Digital Mars - digitalmars.D.bugs - [Issue 2023] New: Returning from foreach body doesn't work as expected.

↑ ↓ ← → d-bugmail puremagic.com writes:

http://d.puremagic.com/issues/show_bug.cgi?id=2023

           Summary: Returning from foreach body doesn't work as expected.
           Product: D
           Version: 2.012
          Platform: PC
        OS/Version: Windows
            Status: NEW
          Severity: normal
          Priority: P2
         Component: DMD
        AssignedTo: bugzilla digitalmars.com
        ReportedBy: samukha voliacable.com


The bug is reproduced by the following code. Instead of making foo return, the
'return' statement in foo's foreach body breaks the loop:

----
struct S
{
    int a[];

    int opApply(int delegate(ref int) dg)
    {        
        foreach (x; a)
        {
            if(dg(x))
                return 1;
        }

        return 0;
    }
}

void foo()
{
    auto s = S([1, 2, 3]);

    foreach (x; s)
    {
        if (x == 1)
            return; // breaks the loop instead of returning from foo.
    }

    assert(false); // should never reach here.
}

void main()
{
    foo();   
}
----


--

Apr 21 2008

↑ ↓ ← → "Jarrett Billingsley" <kb3ctd2 yahoo.com> writes:

<d-bugmail puremagic.com> wrote in message 
news:bug-2023-3 http.d.puremagic.com/issues/...
 http://d.puremagic.com/issues/show_bug.cgi?id=2023

           Summary: Returning from foreach body doesn't work as expected.
           Product: D
           Version: 2.012
          Platform: PC
        OS/Version: Windows
            Status: NEW
          Severity: normal
          Priority: P2
         Component: DMD
        AssignedTo: bugzilla digitalmars.com
        ReportedBy: samukha voliacable.com


 struct S
 {
    int a[];

    int opApply(int delegate(ref int) dg)
    {
        foreach (x; a)
        {
            if(dg(x))
                return 1;
        }

        return 0;
    }
 }



You're implementing the opApply wrong.

if(auto result = dg(x))
    return result;

The result of the delegate is an int, not a bool, for a reason.  There are 
many possible return values, not just 1 or 0; you're just faking out the 
runtime by doing so.

Apr 21 2008

↑ ↓ ← → Max Samukha <nospam nospam.com> writes:

On Mon, 21 Apr 2008 08:46:55 -0400, "Jarrett Billingsley"
<kb3ctd2 yahoo.com> wrote:

<d-bugmail puremagic.com> wrote in message 
news:bug-2023-3 http.d.puremagic.com/issues/...
 http://d.puremagic.com/issues/show_bug.cgi?id=2023

           Summary: Returning from foreach body doesn't work as expected.
           Product: D
           Version: 2.012
          Platform: PC
        OS/Version: Windows
            Status: NEW
          Severity: normal
          Priority: P2
         Component: DMD
        AssignedTo: bugzilla digitalmars.com
        ReportedBy: samukha voliacable.com


 struct S
 {
    int a[];

    int opApply(int delegate(ref int) dg)
    {
        foreach (x; a)
        {
            if(dg(x))
                return 1;
        }

        return 0;
    }
 }



You're implementing the opApply wrong.

if(auto result = dg(x))
    return result;

The result of the delegate is an int, not a bool, for a reason.  There are 
many possible return values, not just 1 or 0; you're just faking out the 
runtime by doing so. 


Oops. Anybody knows why I assumed it's boolean? That is crazy.

Apr 21 2008

↑ ↓ ← → d-bugmail puremagic.com writes:

http://d.puremagic.com/issues/show_bug.cgi?id=2023





------- Comment #1 from wbaxter gmail.com  2008-04-21 07:57 -------
Here's my opportunity to restate my opinion that the programmer should not be
responsible for passing around opaque magic values generated by the compiler. 
At least not in a supposedly modern language.

I wrote a proposal about how to make opApply more user-friendly
Here: digitalmars.D:64686
And a few clarifications here: digitalmars.D:64706

Unfortunately I think it really requires macro() to make it pretty.


--

Apr 21 2008

↑ ↓ ← → d-bugmail puremagic.com writes:

http://d.puremagic.com/issues/show_bug.cgi?id=2023





------- Comment #2 from wbaxter gmail.com  2008-04-21 07:59 -------
(In reply to comment #1)

I think there's some way to get links to NG posts here automatically but I can
never remember what it is...

 I wrote a proposal about how to make opApply more user-friendly
 Here: digitalmars.D:64686


http://www.digitalmars.com/pnews/read.php?server=news.digitalmars.com&group=digitalmars.D&artnum=64686&header

 And a few clarifications here: digitalmars.D:64706


http://www.digitalmars.com/pnews/read.php?server=news.digitalmars.com&group=digitalmars.D&artnum=64706&header


--

Apr 21 2008

↑ ↓ ← → d-bugmail puremagic.com writes:

http://d.puremagic.com/issues/show_bug.cgi?id=2023


samukha voliacable.com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




------- Comment #3 from samukha voliacable.com  2008-04-21 09:46 -------
Maybe it makes sense to post your proposal to bugzilla as enhancement? For now,
I'm closing the bug as the example is invalid.


--

Apr 21 2008