhausted for images bigger than 256x256).
>
> Interestingly, example 10 in the XEP already features an  with a
> bytes value of 78912, which exceeds the limits of unsignedShort.
>
> I'd propose to bump the datatype of bytes up to unsignedInteger, which
> should be sufficient for now.
>
>
That sounds good.


> Do you think this would require a namespace bump?
>
>
Given the example (which, IMO, holds more weight than the schema in our
world), no.


> Related PR: https://github.com/xsf/xeps/pull/812
>
> Happy Hacking!
> Paul
> ___
> Standards mailing list
> Info: https://mail.jabber.org/mailman/listinfo/standards
> Unsubscribe: standards-unsubscr...@xmpp.org
> ___
>
___
Standards mailing list
Info: https://mail.jabber.org/mailman/listinfo/standards
Unsubscribe: standards-unsubscr...@xmpp.org
___


[Standards] Byte count of XEP-0084: User Avatar image data

2019-08-28 Thread Paul Schaub
Hi Standardists!

I noticed, that XEP-0084 was update in version 1.1.2 to allow image
"height"s and "width"s up to 65536 (previously 256). While I agree, that
this is a sensible up to date range for width and height, I think that
the range of the size attribute (bytes) would need to be updated as well.

As of today, the XML Schema for the metadata namespace states the
datatype of bytes to be unsigned short, which is at maximum 65536. 65536
bytes will probably quickly be exhausted for images bigger than 256x256).

Interestingly, example 10 in the XEP already features an  with a
bytes value of 78912, which exceeds the limits of unsignedShort.

I'd propose to bump the datatype of bytes up to unsignedInteger, which
should be sufficient for now.

Do you think this would require a namespace bump?

Related PR: https://github.com/xsf/xeps/pull/812

Happy Hacking!
Paul
___
Standards mailing list
Info: https://mail.jabber.org/mailman/listinfo/standards
Unsubscribe: standards-unsubscr...@xmpp.org
___