On Nov 4, 10:44 pm, "Just Another Victim of the Ambient Morality" <[EMAIL PROTECTED]>
> I believe there is a cure and it's called recursive descent parsing. > It's slow, obviously, but it's correct and, sometimes (arguably, often), > that's more important the execution speed. Recursive decendent parsing is not necessarily slow but from your remarks above I infer you want a general RD parser with backtracking: when one rule doesn't match, try another one to derive the current symbol in the input stream. I'm not sure one needs to start again with a naive approach just to avoid any parser theory. For a user of a parser it is quite important whether she has to wait 50 seconds for a parse to run or 50 milliseconds. I don't like to compromise speed for implementation simplicity here. -- http://mail.python.org/mailman/listinfo/python-list
