XSF Discussion - 2022-12-20

is the standards@ mailing list manually moderated or did I mess up sending an email somehow?

hello everyone

nicoco: you can check the logs. see xmpp.org

emus: you mean archives?

https://xmpp.org/community/mailing-lists/

nicoco: It has been known to be slow. Double check the address tho?

can anyone help me with xpath? this is expected: $ xmllint --xpath '/xep/header/revision/version/text()' xep-0367.xml --nowarning --dtdvalid xep.dtd 0.3 0.2 0.1 0.0.1 this is not expected: $ xmllint --xpath '/xep/header/author/email/text()' xep-0367.xml --nowarning --dtdvalid xep.dtd cpetchell@atlassian.com the output I did expect was: sam@samwhited.com cpetchell@atlassian.com mwild1@gmail.com is xmllint not expanding entities? can I make it somehow?

xmllint has a flag to expand entitites IIRC

You're getting only one item when you expect three?

`xmllint --nonet --noent --loaddtd "$@"`

here is the definition so you don't have to look it up https://github.com/xsf/xeps/blob/master/xep-0367.xml#L28

grepping the man page for entities doesn't show anything that looks relevant

singpolyma, yep, the 2 not showing up are entities which is why I suspect that's the problem ✏

moparisthebest, add `--noent`

haha that did it Zash thanks much

add no entities to get some entities

unclear why "no entities" makes entities work but hey I'm not going to ask

moparisthebest, no entities in _output_ I guess

`--nonet` (no network) is also nice ;)

Guus, you'll probably want that in your script too ^

> --noent > Substitute entity values for entity references. By default, xmllint leaves entity references in place.

I'm guessing, --nonet then means it wont use HTTP PUT for output

> Guus, you'll probably want that in your script too ^ Script more or less successfully parsed all existing XEPs. I'd prefer to merge and iteratively improve when needed.

Though some innocent soul in the future may use an entity and then have to discover --noent all over again lol

Let's cross that bridge if we ever get there. It's been a couple of weeks, but I think that the script actually tests for the inclusion of an entity. That might break?

I am not very interested in spending more cycles on this, for improvements that might maybe prevent a potential future issue. Something something perfect in the way of good something.

If someone wants to have at it, by all means, go for it.

moparisthebest, `tools/xep2md.sh` remembers it

Zash, I'm gonna force-push xep-sla.md away, ok?

we don't typically have .md files and I'd rather not duplicate stuff

Okay

> [FAIL] XPATH value /xep/header/number/text() ('ProtoXEP') does not equal 'XXXX' (but should). Oh no!

Better reject it for having the forbidden value of "xxxx"

everyone's git author email matches with their '/xep/header/author/email/text()' right? I'm going to assume yes and the script will shame people where this isn't true

Zash, thanks, will fix

moparisthebest, don't assume that

why would you have to?

I have too many email addresses to keep them in sync everywhere

Aha! It's from `xep-template.xml`

jonas’: > Is the XEP in Experimental or Deferred state and the PR opener is not an author of the XEP?

Should the script allow lowercase and/or should the template be adjusted?

how else would I determine "PR opener is not an author of the XEP" ?

Zash, fixed

it's just going to be a warning for manual review anyway, but I don't think there's another way?

(allowing lowercase) ✏

moparisthebest, yes

moparisthebest, do the assumption, leave room for some override?

mailmap mebbe?

also I don't like git email addresses as validation for that purpose, because I can just set GIT_AUTHOR_EMAIL

git author vs github account vs True Identity™ ????

well we can start putting PGP keys into author and requiring signed commits

jonas’: that would be what to set to fix it yes

moparisthebest: ooh

until then though this will have to do lol

moparisthebest, please no

You can fetch key from GitHub account too if you ever wanted that way

moparisthebest, you can do it as a heuristic, but I'm fairly certain it'll break for a non-trivial number of people

yea I assume it will too

but please don't shame people when it happens :)

I just assume they'll feel shame when the script says their other email has to approve this change, I can't help it

heh

do things understand the code owner stuff linux uses?

ideally I'd like to have github handles in xep.ent really

ohhh codeowners

interesting idea

also https://git-scm.com/docs/gitmailmap is a thing that might be of use

You can fetch key from GitHub account too if you ever wanted that way

how is approver determined if it's not defined https://xmpp.org/extensions/xep-0072.html

this says council, but the xml has no <approver/>

Is it ever not Council for Standards Track ?

maybe we can just fix old xeps instead of special-casing them ?

jonas’, but how do I deal with the lack of evolve in git after that force-push?

> mercurial.error.ProgrammingError: git support can't handle octopus merges, found a commit with 11 parents :( 😵️

rekt

$ git --no-pager log --format=%aE 7eaa3962a83ebde2607c0a778cda6281981f8444..HEAD nicoco@nicoco.fr truenicoco@users.noreply.github.com why github why

perhaps trying to guess PR author with git commit author is a fools errand

Zash: ignore!

ignore? IGNORE? I can't ignore duplicated commits!

`git branch -f xep-sla zash/xep-sla` or something then

ignore

because your PR will vanish in a second

(or be marked as merged anyway)

I almost wrote a bash function to check if the version has been increased but I don't think it's doable https://paste.rs/SSR

0.0.0.2 0.0.081203 ?????

I guess "version block has been added and version is not the same" is good enough?

checking for unequal version would probably be good enough yee

`0.0.081203` lewat

moparisthebest, could you also check that the version matches `\d+\.\d+\.\d+`? Non-semver-y versions are the bane of my existence. ✏

for new ones only, and hard-error? I like it

yes & yes

thanks a lot

(btw, I didn't expect implementation discussion for the tooling to happen here; I was monitoring editor@ more closely than this place recently)

(so if I missed any questions please repeat)

ah sorry, I mainly asked because bigger audience for xmllint/xpath question then forgot to switch back over :/