Digital Mars - digitalmars.D.announce - dmd 1.056 and 2.040 release

The Anh Tran <trtheanh gmail.com> writes:

First, i would like to say: thank you.

Second, i would like to have your opinion:
"added static/final function implementations to interfaces"

Could you allow normal function imp to interfaces too? Interface will 
look like c++ struct, but the difference is the way inheritance 
organized, so that diamond shape problem does not exist.

And you could simplify something too. For example: for comparison 
between classes, we just inherit Comparable interface, which has 
implementation of many op<, op>, op<=, op>=, ...

(Well, in fact, this is done in Scala, it is named linearization). :)

For example:

class Biology
{
     ......
}

interface Animal
{
     void eat() { ... }
     void run() { ... }
}

interface Bird
{
     void fly() { ... }
}

class Penguin : Biology, Animal, Bird
{
}

The left most class / interface will be the top super. Each interface 
remain will solve its parent at definition site (line "class Penguin : 
Biology, Animal, Bird")
Result: Penguin -> Bird -> Animal -> Biology.

Details of linearization are in page 52, ScalaRef.
http://www.scala-lang.org/sites/default/files/linuxsoft_archives/docu/files/ScalaReference.pdf

Thanks.

Jan 29 2010

Lionello Lunesu <lio lunesu.remove.com> writes:

On 30-1-2010 15:53, The Anh Tran wrote:
 First, i would like to say: thank you.
 
 Second, i would like to have your opinion:
 "added static/final function implementations to interfaces"
 
 Could you allow normal function imp to interfaces too? Interface will
 look like c++ struct, but the difference is the way inheritance
 organized, so that diamond shape problem does not exist.


Use "final".

L.

Jan 30 2010

KennyTM~ <kennytm gmail.com> writes:

On Jan 30, 10 15:13, Walter Bright wrote:
 http://www.digitalmars.com/d/1.0/changelog.html
 http://ftp.digitalmars.com/dmd.1.056.zip


 http://www.digitalmars.com/d/2.0/changelog.html
 http://ftp.digitalmars.com/dmd.2.040.zip

 Thanks to the many people who contributed to this update!


Is the changelog page broken? On the 2.0 changelog it starts with "$(D_S 
D Change Log,", and without a navigation bar.

Jan 30 2010

torhu <no spam.invalid> writes:

On 30.01.2010 09:32, KennyTM~ wrote:
 On Jan 30, 10 15:13, Walter Bright wrote:
  http://www.digitalmars.com/d/1.0/changelog.html
  http://ftp.digitalmars.com/dmd.1.056.zip


  http://www.digitalmars.com/d/2.0/changelog.html
  http://ftp.digitalmars.com/dmd.2.040.zip

  Thanks to the many people who contributed to this update!


 Is the changelog page broken? On the 2.0 changelog it starts with "$(D_S
 D Change Log,", and without a navigation bar.


And 1.056 is missing, only its changelog is present, but under the 1.055 
header.

Jan 30 2010

bearophile <bearophileHUGS lycos.com> writes:

Regarding this:
Bugzilla 3556: version(CTFE)

I have written the following little D2 program:


import std.stdio: printf;

static if (__ctfe) {
    int foo() {
        return 1;
    }
} else {
    int foo() {
        return 2;
    }
}

enum int x1 = foo();

void main() {
    int x2 = foo();
    printf("%d %d\n", x1, x2);
}


But it doesn't work, I don't understand why.
(Isn't the usage of __ctfe normally done at compile time? Otherwise this
feature introduces another very special case in the language).

----------------

Can you tell me the purpose of the following 3 changes?

ModuleInfo changed from class to struct

added static/final function implementations to interfaces

http://dsource.org/projects/dmd/changeset/339

Thank you,
bye,
bearophile

Jan 30 2010

dsimcha <dsimcha yahoo.com> writes:

== Quote from bearophile (bearophileHUGS lycos.com)'s article
 Regarding this:
 Bugzilla 3556: version(CTFE)
 I have written the following little D2 program:
 import std.stdio: printf;
 static if (__ctfe) {
     int foo() {
         return 1;
     }
 } else {
     int foo() {
         return 2;
     }
 }
 enum int x1 = foo();
 void main() {
     int x2 = foo();
     printf("%d %d\n", x1, x2);
 }
 But it doesn't work, I don't understand why.
 (Isn't the usage of __ctfe normally done at compile time? Otherwise this
feature


