Beefy Boxes and Bandwidth Generously Provided by pair Networks
"be consistent"

Re^3: if/else syntax

by Crackers2 (Parson)
on Jun 26, 2009 at 16:10 UTC ( #775099=note: print w/replies, xml ) Need Help??

in reply to Re^2: if/else syntax
in thread if/else syntax

I don't think so. With if/else:

if a if b c else d
could mean
if a { if b { c } else { d } } or if a { if b { c } } else { d }

With the ternary operator that would be

a ? b ? c : d : () and a ? b ? c : () : d

so there's no abmbiguity.

Replies are listed 'Best First'.
Re^4: if/else syntax
by gwadej (Chaplain) on Jun 26, 2009 at 16:44 UTC

    There's actually no ambiguity in the C syntax for your example. The compiler understands it fine. The problem is that the programmer/reader of the code does not always read it the same way.

    if(a) if(b) c(); else d();

    According to what I recall, the else matches with the nearest if in this case. Unfortunately, indentation gives humans fits on this.

    if(a) if(b) c(); else d();

    This does the same thing (else goes with if(b)). As I said, there's nothing ambiguous about it.

    G. Wade
      thanks, thats exactly what I'm trying to say, if the programmer knows how the compiler interprets the code he can use it in the right way.

      The compiler understands it fine.

      Not true. How can the compiler understand it fine when C doesn't define it

      if(a) if(b) c(); else d();


      if(a) { if(b) c(); else d(); }


      if(a) { if(b) c(); } else d();

      Both are right. Neither is right.

      I don't know if that was fixed in recent versions of C.

      There's actually no ambiguity in the C syntax for your example.

      That's not true either. You contradict yourself in the next sentence. You write code for both people and for compilers to read. Ambiguity for either is still ambiguity.

      Anyway, the topic wasn't ambiguity but (the avoidance of) traps. C's if/else definitely has pitfalls.

        Sorry, but you are wrong. Since immemorial times and K&R release 1,
        if(a) if(b) c(); else d();
        is DEFINED as
        if(a) { if(b) c(); else d(); }
        But, this does not mean there is no ambiguity. It's just plain wrong to assume things without the braces because software changes constantly. So, one day you have
        if( a ) if( b ) c(); else d(); else e();
        and then you see some reason for the else d(); to go away, and you stay with
        if( a ) if( b ) c(); else e();
        That does not do what you intended and you can be looking for this bug in another place kilometers/miles away (in the code) because you left out some initialization done by e(); etc etc.

        IOW: You could do that -- because it's really well-defined behaviour for the C language --, but it does not mean you should.

        []s, HTH, Massa (κς,πμ,πλ)

        I'd suggest you check your standard again. It was defined in K&R and in the ANSI/ISO standard.

        Update: Just to make sure I hadn't misremembered, I dug out the ISO C++ standard (which is all I have handy at the moment.) Section 6.4.1 addresses this issue.

        In the second form of if statement (the one including else), if the first substatement is also an if statement then that inner if statement shall contain an else part.

        There was also no contradiction in my later statement. The is no ambiguity in the standard (or compiler) interpretation. But there is a problem with the way the programmer will read it. This is why Perl made the braces mandatory and Python made the indention mandatory.

        G. Wade
Re^4: if/else syntax
by Anonymous Monk on Jun 26, 2009 at 16:27 UTC
    so why we can't have a if/else syntax without brackets that behaves just like the ternary operator ?
    if a { if b { c } else { d } } or if a { if b { c } } else { d }

    Does not matter who the compiler will assume, since if it get implemented we would know it before hand

Log In?

What's my password?
Create A New User
Node Status?
node history
Node Type: note [id://775099]
and all is quiet...

How do I use this? | Other CB clients
Other Users?
Others drinking their drinks and smoking their pipes about the Monastery: (10)
As of 2018-06-22 21:24 GMT
Find Nodes?
    Voting Booth?
    Should cpanminus be part of the standard Perl release?

    Results (124 votes). Check out past polls.