XSF Discussion - 2019-08-03

jonas’, the http api uri in the examples of the api docs of muculumbus is wrong

or not used anymore to be exact

lovetox, which one?

jabber.network

is in the examples

both work

jabber.network will give you a redirect of course

yeah, still i dont think any http lib just autofollows redirects

im also not sure i even want to implement something like that

yeah

I’ll go and fix it soon

yeah no hurry just wanted to mention it

if only we had stable domains maintained by an organisation of sorts.

jonas’, already 2200 mucs on muclumbus

some weeks ago it was only 600

unless there was a glitch, I doubt that’s treu

there are 6.7k public MUCs in the db right now, where did you get the 2.2k from?

thats all i get when i request all

ah, it only returns MUCs with at least one occupant by default

ah yeah :)

but still weird that its always a round number

its exactly 2200

or i have a counting bug

that sounds like a pagination bug

https://xmpp.org/extensions/xep-0060.html talks about a "cache-last-item" feature, but it seems to be undefined.

What's up with that?

does anyone know a case where a error is returned with more than one error condition with namespace urn:ietf:params:xml:ns:xmpp-stanzas

the rfc says an error must contain a condition, but does not say it must be only one

the schema does not allow it at least

Multiple `<text>` is ok tho

for i18n

lovetox, I haven’t found any, when I ran xmpp-parsers on data extracted from the xeps repository.

ok then i will support one error condition and one application error condition

and multiple text elements

parsing errors is a bit weird

you have to iter the childs of error, and you dont really know what you are searching for, you only know what you dont search for

wrong window ✏

It’s one text element, but in multiple different locales.

It’s invalid for instance to have the same locale twice.

that’s still multiple elements on the XML layer

Yes.

But not conceptually.

so i looked into the xml:lang thing

i expected every server issue only one text with the lang i set on the stream

but as someone mentioned, and i saw it myself some server add also the "en" translation

Gajim just takes the first text element and expects it to be in stream lang, so as this is obviously not good i checked expat for xml:lang inheritence

seems the xml standard actually says to inherit that

but i could not find any hint that expat actually does this

BUT its stll very easy on client side

at least with my node objects that my parser creates every node has its parent assigned

so you just traverse up the parents until you find one with an xml:lang attr

so still not seeing the can of worms that i supposed to have opened

I find it actually very beneficial that entities also include "en" besides the receivers xml:lang in error stanzas

Makes it easier to read XMPP traces being posted in your forum etc

I think it would be even nice if the standard recommends doing so

lovetox: lucky you have access to the whole hierarchy of elements from where you need the language. The parser I'm working with doesn't give me the elements above.