http://d.puremagic.com/issues/show_bug.cgi?id=2173

           Summary: foreach iteration not possible for associative array
                    indexed with static array
           Product: D
           Version: 1.031
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P3
         Component: DMD
        AssignedTo: bugzilla@digitalmars.com
        ReportedBy: georg@op-co.de


Trying to foreach through an AA[SA] generates a compiler error:

import std.stdio;
import std.md5;
int main(char[][] args) {
        int[ubyte[16]] md5count;
        foreach (md5, count; md5count)
                writefln("%s: %d", std.md5.digestToString(md5), count);
        return 0;
}

Both gdc and dmd bail out on the foreach statement with:
"test.d(5): Error: cannot have out or ref parameter of type ubyte[16u]"

Possible workarounds:
 * use a wrapper struct for the static array (looks ugly)
 * use foreach (md5; md5count.keys) ... count = md5count[md5]
      (decreases performance almost by 50%)


-- 

http://d.puremagic.com/issues/show_bug.cgi?id=2173


davidl@126.com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |patch




------- Comment #1 from davidl@126.com  2008-06-26 01:39 -------
i bet the code actually meant as following:

mtype.c (2790):
            if (arg->storageClass & (STCout | STCref | STClazy))
            {
---             if (t->ty == Tsarray)
+++             if (t->ty != Tsarray)
                    error(loc, "cannot have out or ref parameter of type %s",
t->toChars());
            }

I hand-crafted a dmd which bypass this check, it can generate correct binary.


-- 

<d-bugmail@puremagic.com> wrote in message news:g3vdj3$114l$1@digitalmars.com...
> http://d.puremagic.com/issues/show_bug.cgi?id=2173

> i bet the code actually meant as following:
>
> mtype.c (2790):
>            if (arg->storageClass & (STCout | STCref | STClazy))
>            {
> ---             if (t->ty == Tsarray)
> +++             if (t->ty != Tsarray)
>                    error(loc, "cannot have out or ref parameter of type
> %s",
> t->toChars());
>            }

Now static arrays are the only type that can be out, ref, or lazy.  I don't think that's much of a fix.  Maybe just get rid of the check entirely.

> I hand-crafted a dmd which bypass this check, it can generate correct binary.

How can you do that?  The backend is closed source.

On Thu, 26 Jun 2008 17:19:41 +0400, Jarrett Billingsley <kb3ctd2@yahoo.com> wrote:

> <d-bugmail@puremagic.com> wrote in message
> news:g3vdj3$114l$1@digitalmars.com...
>> http://d.puremagic.com/issues/show_bug.cgi?id=2173
>
>> i bet the code actually meant as following:
>>
>> mtype.c (2790):
>>            if (arg->storageClass & (STCout | STCref | STClazy))
>>            {
>> ---             if (t->ty == Tsarray)
>> +++             if (t->ty != Tsarray)
>>                    error(loc, "cannot have out or ref parameter of type
>> %s",
>> t->toChars());
>>            }
>
> Now static arrays are the only type that can be out, ref, or lazy.  I don't
> think that's much of a fix.  Maybe just get rid of the check entirely.
>
>> I hand-crafted a dmd which bypass this check, it can generate correct
>> binary.
>
> How can you do that?  The backend is closed source.
>
>

That's not hard to patch je to jne, I assume :)

http://d.puremagic.com/issues/show_bug.cgi?id=2173





------- Comment #2 from davidl@126.com  2008-07-02 09:12 -------
umm, the patch is incorrect..
bypassing the whole test should work.
/*
            if (arg->storageClass & (STCout | STCref | STClazy))
            {
                if (t->ty == Tsarray)
                if (t->ty != Tsarray)
                    error(loc, "cannot have out or ref parameter of type %s",
t->toChars());
            }
*/


-- 

http://d.puremagic.com/issues/show_bug.cgi?id=2173


bearophile_hugs@eml.cc changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |bearophile_hugs@eml.cc
         Resolution|                            |WORKSFORME


--- Comment #3 from bearophile_hugs@eml.cc 2010-02-18 11:09:11 PST ---
I think this bug is now fixed, because fixed-sized arrays are managed by value.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------