 ----------------
 Can you tell me the purpose of the following 3 changes?
 ModuleInfo changed from class to struct
 added static/final function implementations to interfaces
 http://dsource.org/projects/dmd/changeset/339
 Thank you,
 bye,
 bearophile


Because a compile-time __ctfe turned out to be almost impossible to implement.
Therefore, __ctfe is nominally a regular (runtime) variable that evaluatest to
true at compile time and false at runtime.  The proper use is if(__ctfe), not
static if(__ctfe).  However, if(0) statements are thrown out by the code gen, so
there should be no runtime performance hit for using if(__ctfe).

Jan 30 2010

Walter Bright <newshound1 digitalmars.com> writes:

bearophile wrote:
 Can you tell me the purpose of the following 3 changes?
 
 ModuleInfo changed from class to struct


Eliminate unnecessary dependencies on Object's vtbl[]

 added static/final function implementations to interfaces


Support NVI (Non Virtual Inheritance) idiom.

Jan 30 2010

Jacob Carlborg <doob me.com> writes:

On 1/30/10 08:13, Walter Bright wrote:
 http://www.digitalmars.com/d/1.0/changelog.html
 http://ftp.digitalmars.com/dmd.1.056.zip


 http://www.digitalmars.com/d/2.0/changelog.html
 http://ftp.digitalmars.com/dmd.2.040.zip

 Thanks to the many people who contributed to this update!


Very nice.
What happened to the change log for 1.055?
Why can't we get version(CTFE) in D1 also?

Jan 30 2010

grauzone <none example.net> writes:

Jacob Carlborg wrote:
 On 1/30/10 08:13, Walter Bright wrote:
 http://www.digitalmars.com/d/1.0/changelog.html
 http://ftp.digitalmars.com/dmd.1.056.zip


 http://www.digitalmars.com/d/2.0/changelog.html
 http://ftp.digitalmars.com/dmd.2.040.zip

 Thanks to the many people who contributed to this update!


 Very nice.
 What happened to the change log for 1.055?
 Why can't we get version(CTFE) in D1 also?


You can ask this question for all other D2 features backwards compatible 
to D1 too. Why isn't __traits in D1? Why can't D1 have thread local 
variables (I'm not talking about TLS-by-default here)?

Jan 30 2010

Jacob Carlborg <doob me.com> writes:

On 1/30/10 14:24, grauzone wrote:
 Jacob Carlborg wrote:
 On 1/30/10 08:13, Walter Bright wrote:
 http://www.digitalmars.com/d/1.0/changelog.html
 http://ftp.digitalmars.com/dmd.1.056.zip


 http://www.digitalmars.com/d/2.0/changelog.html
 http://ftp.digitalmars.com/dmd.2.040.zip

 Thanks to the many people who contributed to this update!


 Very nice.
 What happened to the change log for 1.055?
 Why can't we get version(CTFE) in D1 also?


 You can ask this question for all other D2 features backwards compatible
 to D1 too. Why isn't __traits in D1? Why can't D1 have thread local
 variables (I'm not talking about TLS-by-default here)?


Because D1 has already got at least two new version identifier since it 
became feature freeze. But I see now that it's actually not a version 
identifier, it's a global bool.

Jan 30 2010

Ary Borenszweig <ary esperanto.org.ar> writes:

Walter Bright wrote:
 
 http://www.digitalmars.com/d/1.0/changelog.html
 http://ftp.digitalmars.com/dmd.1.056.zip
 
 
 http://www.digitalmars.com/d/2.0/changelog.html
 http://ftp.digitalmars.com/dmd.2.040.zip
 
 Thanks to the many people who contributed to this update!


Very nice! Each release kills a lot of bugs and adds small but very 
powerful features.

About the interface functions, this compiles:

--
import std.stdio;

interface One {

   final void foo() {
     writefln("One");
   }

}

interface Two {

   final void foo() {
     writefln("Two");
   }

}

class X : One, Two {
}

class Y : Two, One {
}

void main() {
   X x = new X();
   x.foo(); // prints "One"
   Y y = new Y();
   y.foo(); // prints "Two"
}
--

Is this intended behaviour? Might leads to obscure bugs...

