link parsing
The automatic link parser is rather overzealous. It currently interprets any sequence of alphanumeric characters, a dot/period, and more alphanumeric characters (with no spaces in between) as a link. As a result, any time someone forgets to put a space between two sentences, or deliberately types something like "notbad.jpg", it gets turned into a (broken) link. Furthermore, if the streamer has link filtering turned on, such occurrences will result in the entire message being blocked when there's no actual reason for it to be.
The link parser needs to be smarter. Maybe make it require either two dots/periods with characters before, between, and after; or <characters>://<characters>.<characters> before it'll turn something into a link?
Silvia_Dragoness
shared this idea
-
denvelsamit
commented
The automatic link parser often misinterprets text as links, especially when alphanumeric characters are followed by a dot and more characters, like "notbad.jpg." This can cause broken links and message blocking if link filtering is enabled, even when no actual link is present. To resolve this, the parser should be refined to recognize genuine links by requiring patterns like <characters>://<characters>.<characters> or two dots with characters before, between, and after. For example, a legitimate link like Dairy Queen Menu Prices in Canada (https://thecanadianmenuprices.com/dairy-queen-menu-prices-in-canada/) should be correctly identified, while "notbad.jpg" or sentences without spaces should not be misinterpreted as links.
-
ATRO_1337
commented
In Chrome at least, anything I type with.a.period in it, becomes a URL (clickable) link. This seems like it might be not the safest thing - but I am not sure. Just wanted to see what others' opinions are.