Home freenode/#haskell: Logs Calendar

Logs: freenode/#haskell

←Prev  Next→
Page 1 .. 890 891 892 893 894 895 896 897 898 899 900 .. 5022
— 502,152 events total
2020-10-27 12:47:31 → xff0x joins (~fox@2001:1a81:5219:2300:5d08:4351:f573:c6af)
2020-10-27 12:47:35 <Raito_Bezarius> ha
2020-10-27 12:48:05 <dminuoso> (Note that the first definition is already defined in megaparsec, I just included it for reference)
2020-10-27 12:48:07 <Raito_Bezarius> well, I'll read more the docs until I understand why is it enough/trivial enough
2020-10-27 12:48:13 <Raito_Bezarius> but like
2020-10-27 12:48:20 <dminuoso> lexing and parsing are really the same thing. :)
2020-10-27 12:48:20 <Raito_Bezarius> if I want to add a semicolon to a lexeme if only it's a certain lexeme
2020-10-27 12:48:26 <Raito_Bezarius> I can just really chain lexeme & lexemeSemi ?
2020-10-27 12:48:35 <dminuoso> Raito_Bezarius: Yes, that's why you use both lexeme and lexemeSemi
2020-10-27 12:48:38 <dminuoso> For example:
2020-10-27 12:48:55 → cfricke joins (~cfricke@unaffiliated/cfricke)
2020-10-27 12:49:05 × kish quits (~oracle@unaffiliated/oracle) (Ping timeout: 240 seconds)
2020-10-27 12:49:13 <dminuoso> say you have:
2020-10-27 12:49:23 <dminuoso> lexemeSemi = p <* semi <* spc
2020-10-27 12:49:23 × carlomagno quits (~cararell@148.87.23.7) (Remote host closed the connection)
2020-10-27 12:49:26 <dminuoso> Where spc is your space consumer
2020-10-27 12:49:28 <dminuoso> Then you can define
2020-10-27 12:49:39 <dminuoso> err
2020-10-27 12:49:42 <dminuoso> lexemeSemi p = p <* semi <* spc
2020-10-27 12:49:53 <dminuoso> foo = lexemeSemi (string "foo")
2020-10-27 12:50:00 <dminuoso> which would lex `foo;` plus leading whitespace
2020-10-27 12:50:05 <dminuoso> (or trailing I guess)
2020-10-27 12:50:09 ← PlasmaStrike parts (~mattplasm@38.73.141.198) ("ERC (IRC client for Emacs 28.0.50)")
2020-10-27 12:50:10 <Raito_Bezarius> yes
2020-10-27 12:50:17 <Raito_Bezarius> now if I give foo; can I still get foo; without error?
2020-10-27 12:50:28 <dminuoso> "get foo; without error"?
2020-10-27 12:50:29 × Zetagon quits (~leo@c151-177-52-233.bredband.comhem.se) (Ping timeout: 256 seconds)
2020-10-27 12:50:31 <Raito_Bezarius> sorry
2020-10-27 12:50:36 <Raito_Bezarius> if I give as input `foo;`
2020-10-27 12:50:36 × lpy quits (~nyd@unaffiliated/elysian) (Quit: lpy)
2020-10-27 12:50:41 <Raito_Bezarius> can I still make it so it lex to `foo;`
2020-10-27 12:50:44 <geekosaur> I think Raito_Bezarius wants the opposite of what you're providing
2020-10-27 12:50:50 <Raito_Bezarius> I want both
2020-10-27 12:50:51 → PlasmaStrike joins (~mattplasm@38.73.141.198)
2020-10-27 12:50:55 × thir quits (~thir@p200300f27f0b7e004c18ab60065ea01b.dip0.t-ipconnect.de) (Ping timeout: 240 seconds)
2020-10-27 12:50:58 <Raito_Bezarius> I want semicolon to be automatically added when it's relevant
2020-10-27 12:50:59 <geekosaur> insert a virtual semicolon after some lexemes, but allow it to be explicit as well
2020-10-27 12:51:05 <dminuoso> ahh
2020-10-27 12:51:11 <Raito_Bezarius> something like lexemeSemi (string "foo") <|> lexeme (string "foo;") or something like this
2020-10-27 12:51:15 <dminuoso> Well, I'd just get rid of it in the parser
2020-10-27 12:51:19 <dminuoso> or lexer
2020-10-27 12:51:31 <dminuoso> I see
2020-10-27 12:51:36 <dminuoso> You can make the semi optional
2020-10-27 12:51:43 <dminuoso> lexemeSemi p = p <* optional semi <* spc
2020-10-27 12:52:03 <Raito_Bezarius> but then, this separator won't always appear right?
2020-10-27 12:52:12 <dminuoso> what do you mean by separator?
2020-10-27 12:52:15 <Raito_Bezarius> the ;
2020-10-27 12:52:18 <Raito_Bezarius> it will lex as foo or foo;
2020-10-27 12:52:20 <Raito_Bezarius> right?
2020-10-27 12:52:28 <dminuoso> Raito_Bezarius: What do you mean with "lex as"?
2020-10-27 12:52:32 <dminuoso> It will *consume* both
2020-10-27 12:52:57 <Raito_Bezarius> I mean, my grammar assumes that those semicolons always appear
2020-10-27 12:53:32 <Raito_Bezarius> so the parser will expect semicolons for certain rules, but without automatically adding them explicitly, the rules will fail
2020-10-27 12:53:43 <dminuoso> Im not quite sure what "adding" even means?
