How should I parse keywords when writing a C Compiler?

Question

I am currently in the process of writing a C to Assembly compiler, it is not meant to be practical, but I would like to do it for the educational value. I was wondering when I am testing for keywords, is there any more efficient way rather than just reading in the next word in the file and then running it through a bunch of nested if statements that test for the keywords. Is there any better way?

Answer 1

Your question is actually quite specific. You are asking about how to build the lexical analyzer, also known as the scanner, and how to efficiently and conveniently recognize keywords. The scanner is the first phase of a typical compiler, and it converts the source code, which is a sequence of characters, to a sequence of tokens, where a token is a unit such as a number, an operator or a keyword.

Since keywords match the pattern for general identifiers, a common trick is to put all the keywords in the symbol table, together with information that it is a keyword. Then, when the scanner finds an identifier, it as usual searches the symbol table to see if that identifier has been seen before. If this identifier was a kewyord, it will be found, together with the information about which keyword it is.