Markdown/html parsing issue in clues

[frozenpandaman]

Not sure if the NYT has changed their format or something, where the underlying representation of puzzles is written using Markdown styling rather than HTML, but here is a clue on the NYT site:

And how it shows up on downforacross:

[bdenney]

@frozenpandaman which date was this clue from?

[frozenpandaman]

@bdenney The date I posted it, December 9, 2021 (I just googled it to double-check 😉)

[bdenney]

This is odd, but it appears to be fixed on latest main?

I do see a few HTML encoding issues on other clues though, so perhaps I'll try and address those 😄

[stevenhao]

thanks @bdenney !

sometimes there are different versions of the same puzzle uploaded to the repo -- usually because the puz files were prepared by different software. i think it's pretty safe to detect html escape sequences like & and &lquot; and such.

autodetecting and parsing markdown might be a bit more tricrky; i could see it backfiring in other cases (i.e. where markdown syntax was not intended but a markdown parser picks it up inadvertently). worth a shot though

[frozenpandaman]

Cool, thanks for the info! And glad it's fixed (huh! sorry, don't know the exact puzzle/game URL)

I'll close this for now then but @bdenney, of course do feel free to open a new issue for other parsing stuff. :D

[stevenhao]

i think bdenney's proposal to fix the html parsing is good; i updated this issue to include html in scope too

[frozenpandaman]

also probably related: #243

cc @bdenney