Jan 30 2010

"Robert Jacques" <sandford jhu.edu> writes:

On Sat, 30 Jan 2010 10:56:28 -0500, Ary Borenszweig <ary esperanto.org.ar>  
wrote:

 Walter Bright wrote:
  http://www.digitalmars.com/d/1.0/changelog.html
 http://ftp.digitalmars.com/dmd.1.056.zip
   http://www.digitalmars.com/d/2.0/changelog.html
 http://ftp.digitalmars.com/dmd.2.040.zip
  Thanks to the many people who contributed to this update!


 Very nice! Each release kills a lot of bugs and adds small but very  
 powerful features.

 About the interface functions, this compiles:

 --
 import std.stdio;

 interface One {

    final void foo() {
      writefln("One");
    }

 }

 interface Two {

    final void foo() {
      writefln("Two");
    }

 }

 class X : One, Two {
 }

 class Y : Two, One {
 }

 void main() {
    X x = new X();
    x.foo(); // prints "One"
    Y y = new Y();
    y.foo(); // prints "Two"
 }
 --

 Is this intended behaviour? Might leads to obscure bugs...


This looks like a form of function hijacking, so it should be a  
accepts-invalid bug.

Jan 30 2010

Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:

Ary Borenszweig wrote:
 Walter Bright wrote:
 http://www.digitalmars.com/d/1.0/changelog.html
 http://ftp.digitalmars.com/dmd.1.056.zip


 http://www.digitalmars.com/d/2.0/changelog.html
 http://ftp.digitalmars.com/dmd.2.040.zip

 Thanks to the many people who contributed to this update!


 Very nice! Each release kills a lot of bugs and adds small but very 
 powerful features.
 
 About the interface functions, this compiles:
 
 -- 
 import std.stdio;
 
 interface One {
 
   final void foo() {
     writefln("One");
   }
 
 }
 
 interface Two {
 
   final void foo() {
     writefln("Two");
   }
 
 }
 
 class X : One, Two {
 }
 
 class Y : Two, One {
 }
 
 void main() {
   X x = new X();
   x.foo(); // prints "One"
   Y y = new Y();
   y.foo(); // prints "Two"
 }
 -- 
 
 Is this intended behaviour? Might leads to obscure bugs...


I think this should lead to a compile-time error.

Andrei

Jan 30 2010

Yao G <foo bar.com> writes:

Walter Bright Wrote:

 
 http://www.digitalmars.com/d/1.0/changelog.html
 http://ftp.digitalmars.com/dmd.1.056.zip
 
 
 http://www.digitalmars.com/d/2.0/changelog.html
 http://ftp.digitalmars.com/dmd.2.040.zip
 
 Thanks to the many people who contributed to this update!


Looks like this doesn't work anymore with this release (D2):

---
module Foo;

import core.runtime;
import std.c.windows.windows;

class Bar {}

extern(Windows)
int WinMain(HINSTANCE instance, HINSTANCE, LPSTR cmdLine, int cmdShow )
{
    void exceptionHandler(Throwable e) {
        throw e;
    }

    Runtime.initialize( &exceptionHandler );
    
    auto foo = new Foo();
    
    Runtime.terminate ( &exceptionHandler );

    return 0;
}
---

When I run the program it "segfaults" with the following message (when I debug
it):

Unhandled Exception: EXCEPTION_ACCESS VIOLATION(0xc000005) at
object._moduleCtor2.ModuleInfo

The exception is thrown from the runtime initializer. It used to work with the
previous DMD version (2.039). I think that the change of ModuleInfo from class
to struct somehow affected this (just speculating).

I did a wrapper for the Windows API and now none of my programs run when I use
DMD 2.040. Damn, I was excited by the new  disable attribute. :(

Jan 30 2010

Walter Bright <newshound1 digitalmars.com> writes:

Yao G wrote:
 The exception is thrown from the runtime initializer. It used to work
 with the previous DMD version (2.039). I think that the change of
 ModuleInfo from class to struct somehow affected this (just
 speculating).


Make sure you recompile everything, or you'll get crashes.

Jan 30 2010

Michel Fortin <michel.fortin michelf.com> writes:

$ dmd
Digital Mars D Compiler v2.040
Copyright (c) 1999-2009 by Digital Mars written by Walter Bright

Interestingly, copyright date is still 2009. Is this a bug or a feature?

