Re: Java skeleton bug fixes, new features, and documentation

[Paolo Bonzini]

Di-an JAN wrote:
> This patch for the Java skeleton fixes several bugs, addes some new
> features and adds tables of declarations and action features to the
> manual.  I have no copyright assignment so we'll have to do something
> about that if you want to accept this patch.

I think this is the only reason to delay this after 2.4.  I'll send the
copyright assignment offlist right away.

Thanks -- this basically gives the Java skeleton the real-world testing
that was lacking.

These hunks:

> @@ -205,7 +210,7 @@ b4_lexer_if([[
>  
>    ]b4_locations_if([
>    protected final void yyerror (String s) {
> -    yylexer.yyerror ((Location)null, s);
> +    yylexer.yyerror ((]b4_location_type[)null, s);
>    }
>    protected final void yyerror (]b4_position_type[ loc, String s) {
>      yylexer.yyerror (new ]b4_location_type[ (loc), s);
> @@ -414,7 +419,7 @@ b4_lexer_if([[
>      yycdebug (s + (yytype < yyntokens_ ? " token " : " nterm ")
>             + yytname_[yytype] + " ("]b4_locations_if([
>             + yylocationp + ": "])[
> -           + (yyvaluep == null ? "(null)" : yyvaluep) + ")");
> +           + (yyvaluep == null ? "(null)" : yyvaluep.toString()) + ")");
>    }
>  
>    /**

should be committed right now, as they are really serious bugfixes (much
more so than allowing just 1 lex-param); I will do it.  They are small
enough and  do not need a copyright assignment.

Also, I didn't understand this:

> @@ -99,7 +105,7 @@ b4_token_enums(b4_tokens)
>        return new ]b4_location_type[ (rhs.locationAt (1).begin, 
> rhs.locationAt (n).end);
>      else
>        return new ]b4_location_type[ (rhs.locationAt (0).end);
> -  }]])
> +  }]])[
>  
>    /**
>     * Communication interface between the scanner and the Bison-generated

Where is the corresponding open bracket removed?  Or was there really
none?  This is also a candidate for applying now.

Paolo