search

superseriousbusiness / gotosocial

Fast, fun, small ActivityPub server.
https://docs.gotosocial.org
GNU Affero General Public License v3.0
3.61k stars 304 forks source link

[chore] Markdown - disable smartypants fraction #1028

Closed tsmethurst closed 1 year ago

tsmethurst commented 1 year ago

rn the markdown library we use turns 1/5 into <sup>1</sup>⁄<sub>5</sub> which is actually quite annoying

we should change this behavior in here somewhere + add tests for it: https://github.com/superseriousbusiness/gotosocial/blob/main/internal/text/markdown.go

sargant commented 1 year ago

Seems like SmartyPants treats common fractions differently from arbitrary fractions:

expected: "<p>Some fractions are 1/2 1/3 2/3 1/4 3/4 1/5 2/5 3/5 4/5</p>"
actual  : "<p>Some fractions are ½ 1/3 2/3 ¼ ¾ 1/5 2/5 3/5 4/5</p>"

Should all fractions be disabled, or the common ones that are displayable as single entities OK?

tsmethurst commented 1 year ago

Should all fractions be disabled

imo yes, if it's possible, since i think it's behavior most people won't want

sargant commented 1 year ago

Should all fractions be disabled

imo yes, if it's possible

Annoyingly, the upstream package logic has no option to disable fractions processing entirely without disabling SmartyPants processing as a whole.

Either smart fractions is on, where you get the sup/sub output behaviour, or off, and you get HTML entity processing. A second round of parsing to strip out these blacklisted entities can't differentiate between user inputted entities and smartypants generated ones.

tsmethurst commented 1 year ago

oh that's a bummer, how silly! got any ideas about what the best behavior would be then?

sargant commented 1 year ago

oh that's a bummer, how silly! got any ideas about what the best behavior would be then?

What aspects of SmartyPants parsing are actually needed? Would anything significant (e.g. ampersand escaping) be lost by disabling SmartyPants altogether?

tsmethurst commented 1 year ago

What aspects of SmartyPants parsing are actually needed

I kinda like the aesthetics of having " marks be angled nicely. And automatic em dashes. But actually none of that is necessary at all, so maybe we should just turn it off entirely :P What do you reckon?

sargant commented 1 year ago

As a backend-first project, I think disabling it altogether is sensible - and it's up to whatever frontend is rendering the data to choose to have fancy angled quotes and better dashes.

tsmethurst commented 1 year ago

Alright :) I don't think anyone will mind or even really notice if we disable it, and if that's the only way we can avoid having the silly fraction parsing, then I'm all for it!