Marcin 'Qrczak' Kowalczyk writes:
You can try using Parsec or another similar library for more complex tasks.
IMHO, Parsec is too complicated for something that simply. Stick to the example of implementing a 'lines' that uses \r\n instead of \n, if you will: I don't want any monads, I don't want error messages, I don't need any recursive descent parser. All I need is a simple tokenizer. In the past I have used 'lines' and then chopped off the last character off each line. Fair enough, but I think that there is a more general problem underneath that deserves to be solved. I see this function underneath Parsec. I would, under the right circumstances, use Parsec to write the functions I give to 'tokenize' as arguments! But I wouldn't write 'tokenize' in Parsec.
What if both return True, or if neither returns True?
That's an 'error'.
This is my SplitSeps (formerly Split1), assuming the second function is just the negation of the first one.
Great! Then I second that your function is added to the standard library. :-) Peter