On 8/1/2012 10:53 PM, Jonathan M Davis wrote:
What I was expecting there to be was a type which was a range of tokens. You passed the source string to a function which returned that range, and you iterated over it to process each token. What you appear to have been arguing for is another type which you get the range from which holds additional state (e.g. the indentifier table). So, you're saying now that you only meant that the token type needs to be separate from the range type (which I would have thought would be the only way to do it), or are you indeed saying that there should be another type that the range comes from and which holds additional state?
You could do it either way.