www.digitalmars.com         C & C++   DMDScript  

digitalmars.D.learn - Templated delegate

reply Mitja <odtihmal gmail.com> writes:
I would like to have a template which returns
appropriate delegate, something like this:

module mod1;

import std.algorithm;
import std.array;
import std.stdio;

template DG(RT, T, F)
{
  auto DG(F fun)
  {
    RT delegate(T p)dg;
    dg = fun;
    return dg;
  }
}


void main()
{
  string[] haystack = ["item 1", "item 2", "item 3"];
  string needle = "item 2";

  auto flt = (string s){return s == needle;};

  auto dg = DG!(bool, string, typeof(flt))(flt);

  auto result = filter!(dg)(haystack);

  writeln(result);
}

The thing is I don't know what am I doing wrong here,
because sometimes works and sometimes doesn't.
For instance, if I'd added another file along with mod1.d, let's say,
mod2.d which contained:

module mod2;

import std.conv;
import std.string;
import std.random;

and compiled it like this: dmd mod2.d mod1.d, the program would produce
segmentation fault.
It works fine when packages in mod2 are omitted.

What would be the correct way for templated delegate?
Jan 10 2011
next sibling parent reply Stewart Gordon <smjg_1998 yahoo.com> writes:
On 10/01/2011 13:02, Mitja wrote:
 I would like to have a template which returns
 appropriate delegate, something like this:

For what purpose? All you seem to have is a long-winded identity function. <snip>
 and compiled it like this: dmd mod2.d mod1.d, the program would produce
 segmentation fault.

I can't reproduce (DMD 2.051, Windows). In what setup are you experiencing the problem?
 It works fine when packages in mod2 are omitted.

When which packages in mod2 are omitted? I can only make out that this is a compiler bug that needs to be investigated.
 What would be the correct way for templated delegate?

It depends on what you're trying to do. Stewart.
Jan 10 2011
parent Mitja <odtihmal gmail.com> writes:
Stewart Gordon Wrote:

 On 10/01/2011 13:02, Mitja wrote:
 I would like to have a template which returns
 appropriate delegate, something like this:

For what purpose? All you seem to have is a long-winded identity function. <snip>
 and compiled it like this: dmd mod2.d mod1.d, the program would produce
 segmentation fault.

I can't reproduce (DMD 2.051, Windows). In what setup are you experiencing the problem?

Debian GNU/Linux 5.0, kernel 2.6.26-2-686 #1 SMP i686. dmd is D 2.051. As well as on Windows XP as guest OS in VirtualBox. XP prints out: "object.Error: Access Violation"
 
 It works fine when packages in mod2 are omitted.

When which packages in mod2 are omitted? I can only make out that this is a compiler bug that needs to be investigated.
 What would be the correct way for templated delegate?

It depends on what you're trying to do. Stewart.

Jan 10 2011
prev sibling parent reply "Simen kjaeraas" <simen.kjaras gmail.com> writes:
Mitja <odtihmal gmail.com> wrote:

 I would like to have a template which returns
 appropriate delegate, something like this:

So like this? module mod1; import std.algorithm; import std.functional; import std.stdio; void main( ) { auto haystack = ["a","b","c"]; auto needle = "b"; auto flt = (string s){return s == needle;}; auto dg = toDelegate( flt ); auto result = filter!dg(haystack); writeln(result); } I would have to argue that the delegate is hardly templated, though. Not sure what that means.
 What would be the correct way for templated delegate?

Depending on what you mean by 'templated delegate', I would say std.functional's toDelegate, as used above. Perhaps if you explained what you mean by 'templated delegate', I could give a better answer. -- Simen
Jan 10 2011
next sibling parent Mitja <odtihmal gmail.com> writes:
Thanks, toDelegate is what I've been looking for.

It still segfaults, though.
The setup:
=======
module mod1;

import std.algorithm;
import std.functional;
import std.stdio;

void main( ) {
   auto haystack = ["a","b","c"];
   auto needle = "b";
   auto flt = delegate(string s){return s == needle;};
   auto dg = toDelegate( flt );
   auto result = filter!dg(haystack);
   writeln(result);
}
=========
module mod2;

import std.string;
// or import std.conv;
// or import std.random;
========
dmd mod2.d mod1.d compiles, but compiled program segfaults.
System is Debian GNU/Linux 5.0 on 2.6.26-2-686 #1 SMP i686.
dmd is D 2.051.


Simen kjaeraas Wrote:

 Mitja <odtihmal gmail.com> wrote:
 
 I would like to have a template which returns
 appropriate delegate, something like this:

So like this? module mod1; import std.algorithm; import std.functional; import std.stdio; void main( ) { auto haystack = ["a","b","c"]; auto needle = "b"; auto flt = (string s){return s == needle;}; auto dg = toDelegate( flt ); auto result = filter!dg(haystack); writeln(result); } I would have to argue that the delegate is hardly templated, though. Not sure what that means.
 What would be the correct way for templated delegate?

Depending on what you mean by 'templated delegate', I would say std.functional's toDelegate, as used above. Perhaps if you explained what you mean by 'templated delegate', I could give a better answer. -- Simen

Jan 10 2011
prev sibling parent Stanislav Blinov <blinov loniir.ru> writes:
10.01.2011 21:08, Mitja пишет:
 Thanks, toDelegate is what I've been looking for.

 It still segfaults, though.
 The setup:
 =======
 module mod1;

 import std.algorithm;
 import std.functional;
 import std.stdio;

 void main( ) {
     auto haystack = ["a","b","c"];
     auto needle = "b";
     auto flt = delegate(string s){return s == needle;};
     auto dg = toDelegate( flt );
     auto result = filter!dg(haystack);
     writeln(result);
 }
 =========
 module mod2;

 import std.string;
 // or import std.conv;
 // or import std.random;
 ========
 dmd mod2.d mod1.d compiles, but compiled program segfaults.
 System is Debian GNU/Linux 5.0 on 2.6.26-2-686 #1 SMP i686.
 dmd is D 2.051.

not so long ago: in 2.051 functions with 'auto' result type and at least one parameter cause segfault during compilation when they are in a separate module. I can't remember issue number right now (and have no internet connection to find it), but it's pretty recent.
Jan 11 2011