Re: what scanner scheme is efficient?

> Is scanning/parsing _speed_ ever reason to go toward using a hash
> table for identifying keywords?


Sure. In C like languages, really fast compilation may require an
integrated preprocessor, and since even keywords can be #define'd to
something else, it would be wrong to specifically scan for them as such.
Therefore, the best approach is often to recognise an identifier
([A-Za-z_][A-Za-z_0-9]*), then lookup (perhaps via hash table) several
properties of the identifier:
* is it a macro name? (and perhaps its macro definition)
* is it a keyword? (and its keyword token)
* other useful attributes of the id (id known to be, or definitely not be,
a type name, class name, member name, etc.)


Jan Gray
Redmond, WA
--

Return to the comp.compilers page.
Search the comp.compilers archives again.