On Tue, 5 Apr 2011, Eugeniy Meshcheryakov wrote:
> Hello,
>
> I'm trying to check if new version of coccinelle (0.2.5) works correctly,
> and got strange results with python_regexp.cocci example:
>
> % spatch -sp_file /tmp/python_regexp.cocci python_regexp.c
> init_defs_builtins: /usr/share/coccinelle/standard.h
> File "/tmp/python_regexp.cocci", line 3, column 15, charpos = 45
> around = '', whole content = m = re.compile('_new$')
> Fatal error: exception Failure("lexing: empty token")
>
> Is it expected result, or is there something wrong with how I compile
> the package? Other examples that use python seem to work fine (after
> correcting syntax errors).
A patch that fixes the problem is below.
julia
diff -u -p a/parsing_cocci/lexer_script.mll b/parsing_cocci/lexer_script.mll
--- a/parsing_cocci/lexer_script.mll 2011-04-02 10:39:29.000000000 +0200
+++ b/parsing_cocci/lexer_script.mll 2011-04-05 22:03:04.000000000 +0200
@@ -46,18 +46,15 @@ rule token = parse
| "/" { TScriptData (tok lexbuf) }
| "//" [^ '\n']* { token lexbuf } (* skip SmPL comments *)
| '"' { TScriptData (Printf.sprintf "\"%s\"" (string lexbuf)) }
- | "'" { TScriptData (Printf.sprintf "'%s'" (char lexbuf)) }
+ | "'" { TScriptData (Printf.sprintf "'%s'" (cstring lexbuf)) }
| eof { EOF }
| _ { raise (Lexical ("unrecognised symbol, in token rule:"^tok lexbuf)) }
(* These are C strings. Perhaps they require some adjustment. *)
and string = parse
| '"' { "" }
- | (_ as x) { Common.string_of_char x ^ string lexbuf }
- | ("\\" _) as x { x ^ string lexbuf }
+ | (_ as x) { (String.make 1 x) ^ string lexbuf }
-and char = parse
- | (_ as x) "'" { String.make 1 x }
- | (("\\" (oct | oct oct | oct oct oct)) as x "'") { x }
- | (("\\x" (hex | hex hex)) as x "'") { x }
- | (("\\" _ ) as x "'") { x }
+and cstring = parse
+ | "'" { "" }
+ | (_ as x) { (String.make 1 x) ^ cstring lexbuf }
_______________________________________________
Cocci mailing list
[email protected]
http://lists.diku.dk/mailman/listinfo/cocci
(Web access from inside DIKUs LAN only)
