In my previous email I perhaps should have also reported a second type of error: example.hs:26:21: Couldn't match expected type ‘Text.Parsec.Token.GenTokenParser s0 u0 m0’ with actual type ‘[Char]’ In the first argument of ‘prefix’, namely ‘"-"’ In the expression: prefix "-" negate In the expression: [prefix "-" negate, prefix "+" id] On Fri, Feb 26, 2016 at 10:05 PM, Jeffrey Brown wrote:

Thanks, Will. I had tried that, and got a lot of errors like this:

example.hs:20:26: Couldn't match type ‘Char’ with ‘()’ Expected type: [()] Actual type: [Char] In the first argument of ‘symbol’, namely ‘"("’ In the first argument of ‘between’, namely ‘(symbol "(")’ In the expression: between (symbol "(") (symbol ")")

On Fri, Feb 26, 2016 at 8:20 PM, William Yager wrote:

...

https://hackage.haskell.org/package/parsec-3.1.9/docs/Text-Parsec-Token.html... ?

--Will

On Fri, Feb 26, 2016 at 10:08 PM, Jeffrey Brown wrote:

...

That still won't compile, because GHC does not know what reservedOp means. Does reservedOp refer to something that no longer exists, or have I just not found it?

-- Jeffrey Benjamin Brown

-- Jeffrey Benjamin Brown

Megaparsec does away with Parsec's LanguageDef machinery (which provides reservedOp). This is a double edge sword - as Megaparsec's author notes Parsec lexers (i.e LanguageDef based lexers) are inflexible especially if you need whitespace sensitive parsing; but they are very handy for the simple case of whitespace insensitive parsing. For the expression parser, Megaparsec's documentation is wrong[*] and probably it should use symbol rather than reservedOp. Note that symbol is slightly different in Megaparsec as it's a plain combinator (rather than one instantiated from a first class module as in Parsec) so it takes two args rather than one. [*] Well, likely wrong - I haven't got round to using Megaparsec yet. On 27 February 2016 at 17:04, Özgür Akgün wrote:

Hi.

On 27 February 2016 at 04:20, William Yager wrote:

...

https://hackage.haskell.org/package/parsec-3.1.9/docs/Text-Parsec-Token.html... ?

If I understand it correctly, Jeffrey is asking about using Megaparsec, and this link is to the parsec combinator with the same name.

I am very surprised to find out that Megaparsec does not provide[*] the same combinator. Especially since they keep the same example!

[*] At least it is not listed here: http://hackage.haskell.org/package/megaparsec-4.4.0/docs/doc-index-All.html

-- Özgür Akgün

_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe

I'm trying this again! The comment for Megaparsec.makeExprParser refers to "parens". When I grep for "parens" in Megaparsec I get only two occurences, both of them in comments: ./Lexer.hs:-- > parens = between (symbol "(") (symbol ")") ./Expr.hs:-- > term = parens expr <|> integer "term" If I try defining parens as in the comment from Lexer.hs, I get these errors: <interactive>:37:33: Couldn't match type ‘Char’ with ‘()’ Expected type: [()] Actual type: [Char] In the first argument of ‘symbol’, namely ‘"("’ In the first argument of ‘between’, namely ‘(symbol "(")’ In the expression: between (symbol "(") (symbol ")") <interactive>:37:46: Couldn't match type ‘Char’ with ‘()’ Expected type: [()] Actual type: [Char] In the first argument of ‘symbol’, namely ‘")"’ In the second argument of ‘between’, namely ‘(symbol ")")’ In the expression: between (symbol "(") (symbol ")") [1] https://hackage.haskell.org/package/megaparsec-5.0.1/docs/Text-Megaparsec-Ex... On Mon, Feb 29, 2016 at 2:05 AM, Özgür Akgün wrote:

On 29 February 2016 at 08:22, Stephen Tetley wrote:

...

For the expression parser, Megaparsec's documentation is wrong[*] and probably it should use symbol rather than reservedOp. Note that symbol is slightly different in Megaparsec as it's a plain combinator (rather than one instantiated from a first class module as in Parsec) so it takes two args rather than one.

[*] Well, likely wrong - I haven't got round to using Megaparsec yet.

You are right: https://github.com/mrkkrp/megaparsec/commit/ 750adb7c70392c3195eda12d816f4a1a2305321e

-- Özgür Akgün

_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe

-- Jeff Brown | Jeffrey Benjamin Brown Website https://msu.edu/~brown202/ Facebook https://www.facebook.com/mejeff.younotjeff LinkedIn https://www.linkedin.com/in/jeffreybenjaminbrown (InMail is unreliable) Github https://github.com/jeffreybenjaminbrown