-- 
Michel Fortin
michel.fortin michelf.com
http://michelf.com/

Jan 30 2010

Walter Bright <newshound1 digitalmars.com> writes:

Michel Fortin wrote:
 $ dmd
 Digital Mars D Compiler v2.040
 Copyright (c) 1999-2009 by Digital Mars written by Walter Bright
 
 Interestingly, copyright date is still 2009. Is this a bug or a feature?
 


It takes a while for me to get all the years changed.

Jan 30 2010

Don <nospam nospam.com> writes:

strtr wrote:
 Walter Bright Wrote:
 
 http://www.digitalmars.com/d/1.0/changelog.html
 http://ftp.digitalmars.com/dmd.1.056.zip


 http://www.digitalmars.com/d/2.0/changelog.html
 http://ftp.digitalmars.com/dmd.2.040.zip

 Thanks to the many people who contributed to this update!


 Do you ever find new bugs while fixing other?


Yes. It's a big problem with forward references, because they can affect 
unrelated parts of the compiler. I think that's the reason that Walter's 
been slow to apply patches for forward reference bugs.
Fortunately, most other bugs aren't like that. The progress is real.

Jan 31 2010

"Simen kjaeraas" <simen.kjaras gmail.com> writes:

Walter Bright <newshound1 digitalmars.com> wrote:

 http://www.digitalmars.com/d/1.0/changelog.html
 http://ftp.digitalmars.com/dmd.1.056.zip


 http://www.digitalmars.com/d/2.0/changelog.html
 http://ftp.digitalmars.com/dmd.2.040.zip

 Thanks to the many people who contributed to this update!


D2 changelog points  disable to attribute.html#deprecated,
should be attribute.html#disable

-- 
Simen

Jan 30 2010

Michel Fortin <michel.fortin michelf.com> writes:

On 2010-01-30 02:13:48 -0500, Walter Bright <newshound1 digitalmars.com> said:

 http://www.digitalmars.com/d/2.0/changelog.html
 http://ftp.digitalmars.com/dmd.2.040.zip


It's great to have TLS working on Mac OS X. But it looks like it 
suffers from the same linking problem as the module info section. I've 
added some useful observations to bugzilla that might help fix the 
issue in case you want to revisit it:

http://d.puremagic.com/issues/show_bug.cgi?id=3453


-- 
Michel Fortin
michel.fortin michelf.com
http://michelf.com/

Jan 30 2010

Walter Bright <newshound1 digitalmars.com> writes:

Michel Fortin wrote:
 It's great to have TLS working on Mac OS X. But it looks like it suffers 
 from the same linking problem as the module info section. I've added 
 some useful observations to bugzilla that might help fix the issue in 
 case you want to revisit it:
 
 http://d.puremagic.com/issues/show_bug.cgi?id=3453


You have to recompile *everything* with the new dmd, or anything 
dependent on the ModuleInfo will not work.

Jan 30 2010

Michel Fortin <michel.fortin michelf.com> writes:

On 2010-01-30 22:35:28 -0500, Walter Bright <newshound1 digitalmars.com> said:

 Michel Fortin wrote:
 It's great to have TLS working on Mac OS X. But it looks like it 
 suffers from the same linking problem as the module info section. I've 
 added some useful observations to bugzilla that might help fix the 
 issue in case you want to revisit it:
 
 http://d.puremagic.com/issues/show_bug.cgi?id=3453


 You have to recompile *everything* with the new dmd, or anything 
 dependent on the ModuleInfo will not work.


This has nothing to do with a partially recompiled program. It's about 
something interesting I added to this bug report. And trust me, I 
compiled the test case from scratch before using dumpobj and objdump on 
the object files and the linked executables.


-- 
Michel Fortin
michel.fortin michelf.com
http://michelf.com/

Jan 30 2010

Walter Bright <newshound1 digitalmars.com> writes:

Michel Fortin wrote:
 This has nothing to do with a partially recompiled program. It's about 
 something interesting I added to this bug report. And trust me, I 
 compiled the test case from scratch before using dumpobj and objdump on 
 the object files and the linked executables.


All right, except that some others had the same problem, and discovered 
that they hadn't recompiled everything, and the problem went away when 
they did.

Jan 30 2010

Walter Bright <newshound1 digitalmars.com> writes:

Michel Fortin wrote:
 It's about 
 something interesting I added to this bug report.


