#dev 2023-07-16

2023-07-16 UTC
cerealp0rt, tei_ and [manton] joined the channel
#
[manton] [snarfed] Finally tracked down that Webmention problem… The Microformats parsing library is actually getting tripped up by your “http://public_at_ryanb.org” mailto URL when it’s looking at links on the page.
#
[manton] I’m deploying a dumb quick fix and then will investigate in more detail.
#
[snarfed] argh sorry!
tei_, vladimyr, [tantek], [manton]1, [KevinMarks]1, [fluffy]1, UKn0Me, superkuh, gRegor, [jgarber]1, tei_1, sebbu2, geoffo and [capjamesg] joined the channel
#
[fluffy] I seem to recall a lot of weird failures in various mf2 parsers and webmention receivers that are trying to hyper-vigilantly validate data that’s got some antispam/privacy-protecting measures in it. IMO people shouldn’t be trying to validate the u-email property because there are many good reasons to have a non-validating address there.
#
[fluffy] In my u-email I URL-encode the @ and . which has the nice features of allowing browsers to still decode the address while tripping up most/all spam harvesters. But it also causes problems with some webmention receivers.
tei_1, geoffo, tei_, bret_ and gRegor joined the channel
#
gRegor [manton], which parser was that? Guessing since it was a rel-me value it was part of the finding the authorship, though it should be leaving that with mailto: scheme
#
gRegor Actually I'm not sure what was going on. Unless the markup changed, the h-entry has an author property. Only see the mailto in the rel-mes