Simple zero-or-more (*) example fails

[malcolmsparks]

I'm not sure why this works :-

((parser
   {:main :name}
   :nameStartChar #"[\w]"
   :nameChar #"[\w]*"
   :name [:nameStartChar :nameChar]
   )
 "foo")

returning

#net.cgrand.parsley.Node{:tag :net.cgrand.parsley/root, :content [#net.cgrand.parsley.Node{:tag :name, :content [#net.cgrand.parsley.Node{:tag :nameStartChar, :content ["f"]} #net.cgrand.parsley.Node{:tag :nameChar, :content ["oo"]}]}]}

whereas this produces an error :-

((parser
   {:main :name}
   :nameStartChar #"[\w]"
   :nameChar #"[\w]"
   :name [:nameStartChar :nameChar*]
   )
 "foo")

at state #{[:name_repeat+_4476 1 [:nameChar]] [:name 1 nil] [:nameChar 1 [#"[\w]"]] [:name_repeat+_4476 2 (:name_repeat+_4476 :nameChar)] [:name 2 (:name_repeat+_4476)]}
 shift/reduce conflict {#"[\w]" #{[:nameChar 1 nil]}}
#{[:name 1 nil]}
  [Thrown class java.lang.Exception]

Backtrace:
  0:      lrplus.clj:220 net.cgrand.parsley.lrplus/transitions

Is this a bug? If not, would you be able to explain why I'm getting this error?

btw. Thanks for parsley and regex, they are wonderful.

[ckirkendall]

I am also running into this same issue. It looks like in paraedit, Laurent, had an eof function to give the parse tree an ending node to match. However, it looks like you made lr+ throw a run time exception when you add you own matcher function.

[ska2342]

I would love to understand this, too. An even smaller example:

Fails: ((p/parser :thing [:name*] :name #"[a-z]" ) "foo")

Works: ((p/parser :thing [:name] :name #"[a-z]*" ) "foo")

Regards, Stefan

[cgrand]

Ok it's #1 on my todo list for parsley but since is a sizeable update...

Parsley is slightly more expressive than LR(0) but it is not LR(1) so many common cases such as repetitions can't be handled without end delimiters (the end delimiter explicitly marking the end of the separated repetition). "Historically," parsley has been built to prove this incremental parsing approach and act as a parser for CCW and LR(0) is enough for a lisp (where every repetition has an end delimiter (the top-level repetition being ended by EOF.

On Thu, Feb 14, 2013 at 3:19 PM, Stefan Kamphausen <[email protected]

wrote:

I would love to understand this, too. An even smaller example:

Fails: ((p/parser :thing [:name*] :name #"[a-z]" ) "foo")

Works: ((p/parser :thing [:name] :name #"[a-z]*" ) "foo")

Regards, Stefan

— Reply to this email directly or view it on GitHubhttps://github.com/cgrand/parsley/issues/6#issuecomment-13551119.

On Clojure http://clj-me.cgrand.net/ Clojure Programming http://clojurebook.com Training, Consulting & Contracting http://lambdanext.eu/

[ska2342]

Salut,

no need to hurry, I am just exploring ways how to parse a rather complicated format. Seems like Parsley won't fit. No problem. Thanks for taking the time to answer anyway.

Best, Stefan