andreas.raab at gmx.de
Sat Sep 17 22:44:08 UTC 2005
Alexander Lazarević wrote:
> In addition there could be a new
> AmbiguousNumericLiteralWarning that could be raised whenever a numerical
> literal is parsed that has base>14 and contains an e character. The
> default action could be to get a reassurance that the number with
> exponent is what is wanted. Maybe too much ado though.
Actually, that's a *great* idea if you generalize it a little, e.g.,
make it so that you get the warning whenever a lower-case character is
encountered where it could be interpreted as a digit under the current
base[*]. This would generally cover the problem of people not being
aware that they need to spell the digits as upper case. It sure as hell
would've solved my problem ;-)
[*] As a matter of fact this might be signaled _whenever_ a character is
encountered that cannot be uniquely interpreted in a numeric literal.
For example, if you write "10F" the compiler will perfectly interpret
that as "10 F" where, clearly, either that needs to be 16r10F or I'm
such a lazy guy that I don't put white spaces where they belong. In
either case, a warning would be adequate.
More information about the Squeak-dev