Re: Token value in custom error reporting

[Hans Åberg]

> On 19 Jun 2020, at 07:16, Akim Demaille <akim@lrde.epita.fr> wrote:
> 
>> Le 18 juin 2020 à 20:46, Hans Åberg <haberg-1@telia.com> a écrit :
>> 
>> Otherwise, in your link above you suggest not using the semantic value in 
>> error messages, but when using locations, it contains the token 
>> delimitations. So there seems to be no advantage letting the lexer 
>> generating the error.
> 
> It is still useful for the scanner to emit the error message, because the 
> parser has no idea what is wrong.  Granted, it knows _where_ it's wrong, but 
> not _why_:
> - invalid character?
> - not-closed string?
> - invalid escape sequence?
> - out-of-range literal number?
> - etc.
> 
> The scanner faces the error, *it* should say what's wrong.

Actually, I throw exceptions from the lexer, too. Like:
   throw mli::my_parser::syntax_error(yylloc, "Directive syntax error.”);

In addition, I have warnings, which Bison does not support. A snippet that use 
both errors and warnings:
        if (unused_variable != true)
          mli::my_parser::error(@x, "warning: " + ds);
        else
          throw syntax_error(@x, ds);

The error function checks if the string starts with “warning” and then does not 
write the “error: “ part.