www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - emplace, scope, enforce [Was: Re: Manual...]

reply bearophile <bearophileHUGS lycos.com> writes:
Andrei Alexandrescu:

 emplace(), defined in std.conv, is relatively new. I haven't yet added
 emplace() for class objects, and this is as good an opportunity as any:
 http://www.dsource.org/projects/phobos/changeset/1752

Thank you, I have used this, and later I have done few tests too. The "scope" for class instantiations can be deprecated once there is an acceptable alternative. You can't deprecate features before you have found a good enough alternative. --------------------- A first problem is the syntax, to allocate an object on the stack you need something like: // is testbuf correctly aligned? ubyte[__traits(classInstanceSize, Test)] testbuf = void; Test t = emplace!(Test)(cast(void[])testbuf, arg1, arg2); That is too much worse looking, hairy and error prone than: scope Test t = new Test(arg1, arg2); I have tried to build a helper to improve the situation, like something that looks: Test t = StackAlloc!(Test, arg1, arg2); But failing that, my second try was this, not good enough: mixin(stackAlloc!(Test, Test)("t", "arg1, arg2")); --------------------- A second problem is that this program compiles with no errors: import std.conv: emplace; final class Test { int x, y; this(int xx, int yy) { this.x = xx; this.y = yy; } } Test foo(int x, int y) { ubyte[__traits(classInstanceSize, Test)] testbuf = void; Test t = emplace!(Test)(cast(void[])testbuf, x, y); return t; } void main() { foo(1, 2); } While the following one gives: test.d(13): Error: escaping reference to scope local t import std.conv: emplace; final class Test { int x, y; this(int xx, int yy) { this.x = xx; this.y = yy; } } Test foo(int x, int y) { scope t = new Test(x, y); return t; } void main() { foo(1, 2); } So the compiler is aware that the scoped object can't escape, while using emplace things become more bug-prone. "scope" can cause other bugs, time ago I have filed a bug report about one problem, but it avoids the most common bug. (I am not sure the emplace solves that problem with scope, I think it shares the same problem, plus adds new ones). --------------------- A third problem is that the ctor doesn't get called: import std.conv: emplace; import std.c.stdio: puts; final class Test { this() { } ~this() { puts("killed"); } } void main() { ubyte[__traits(classInstanceSize, Test)] testbuf = void; Test t = emplace!(Test)(cast(void[])testbuf); } That prints nothing. Using scope it gets called (even if it's not present!). --------------------- This is not a problem of emplace(), it's a problem of the dmd optimizer. I have done few tests for the performance too. I have used this basic pseudocode: while (i < Max) { create testObject(i, i, i, i, i, i) testObject.doSomething(i, i, i, i, i, i) testObject.doSomething(i, i, i, i, i, i) testObject.doSomething(i, i, i, i, i, i) testObject.doSomething(i, i, i, i, i, i) destroy testObject i++ } Coming from here: http://www.drdobbs.com/java/184401976 And its old timings: http://www.ddj.com/java/184401976?pgno=9 The Java version of the code is simple: final class Obj { int