What to do with multi-character operators when parsing ?

[manasij7479]

I thought that I'd skip multi-character operators entirely, but am having difficulties to come up with enough single-char operators for covering the ground.

My idea was to make the tokens run through another function ...say.. the 'joiner' which would do stuff like..
*Join multi character operators together.
*Convert integers separated by a ' . ' into doubles ..etc

But I can't imagine what to do when the joining of the operators isn't necessary.

For example :
When the input is a mathematical expression in rpn(postfix) notation, two '+' `s together does not necessarily mean an increment operator, but it could mean so at times!

[whiteflags]

Your parser should be smart enough to take them as one token. Note that lexical parsing is kind of a big deal and there is a lot of literature on the subject. From your problem, I think you don't have a context free language to parse, so your lexer should look at the context of trouble statements to decide how to parse them. You've given me kind of a poor example, but if the statement is a math expression then the RPN must have at least 4 other operands. If it's not exactly that, then you must be using postfix increment. If it's not that, then there is a syntax error.

[laserlight]

Basically, you are asking how to handle a case where there can be ambiguity. One way is to specify the grammar such that it is not ambiguous for the parser that you wish to implement. Another way is to perform lookahead such that the parser can resolve the ambiguity with the information gained from the lookahead.