Uwe Stöhr <[EMAIL PROTECTED]> writes:
>> I think the new method is not needed because direct code is almost as
>> compact.
>> Moreover, it is better IMO to keep the Parser interface small.
>
> But getting an option that is in parentheses will surely be useful for
> the future as well, therefore I didn't use the direct code but
> introduced this function in parser.cpp. The function is as small as
> possible and easy to understand.
Your getParenthesesOpt function has a problem: since it is modelled
after getOpt(), it allows for an optional argument. But () arguments
are not optional in general. So at least this function should be
renamed to Parser::getFullParentheseArg() and should fail (how?) when
the opening parenthesis is not found. Still, it would have semantics
different from the other functions.
I think therefore that this function is very misleading. What could be
useful for ERT, though, is a version of getFullArg that returns the
outer delimiters. Moreover, one could make use of the currently unused
FLAG_OPTION = 1 << 11, // read [...] style option
FLAG_BRACED = 1 << 12, // read {...} style argument
enum flags instead of passing explicitly the left and right
delimiters.
Note also that the new code just needs one more line, so there is not
much lost.
JMarc
