hi jmv

[Kalahari Inkantation]

also i wrote a test script that parses out quotes but it gets kind of tricky with nested quotes

is there something that makes nested quotes harder for it to ignore than single quotes

[Daddy]

also i wrote a test script that parses out quotes but it gets kind of tricky with nested quotes

is there something that makes nested quotes harder for it to ignore than single quotes

welll i was goign to say it only cought the first [/quote] but i fixed that

Code Select

$re = '`\[quote.*\].*\[\/quote\]`s';
also looks like i fixed the snipped quote issue

[Daddy]

[[[Lord Kazm]]]

Testing that now

cuck

you should write one and unleash it on boyah

have it make a nonsensical markov'd boyahpost a day

this is a test

[Daddy]

nope nvm.

it wouldn't be able to parse out anything but the last line of that previous post.

The sentence between the two quotes would get filtered out.

[bluaki]

this is a textbook example of something you can't totally solve with regular expressions, counting is a fundamental limitation of regular languages

this is also why syntax highlighting systems that exclusively rely on regex are bad

you have to basically search for every instance of [quote] and [/quote] and count the number of open quote tags for each part of the post

pseudocode for this is something like:

Code Select

pos = 0
while match.search(post + pos, "[quote]"):
    append everything between pos and match.offset-1 to the output
    nesting = 1
    for each match of either "[quote]" or "[/quote]":
        if it's a "[quote]": nesting++
        if it's a "[/quote]": nesting--
        if nesting == 0: break
    pos = match.offset + match.length
append everything between pos and the end of the post to the output

if you want to match inside the brackets of a start-of-quote, you need to exclude the end bracket or else you'll continue matching up to the last bbcode tag before the end quote

Code Select

\[quote[^]]*\]

[Daddy]

this is a textbook example of something you can't totally solve with regular expressions, counting is a fundamental limitation of regular languages

this is also why syntax highlighting systems that exclusively rely on regex are bad

you have to basically search for every instance of [quote] and [/quote] and count the number of open quote tags for each part of the post

pseudocode for this is something like:

Code Select

pos = 0
while match.search(post + pos, "[quote]"):
    append everything between pos and match.offset-1 to the output
    nesting = 1
    for each match of either "[quote]" or "[/quote]":
        if it's a "[quote]": nesting++
        if it's a "[/quote]": nesting--
        if nesting == 0: break
    pos = match.offset + match.length
append everything between pos and the end of the post to the output

if you want to match inside the brackets of a start-of-quote, you need to exclude the end bracket or else you'll continue matching up to the last bbcode tag before the end quote

Code Select

\[quote[^]]*\]

but then you have things like

Code Select

[quote author=cuckman] that needs to accounted for, though tht's just an or statement rly

[Daddy]

excuse me but if im going to compete i need posts to reply to

[bluaki]

but then you have things like

Code Select

[quote author=cuckman] that needs to accounted for, though tht's just an or statement rly

In that pseudocode I used "[quote]" for clarity, but if you replace every instance of that with the regular expression at the end of my post, it'll handle optional attributes like author and date. I just replaced the ".*" in your expression with "[^]]*" to fix the problem of your pattern matching way too far past the first right-bracket.

[Kalahari Inkantation]

excuse me but if im going to compete i need posts to reply to

reply to blu because i think she's onto something

boyah, with its rich ten-year history of wacky posts, deserves its very own markov chain

[Daddy]

yes it does

[Kalahari Inkantation]

bump

is a boyah markov chain even really within the realm of feasibility lol

[Kalahari Inkantation]

what if we were to automatically mirror random boyah posts without formatting to some secure database (like, probably even a basic google drive account accessible only to the bot and its keeper would suffice), and had our markov bot draw from that instead of directly from boyah

that way it wouldn't have to parse quotes at all