2020-10-27 12:53:54 <dminuoso> We're in a parser, we dont generate a string, we consume it
2020-10-27 12:53:56 <Raito_Bezarius> in this case, I'm looking at Julia
2020-10-27 12:54:09 <Raito_Bezarius> if I have, `while foo end`
2020-10-27 12:54:15 → carlomagno joins (~cararell@148.87.23.12)
2020-10-27 12:54:23 <Raito_Bezarius> it should appear as `while foo end;`
2020-10-27 12:54:29 <dminuoso> what do you mean by "appear"?
2020-10-27 12:54:56 <dminuoso> A parser usually transforms `while foo end;` into some concrete or abstract syntax tree
2020-10-27 12:55:02 <Raito_Bezarius> Yes
2020-10-27 12:55:08 <Raito_Bezarius> Hm, maybe I am creating a XY problem or something
2020-10-27 12:55:54 → Criggie1 joins (~Criggie@154.13.1.56)
2020-10-27 12:56:16 <Raito_Bezarius> dminuoso: thanks for the ideas, I'll try to go for implementation first and expose the precise problem when I encounter it
2020-10-27 12:56:19 <Raito_Bezarius> That was very helpful
2020-10-27 12:56:20 <geekosaur> if you want to parse something and then prettyprint it, just have the prettyprinter always print the semicolons while rendering the AST back into text
2020-10-27 12:56:34 <Raito_Bezarius> I'm not in the prettyprint phase though I plan to do so
2020-10-27 12:56:42 × nbloomf quits (~nbloomf@2600:1700:ad14:3020:d1e5:1a77:fd52:7b88) (Quit: My MacBook has gone to sleep. ZZZzzz…)
2020-10-27 12:56:46 <Raito_Bezarius> I was rather adapting a set of rules from a grammar into megaparsec parsing code
2020-10-27 12:57:00 <Raito_Bezarius> and the set of rules expects semicolon because the lexical analysis suggests to automatically add them
2020-10-27 12:57:02 <dminuoso> Raito_Bezarius: So this is where things can very very quickly become complicated.
2020-10-27 12:57:16 <dminuoso> Do you need to preserve the exact input?
2020-10-27 12:57:21 <Raito_Bezarius> No I do not need
2020-10-27 12:57:34 × perry69420 quits (6ee39f85@110.227.159.133) (Ping timeout: 245 seconds)
2020-10-27 12:57:39 <Raito_Bezarius> I didn't want to go for custom lexer code because I wanted to benefit from megaparsec source position stuff
2020-10-27 12:57:48 <dminuoso> Raito_Bezarius: Okay. I think there may be some confusion here of what the output of a parser really is
2020-10-27 12:57:58 <Raito_Bezarius> I want to output an AST right now
2020-10-27 12:58:12 <Raito_Bezarius> I have an AST in terms of data
2020-10-27 12:58:13 <dminuoso> Raito_Bezarius: Do you represent semicolons in that AST?
2020-10-27 12:58:19 <Raito_Bezarius> No, so I can just discard them
2020-10-27 12:58:32 <Raito_Bezarius> But I was afraid of stuff like ambiguous things in the grammar
2020-10-27 13:00:59 → acarrico joins (~acarrico@dhcp-68-142-39-249.greenmountainaccess.net)
2020-10-27 13:01:31 → djellemah joins (~djellemah@2601:5c2:100:96c:e008:b638:39fe:6a54)
2020-10-27 13:03:45 → britva joins (~britva@31-10-157-156.cgn.dynamic.upc.ch)
2020-10-27 13:04:27 → hyperisco joins (~hyperisco@d192-186-117-226.static.comm.cgocable.net)
2020-10-27 13:05:09 <Raito_Bezarius> maybe stupid question, I'm trying to have "some digitChar" but I cannot because "Illegal equational constraint Token s ~ Char", should I just enable TypeFamilies or GADTs?
2020-10-27 13:05:16 <Raito_Bezarius> or am I doing something wrong?
2020-10-27 13:05:22 <Raito_Bezarius> integer = some digitChar <?> "an integer"
2020-10-27 13:05:33 → nbloomf joins (~nbloomf@2600:1700:ad14:3020:d1e5:1a77:fd52:7b88)
2020-10-27 13:06:08 × plutoniix quits (~q@175.176.222.7) (Quit: Leaving)
2020-10-27 13:07:09 <dminuoso> Raito_Bezarius: What's the type signature of integer?
2020-10-27 13:07:33 <dminuoso> Or.. that error mmm
2020-10-27 13:07:36 → ClaudiusMaximus joins (~claude@198.123.199.146.dyn.plus.net)
2020-10-27 13:07:36 × ClaudiusMaximus quits (~claude@198.123.199.146.dyn.plus.net) (Changing host)
2020-10-27 13:07:36 → ClaudiusMaximus joins (~claude@unaffiliated/claudiusmaximus)
2020-10-27 13:07:38 <dminuoso> Which ghc are you using?
2020-10-27 13:07:52 → vacm joins (~vacwm@70.23.92.191)
2020-10-27 13:07:53 <Raito_Bezarius> I want something like (MonadParsec m, …) => m [Char]
2020-10-27 13:07:56 <Raito_Bezarius> GHC8 dminuoso
←Prev  Next→
Page 1 .. 890 891 892 893 894 895 896 897 898 899 900 .. 5022

All times are in UTC.