Conversation:
Notices
-
Nicely done. Luckily, the effect only goes as far as the end of the encapsulating HTML block-level element. Maaaybe the !gnusocial group would be interested to know that funny Unicode characters in a user's display name can wreak a small degree of havoc on page rendering. I don't know if they should be filtered or not, thought. There are legitimate uses for RIGHT-TO-LEFT-OVERRIDE (U+202E) in names ...aren't there?
-
@laemeur Hm, apparently we only remove those things on stuff run through common_purify() (i.e. notice HTML content), in: function common_remove_unicode_formatting($text) { // Strip Unicode text formatting/direction codes // this is pretty dangerous for visualisation of text and can be used for mischief return preg_replace('/[\\x{200b}-\\x{200f}\\x{202a}-\\x{202e}]/u', …
-
@laemeur And if it's the "userX likes this", it's something that has (or at least should've) passed through common_purify.
Perhaps it doesn't do that on some of those auto-generated things, but then it's a bug that should be fixed. But at least it does that when spreading to other servers, so potential confusion harm is minimised.
-
This is pretty. cc: @laemeur https://social.umeahackerspace.se/attachment/57725