[antlr-interest] using RandomPhrase feature

[Terence Parr]

On Jul 21, 2009, at 9:46 PM, <harsha.nagesh at thomsonreuters.com> <harsha.nagesh at thomsonreuters.com 
 > wrote:

> I am playing with the RandomPhrase feature in antlr to generate random
> phrases based on a grammar. Infact I took the Expr grammar on antlr
> website to try out this feature. Firstly I had to bump up the heap  
> size
> to more than 512mb to get any output from the program. After that I  
> did
> get a few expressions generated and sometimes it simply exited without
> printing anything (perhaps the expression generated just matched  
> NEWLINE
> in the Expr grammar). Sometimes it went out of memory.

It is only a prototype. Unfortunately, it uses the stack to dig down  
deeper into the state machine. This will only work for so long of a  
string before it explodes stack. :(

Ter
>
> I am wondering if anybody has suggestions on how to control the  
> behavior
> of this feature ? Perhaps feed our own lexical tokens to the generator
> and control the length of the expression ?
>
> Thanks
> Harsha
>
> List: http://www.antlr.org/mailman/listinfo/antlr-interest
> Unsubscribe: http://www.antlr.org/mailman/options/antlr-interest/your-email-address