[antlr-interest] Grammar composition in v3, multiple packages, no joy

Jim Idle jimi at temporal-wave.com
Wed Feb 10 11:30:33 PST 2010


> -----Original Message-----
> From: antlr-interest-bounces at antlr.org [mailto:antlr-interest-
> bounces at antlr.org] On Behalf Of Christian Nentwich
> Sent: Wednesday, February 10, 2010 11:17 AM
> To: antlr-interest at antlr.org
> >
> > I am not sure what you mean by scale to be honest. Performing an
> upcase in
> > LA isn't much of an overhead compared to everything else. I use this
> on my
> > SQL grammars with no problems and there are hundreds of keywords to
> > recognize. But if you prefer JFlex, then so be it :-)
> >
> > Jim
> >
> 
> You saying that made me somewhat suspicious that I might have
> overlooked
> something, and indeed it seems so; I made an assumption that the
> character
> data of tokens was being converted to upper case by that solution, but
> on
> second reading it seems that this is only done for matching purposes. I
> have
> lots of stuff (quoted strings, etc) that must not be touched. Hm. I may
> rethink :)

Yes - the common case is that you don't want to change the input code, just test it in a case insensitive way. That's why the solution just matches and leaves the input alone.

Jim






More information about the antlr-interest mailing list