Tim Van Holder wrote:
> Guillaume Rousse wrote:
>> Which is really not meaningful for user :/ I'd prefer to at least
>> display current line, and avoid refering to internal grammar symbols.
>
> Part 2 - not referring to internal grammar symbols: use string aliases
> for your tokens, like so:
>
>
Guillaume Rousse wrote:
> Which is really not meaningful for user :/ I'd prefer to at least
> display current line, and avoid refering to internal grammar symbols.
Part 2 - not referring to internal grammar symbols: use string aliases
for your tokens, like so:
%token MY_IDENTIFIER "identifier"
%
Satya wrote:
> hi,
> As I understand, you want to refer to the user's exact input symbol in
> the error message rather than the grammar symbol associated with it..
> ?
Rather the whole input line than just an isolated part of it.
> Well, flex provides a global variable 'yytext' that contains the v
hi,
As I understand, you want to refer to the user's exact input symbol in
the error message rather than the grammar symbol associated with it..
?
Well, flex provides a global variable 'yytext' that contains the value
of the symbol. Suppose you have a definition like this in your lexer:
[0-9]+
Hello.
I'm trying for two days now to produce better output messages in case of
parsing failure. I'm using flex + bison for a standard (non-reentrant)
parser.
I've had to redefine YY_INPUT in scanner.l, as I'm parsing small,
line-based messages, and default version was blocking:
#define YY_INPUT
