[antlr-interest] Best practice to handle Lexer backtracking demand
Joachim Schrod
jschrod at acm.org
Sun Aug 15 01:45:15 PDT 2010
Gerald Rosenberg wrote:
>
>> Input: aaaaXXXXbbbb wwwXXX ddddYYYzzzz
>>
> How is XXXX guaranteed to be unambiguous with any other fragment of
> aaaaXXXXbbbb? That is, how can you be sure that a fragment like aaaX or
> XXXXb will never match a different start marker.
The data generating service guarantees it. (It escapes characters
if any complete marker substring happens to be in the data.)
> Is there a case
> distinction, as implied, or something more interesting? Is the
> distinction the same for the end marker?
No and no. The markers are strings like `prenames', `prenamee',
`surnames', etc. That's why I used that example in my original email.
Joachim
--
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Joachim Schrod Email: jschrod at acm.org
Roedermark, Germany
More information about the antlr-interest
mailing list