i1, i2, i3, i4, i5, i6; Obj(int ii1, int ii2, int ii3, int ii4, int ii5, int ii6) { this.i1 = ii1; this.i2 = ii2; this.i3 = ii3; this.i4 = ii4; this.i5 = ii5; this.i6 = ii6; } void doSomething(int ii1, int ii2, int ii3, int ii4, int ii5, int ii6) { } } class Test { public static void main(String args[]) { final int N = 100_000_000; int i = 0; while (i < N) { Obj testObject = new Obj(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); // testObject = null; // makes no difference i++; } } } This is a D version that uses emplace() (if you don't use emplace here the performance of the D code is very bad compared to the Java one): // program #1 import std.conv: emplace; final class Test { // 32 bytes each instance int i1, i2, i3, i4, i5, i6; this(int ii1, int ii2, int ii3, int ii4, int ii5, int ii6) { this.i1 = ii1; this.i2 = ii2; this.i3 = ii3; this.i4 = ii4; this.i5 = ii5; this.i6 = ii6; } void doSomething(int ii1, int ii2, int ii3, int ii4, int ii5, int ii6) { } } void main() { enum int N = 100_000_000; int i; while (i < N) { ubyte[__traits(classInstanceSize, Test)] buf = void; Test testObject = emplace!(Test)(cast(void[])buf, i, i, i, i, i, i); // Test testObject = new Test(i, i, i, i, i, i); // scope Test testObject = new Test(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); testObject = null; i++; } } The Java code (server) runs in about 0.25 seconds here. The D code (that doesn't do heap allocations at all) run in about 3.60 seconds. With a bit of experiments I have seen that emplace() doesn't get inlined, and the cause is it contains enforce(). enforce contains a throw, and it seems dmd doesn't inline functions that can throw, you can test it with a little test program like this: import std.c.stdlib: atoi; void foo(int b) { if (b) throw new Throwable(null); } void main() { int b = atoi("0"); foo(b); } So if you comment out the two enforce() inside emplace() dmd inlines emplace() and the running time becomes about 2.30 seconds, less than ten times slower than Java. If emplace() doesn't contain calls to enforce() then the loop in main() becomes (dmd 2.047, optmized build): L1A: push dword ptr 02Ch[ESP] mov EDX,_D10test6_good4Test7__ClassZ[0Ch] mov EAX,_D10test6_good4Test7__ClassZ[08h] push EDX push ESI call near ptr _memcpy mov ECX,03Ch[ESP] mov 8[ECX],EBX mov 0Ch[ECX],EBX mov 010h[ECX],EBX mov 014h[ECX],EBX mov 018h[ECX],EBX mov 01Ch[ECX],EBX inc EBX add ESP,0Ch cmp EBX,05F5E100h jb L1A (The memcpy is done by emplace to initialize the object before calling its ctor. You must perform the initialization because it needs the pointer to the virtual table and monitor. The monitor here was null. I think a future LDC2 can optimize away more stuff in that loop, so it's not so bad). If you use this in program #1: scope Test testObject = new Test(i, i, i, i, i, i); It runs in about 6 seconds (also because the ctor is called even if's missing). If in program #1 you use just new, without scope, the runtime is about 27.2 seconds, about 110 times slower than Java. Bye, bearophile
Jul 20 2010
next sibling parent reply "Rory McGuire" <rmcguire neonova.co.za> writes:
On Wed, 21 Jul 2010 03:58:33 +0200, bearophile <bearophileHUGS lycos.com>  
wrote:

