XSF Discussion - 2021-05-19

https://matrix.org/blog/2021/05/17/the-matrix-space-beta

nice

app for what

XSF application deadline is 23rd, right? ✏

Chan Shen, your friends sound like mine :D but some of them use XMPP just for me.

good friends then

qrpnxz, yes otherwise it is dificult.

Some drama happening around Freenode: https://fuchsnet.ch/freenode-resign-letter.txt

I'm wondering if this might spur efforts around providing IRC and XMPP services under the same umbrella.

on that matter I read a couple of times 'it's time to migrate to matrix' in many freenode channels I participate

XMPP-ADDR states: ``` An entity that prepares a string for inclusion in an XMPP domainpart slot MUST ensure that the string consists only of Unicode code points that are allowed in NR-LDH labels or U-labels. [...] This implies that the string MUST NOT include A-labels as defined in [RFC5890]; ``` Ummm, how is this implied? All codepoints used in A-labels may be used in NR-LDH- or U-labels.

Heh, "implies" ... "MUST NOT"

What was A-labels again? Plain ASCII, possibly punycoded?

A-labels are "xn--PUNYCODE"

which is indeed ascii and hyphens

it's ascii and hyphens all the way down.

🤪️

So I guess it says that xmpp:user@xn--blah is forbidden

good: xmpp:nödåtgärd bad: xmpp:xn--ndtgrd-euaf5m

right

which is sad

less work for me, but sad

Why?

why less work? because i don't have to support puny ofc

why sad? i think puny is more efficient in some cases, and also it's helpful for some systems ig idk

What do you mean by "efficient"?

You still need to care about it nearer the DNS layer.

It only exists for backwards-compatibility with DNS

Sam, that sometimes puny would allow encoding more code points than a U-label.

Everything in XMPP is Unicode, UTF-8-encoded. It would be a bit weird to pass around punycoded stuff, especially as these are user-facing strings.

I'm still not entirely sure what you mean, so I may be mimsunderstanding you, but that's almost certainly not true. Punycode is encoding Unicode, UTF-8 is also encoding Unicode. You can't have more Unicode in one vs another and UTF-8 is definitely one of the most efficient and widely used ways to encode Unicode in general

all A-label are transformable to U-label, so display is not an issue, but yeah better to just cut fat and only do U-label ig

Yah, that's a better way to say it

Sam, the IDNA standard says so so i'm not gonna contradict it lol

also UTF-8 is not very efficient for high code points like say chinese. If most of your text is chinese you might wanna do UTF-32 except that compression exists so don't do that xD.

Okay, I see what you're saying, I suppose in very specific circumstances it could save some space. Thanks.

you really don't want people to enter A-labels when asked for their XMPP domain (or JID), do you?

i don't have a problem

qrpnxz, I know. I was tyring to point out that I don't believe you have a reason to be sad :)