XEP-0353/JMI: Fix specification and xml schema inconsistencies

[lnjX]

Currently the schema says some elements MUST contain a '<reason/>' from urn:xmpp:jingle-message:0 and the examples and specification text said that some elements SHOULD contain a <reason/> element from urn:xmpp:jingle:1.

This does:

• SHOULD -> MUST contain reason
• Use reason from urn:xmpp:jingle:1 in both XML schema and examples

• XEP-0353/JMI: Remove redefinition of jingle reason element in schema
• XEP-0353/JMI: Make usage of reason element a MUST
• XEP-0353/JMI: Add version block for 0.7.0

[lnjX]

Ping @tmolitor-stud-tu :)

[lnjX]

Ping @fippo @stpeter @tmolitor-stud-tu

and @iNPUTmice :)

[tmolitor-stud-tu]

Unfortunately, afaik, the SHOULD --> MUST requires a namespace bump and we should avoid this.
Doing it the other way round (MUST-->SHOULD) would be okay, though.

[lnjX]

@tmolitor-stud-tu I see. :/ However, the XML schema already had minOccurs=1 maxOccurs=1 before, so the schema already had a MUST before. This is inconsistent, but both conditions, those from the text and those from the XML schema need to be met, or am I wrong?

If so, then a JMI element without the reason element was allowed by the text but not by the XML schema, so it was not valid.
If we also change the text to MUST, then this does not change the behavior, an element without the reason element is still invalid.

However, this is maybe a bit pedantic and I can totally understand if you say this leads to confusion and making the MUST from the schema a SHOULD would be better than the current situation. :)

[stpeter]

What do the existing implementations generate and expect?

[lnjX]

xmpp-parses/Rust, slixmpp, movim, dino all ignore the reason element and also don't generate it. I didn't have a look into the code of the other implementations, but it seems like making it optional is the way to go then.

I'll adjust the PR.

[lnjX]

Sorry that this took so long, did the adjustments now. :)