what -I- consider a bug in *cabe

[Kenneth Brody]

Quoting Jay R. Ashworth (Sun, 3 Jun 2007 10:26:43 -0400):

> On Sat, Jun 02, 2007 at 03:20:45PM -0400, Fairlight wrote:
[...]
> > required."  Emphasis on -sole string literal-.  I take your point
> > about runtime expression evaluation, but if it were just a single
> > string literal there, there's no evaluating to be done--it's correct
> > or it's not.
>
> "a quoted literal that can only be numeric" is a syntactical particle
> that... I don't think I've ever seen in any other language.  So it
> doesn't surprise me that the lexer misses it.

Remember, the syntax says that an expression goes there.  A "single
string literal" happens to be one type of expression.  The contents
of that string are irrelevent to the syntax.

Now, should a semantic check be performed at compile time as well?
It's possible to make an exception for the case of an expression
being a "single string literal", but that would have to go beyond
mere type-checking.  Remember, but "$" and "4" are simple literal
strings.  The "4" is not a numeric field, but rather a string that
happens to contain a value number.  How much checking should be
performed?  Should SQRT("-1") be flagged at compile time as well?
What about MID(field,"4.5","7.3")?

Should every place where an expression is allowed, check to see if
that expression consists of a "single string literal", and validate
the string at compile time?

--
KenBrody at BestWeb dot net        spamtrap: <g8ymh8uf001 at sneakemail.com>
http://www.hvcomputer.com
http://www.fileProPlus.com