[antlr-interest] problem with antlr-2.7.5

[Peggy Fieland]

I doubt there are problems with the examples but I'll check when I get into work --
  the input with the problem is very large (an oracle package) and the parser in question is
  also very large. I've also modified antlr 2.7.5 to support unicode (using your example as a starting point.)
   
  I'll be spending today narrowing down the problem 
   
  I doubt it's an actual antlr problem -- I'd guess it's either a problem in my grammar exposed
  by the compiler change or a gcc 4.1.1 code generator bug ...
   
  <sigh>.  I'm hoping it's a problem with my grammar.
   
  
Ric Klaren <ric.klaren at gmail.com> wrote:
  Hi,

On 2/15/07, Peggy Fieland wrote:
> Everything works fine when built with gcc 3.2.2 (antlr and test driver,
> parser input) but when built with gcc 4.1.1 (antlr, test driver) that same
> input loops infinitely in the c++ runtime --
> I think in token buffer delete -- anyway somewhere in the ref/deref stuff.

Hmm weird, I think I've used antlr generated code on gcc 4.1.1 without
problems, at least I've got 4.1.1 on a fedora box at home. I can do a
check when I get back from work.

Are there also problems with examples supplied with antlr?

Cheers,

Ric



Margaret Fieland
    www.margaretfieland.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.antlr.org/pipermail/antlr-interest/attachments/20070215/0b55197c/attachment-0001.html