I submitted a patch for it, 363.

Feb 03 2010

Michel Fortin <michel.fortin michelf.com> writes:

On 2010-02-04 01:39:44 -0500, Walter Bright <newshound1 digitalmars.com> said:

 Michel Fortin wrote:
 It's about something interesting I added to this bug report.


 I submitted a patch for it, 363.


Thanks.

The linker keeps sections in the right order now... except for 
__tlscoal_nt. If I'm not mistaken, this section should probably get the 
same treatment as the *_beg and *_end sections.

-- 
Michel Fortin
michel.fortin michelf.com
http://michelf.com/

Feb 04 2010

Walter Bright <newshound1 digitalmars.com> writes:

Michel Fortin wrote:
 On 2010-02-04 01:39:44 -0500, Walter Bright <newshound1 digitalmars.com> 
 said:
 
 Michel Fortin wrote:
 It's about something interesting I added to this bug report.


 I submitted a patch for it, 363.


 Thanks.
 
 The linker keeps sections in the right order now... except for 
 __tlscoal_nt. If I'm not mistaken, this section should probably get the 
 same treatment as the *_beg and *_end sections.
 


I think you're right.

Feb 04 2010

Walter Bright <newshound1 digitalmars.com> writes:

Michel Fortin wrote:
 On 2010-02-04 01:39:44 -0500, Walter Bright <newshound1 digitalmars.com> 
 said:
 
 Michel Fortin wrote:
 It's about something interesting I added to this bug report.


 I submitted a patch for it, 363.


 Thanks.
 
 The linker keeps sections in the right order now... except for 
 __tlscoal_nt. If I'm not mistaken, this section should probably get the 
 same treatment as the *_beg and *_end sections.
 


Changeset 365

Feb 04 2010

Leandro Lucarella <llucax gmail.com> writes:

Simen kjaeraas, el 30 de enero a las 23:23 me escribiste:
 Walter Bright <newshound1 digitalmars.com> wrote:
 
http://www.digitalmars.com/d/1.0/changelog.html
http://ftp.digitalmars.com/dmd.1.056.zip


http://www.digitalmars.com/d/2.0/changelog.html
http://ftp.digitalmars.com/dmd.2.040.zip

Thanks to the many people who contributed to this update!


 D2 changelog points  disable to attribute.html#deprecated,
 should be attribute.html#disable


I reported that in the beta ML but it seems that wasn't important enough
to fix :S

-- 
Leandro Lucarella (AKA luca)                     http://llucax.com.ar/
----------------------------------------------------------------------
GPG Key: 5F5A8D05 (F8CD F9A7 BF00 5431 4145  104C 949E BFB6 5F5A 8D05)
----------------------------------------------------------------------

Jan 30 2010

BLS <windevguy hotmail.de> writes:

On 30/01/2010 08:13, Walter Bright wrote:
 http://www.digitalmars.com/d/1.0/changelog.html
 http://ftp.digitalmars.com/dmd.1.056.zip


 http://www.digitalmars.com/d/2.0/changelog.html
 http://ftp.digitalmars.com/dmd.2.040.zip

 Thanks to the many people who contributed to this update!


Instead of complaining (as usual) I would like to say "Thank You" for 
opening the complete D source code.
It is pretty clear that since D is open sourced we got much more fixes, 
and, not to forget, another level of ideas regarding future development 
tasks.

well done. Bjoern

Feb 05 2010

grauzone <none example.net> writes:

BLS wrote:
 On 30/01/2010 08:13, Walter Bright wrote:
 http://www.digitalmars.com/d/1.0/changelog.html
 http://ftp.digitalmars.com/dmd.1.056.zip


 http://www.digitalmars.com/d/2.0/changelog.html
 http://ftp.digitalmars.com/dmd.2.040.zip

 Thanks to the many people who contributed to this update!


 Instead of complaining (as usual) I would like to say "Thank You" for 
 opening the complete D source code.


Don't forget thanking Walter for using a public version controlled 
repository.

 It is pretty clear that since D is open sourced we got much more fixes, 
 and, not to forget, another level of ideas regarding future development 
 tasks.
 
 well done. Bjoern

Feb 05 2010

Trass3r <un known.com> writes:

 Don't forget thanking Walter for using a public version controlled  
 repository.


Hell yeah, thank god

Feb 05 2010