#dev 2021-05-09

2021-05-09 UTC
samwilson, [snarfed] and dckc joined the channel
#
[snarfed]
from #indieweb, trying to find an authoritative place that explains that IDNA2008 (ie which Unicode chars are allowed in domain names) only applies to gTLDs and others, but not all ccTLDs
#
[snarfed]
haven’t found anything authoritative yet. seems like it would be somewhere under https://www.icann.org/resources/pages/cctlds-21-2012-02-25-en , but the docs there descend into big long legalese pretty quickly
#
[snarfed]
‘Since 2000, ICANN has also been working with managers of ccTLDs (the two-letter TLDs that have been established for countries and some territories) to document their relationship with ICANN. These relationships are more complex, because of the varying circumstances (in terms of type of organization, policies followed, economics, language, culture, legal environment, and relations with governments) of different ccTLDs and the organizations
#
[snarfed]
that operate them. An additional factor to be addressed is the role, recognized in the June 1998 U.S. Government White Paper, that national governments have in “manag[ing] or establish[ing] policy for their own ccTLDs.“’
#
[snarfed]
i’m also curious about emoji subdomains, eg https://scotthelme.co.uk/https-scotthelme-co-uk/ and https://aye.sh/blog/emoji-%F0%9F%99%8B%F0%9F%8C%8F . what’s the authority on which characters are allowed in third level and lower domain labels like those?
#
[snarfed]
browsers convert them to punycode, since DNS is just ASCII, so maybe it’s just the major browser vendors’ consensus?
#
[snarfed]
ugh so many questions
jeremycherfas, samwilson and [snarfed] joined the channel
#
[snarfed]
ok, i wrote up my research and questions for the Python requests package in https://github.com/psf/requests/issues/3687#issuecomment-835688746
#
Loqi
[snarfed] I ended up doing more research here, and I'm curious about a design decision. Was it a deliberate choice to build in just IDNA2008 and not full Punycode? Or was `idna` the only mature package you found, and it only supported IDNA2008? Or something el...
justBull, doosboox, oodani, treora, dianoetic, barnabywalters, [Rose], [tw2113_Slack_], [jacky], [jeremycherfas], [schmarty] and sparseMatrix joined the channel
#
sparseMatrix
o/
#
sparseMatrix
I finally got Beautiful Soup to give me the reMe hrefs outta my <head />
#
sparseMatrix
my first stupid human trick with Beautiful Soup
#
sparseMatrix
*relMe
#
sparseMatrix
time to find out what happens next in the ol' login workflow
leo60228, alex11, shoesNsocks, [tantek], [KevinMarks], dianoetic and sparseMatrix joined the channel
#
sparseMatrix
So, I'm working my way through the finer details of this process (the relMe login), and it seems to me the specs suggest a rather aggressive approach to finding an h-card anywhere on just about any site referenced in a relMe link anywhere in the document, as well as any explicit URI tucked into a form
#
sparseMatrix
...and that an h-card need not contain just the h-card tags but can really be any chunk of html so long as that html contains *at least* the h-card tags
#
sparseMatrix
i.e., you can paste your h-card into the head of every page on your site and any of them should work
#
Loqi
Tantek Çelik
#
sparseMatrix
Hi there Loqi
sumner joined the channel
#
aaronpk
in general, microformats parsing doesn't care too much about where it is in the HTML, it helps to decouple thinking about consuming the microformats data from consuming the HTML
#
aaronpk
as a practical example, drop a URL into one of the online test parsers and think only about consuming the JSON data returned
#
aaronpk
pretend the HTML doesn't exist, that's the nice thing about microformats
#
sparseMatrix
Yeah I was just thinking I should be using <html /> as my context vs. <head /> which where I got it working
#
sparseMatrix
then I'd judt have all the relMe links in the document
#
sparseMatrix
BUT, what I was finding especially interesting in the spec was that the hrefs should be searched recursively
dianoetic[m] joined the channel
#
sparseMatrix
for instance if a given relMe link goes to a document instead of an endpoint, the document should be searched for relMe links ad finitum and the process repeated, until auth happens or all relMe links are exhausted