Re: String concatenation design




"Barry" <barryg@xxxxxxxxxxxxxxxxxx> wrote in message
news:122nfiira6b9l3d@xxxxxxxxxxxxxxxxxxxxx

<andy@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote in message
news:1143553103.017921.221240@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

Barry wrote:
<andy@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote in message
news:1143540676.814646.118800@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

Anton Treuenfels wrote:
I have a numeric expression parser that allows literal strings to
be
used in
logical expressions (such as comparisons) because the results of
those
are
numeric. I gave the relevant string operators a very high
precedence
because
that seemed the easiest way to prevent numeric operators from
engaging
string operands.

So the expression

"ABC" < "DEF" + 1

is parsed as

("ABC" < "DEF") + 1

and yields the value 2.

why does boolean + numeric -> numeric.

boolean + numeric could go to boolean or be disallowed IMO.
Or is the current semantic particularly useful and if so where?

regards
Andy Little


Who said anything about booleans?

What else should a less-than comparison return?

cheers
Andy Little


In C it returns an int, in perl it is just a scalar.


Actually, I should have said the result is an int. It doesn't "return"
anything :-).


.

Relevant Pages

  • Re: String concatenation design
    ... Barry wrote: ... Anton Treuenfels wrote: ... and yields the value 2. ... boolean + numeric could go to boolean or be disallowed IMO. ...
    (comp.programming)
  • Re: String concatenation design
    ... Anton Treuenfels wrote: ... string operands. ... and yields the value 2. ... boolean + numeric could go to boolean or be disallowed IMO. ...
    (comp.programming)
  • Re: String concatenation design
    ... Anton Treuenfels wrote: ... string operands. ... and yields the value 2. ... boolean + numeric could go to boolean or be disallowed IMO. ...
    (comp.programming)
  • Re: assert-like debug { /* code */ } feature?
    ... operand, and yields that value as the value of its expression. ... the value of the expression) is a boolean, ... I don't think "operator is a predicate" idea has any merit. ...
    (comp.lang.java.programmer)