 Andrei Alexandrescu:

 emplace(), defined in std.conv, is relatively new. I haven't yet added
 emplace() for class objects, and this is as good an opportunity as any:
 http://www.dsource.org/projects/phobos/changeset/1752

Thank you, I have used this, and later I have done few tests too. The "scope" for class instantiations can be deprecated once there is an acceptable alternative. You can't deprecate features before you have found a good enough alternative. --------------------- A first problem is the syntax, to allocate an object on the stack you need something like: // is testbuf correctly aligned? ubyte[__traits(classInstanceSize, Test)] testbuf = void; Test t = emplace!(Test)(cast(void[])testbuf, arg1, arg2); That is too much worse looking, hairy and error prone than: scope Test t = new Test(arg1, arg2); I have tried to build a helper to improve the situation, like something that looks: Test t = StackAlloc!(Test, arg1, arg2); But failing that, my second try was this, not good enough: mixin(stackAlloc!(Test, Test)("t", "arg1, arg2")); --------------------- A second problem is that this program compiles with no errors: import std.conv: emplace; final class Test { int x, y; this(int xx, int yy) { this.x = xx; this.y = yy; } } Test foo(int x, int y) { ubyte[__traits(classInstanceSize, Test)] testbuf = void; Test t = emplace!(Test)(cast(void[])testbuf, x, y); return t; } void main() { foo(1, 2); } While the following one gives: test.d(13): Error: escaping reference to scope local t import std.conv: emplace; final class Test { int x, y; this(int xx, int yy) { this.x = xx; this.y = yy; } } Test foo(int x, int y) { scope t = new Test(x, y); return t; } void main() { foo(1, 2); } So the compiler is aware that the scoped object can't escape, while using emplace things become more bug-prone. "scope" can cause other bugs, time ago I have filed a bug report about one problem, but it avoids the most common bug. (I am not sure the emplace solves that problem with scope, I think it shares the same problem, plus adds new ones). --------------------- A third problem is that the ctor doesn't get called: import std.conv: emplace; import std.c.stdio: puts; final class Test { this() { } ~this() { puts("killed"); } } void main() { ubyte[__traits(classInstanceSize, Test)] testbuf = void; Test t = emplace!(Test)(cast(void[])testbuf); } That prints nothing. Using scope it gets called (even if it's not present!). --------------------- This is not a problem of emplace(), it's a problem of the dmd optimizer. I have done few tests for the performance too. I have used this basic pseudocode: while (i < Max) { create testObject(i, i, i, i, i, i) testObject.doSomething(i, i, i, i, i, i) testObject.doSomething(i, i, i, i, i, i) testObject.doSomething(i, i, i, i, i, i) testObject.doSomething(i, i, i, i, i, i) destroy testObject i++ } Coming from here: http://www.drdobbs.com/java/184401976 And its old timings: http://www.ddj.com/java/184401976?pgno=9 The Java version of the code is simple: final class Obj { int i1, i2, i3, i4, i5, i6; Obj(int ii1, int ii2, int ii3, int ii4, int ii5, int ii6) { this.i1 = ii1; this.i2 = ii2; this.i3 = ii3; this.i4 = ii4; this.i5 = ii5; this.i6 = ii6; } void doSomething(int ii1, int ii2, int ii3, int ii4, int ii5, int ii6) { } } class Test { public static void main(String args[]) { final int N = 100_000_000; int i = 0; while (i < N) { Obj testObject = new Obj(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); // testObject = null; // makes no difference i++; } } } This is a D version that uses emplace() (if you don't use emplace here the performance of the D code is very bad compared to the Java one): // program #1 import std.conv: emplace; final class Test { // 32 bytes each instance int i1, i2, i3, i4, i5, i6; this(int ii1, int ii2, int ii3, int ii4, int ii5, int ii6) { this.i1 = ii1; this.i2 = ii2; this.i3 = ii3; this.i4 = ii4; this.i5 = ii5; this.i6 = ii6; } void doSomething(int ii1, int ii2, int ii3, int ii4, int ii5, int ii6) { } } void main() { enum int N = 100_000_000; int i; while (i < N) { ubyte[__traits(classInstanceSize, Test)] buf = void; Test testObject = emplace!(Test)(cast(void[])buf, i, i, i, i, i, i); // Test testObject = new Test(i, i, i, i, i, i); // scope Test testObject = new Test(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); testObject = null; i++; } } The Java code (server) runs in about 0.25 seconds here. The D code (that doesn't do heap allocations at all) run in about 3.60 seconds. With a bit of experiments I have seen that emplace() doesn't get inlined, and the cause is it contains enforce(). enforce contains a throw, and it seems dmd doesn't inline functions that can throw, you can test it with a little test program like this: import std.c.stdlib: atoi; void foo(int b) { if (b) throw new Throwable(null); } void main() { int b = atoi("0"); foo(b); } So if you comment out the two enforce() inside emplace() dmd inlines emplace() and the running time becomes about 2.30 seconds, less than ten times slower than Java. If emplace() doesn't contain calls to enforce() then the loop in main() becomes (dmd 2.047, optmized build): L1A: push dword ptr 02Ch[ESP] mov EDX,_D10test6_good4Test7__ClassZ[0Ch] mov EAX,_D10test6_good4Test7__ClassZ[08h] push EDX push ESI call near ptr _memcpy mov ECX,03Ch[ESP] mov 8[ECX],EBX mov 0Ch[ECX],EBX mov 010h[ECX],EBX mov 014h[ECX],EBX mov 018h[ECX],EBX mov 01Ch[ECX],EBX inc EBX add ESP,0Ch cmp EBX,05F5E100h jb L1A (The memcpy is done by emplace to initialize the object before calling its ctor. You must perform the initialization because it needs the pointer to the virtual table and monitor. The monitor here was null. I think a future LDC2 can optimize away more stuff in that loop, so it's not so bad). If you use this in program #1: scope Test testObject = new Test(i, i, i, i, i, i); It runs in about 6 seconds (also because the ctor is called even if's missing). If in program #1 you use just new, without scope, the runtime is about 27.2 seconds, about 110 times slower than Java. Bye, bearophile

Takes 18m27.720s in PHP :)
Jul 21 2010
next sibling parent reply Rory Mcguire <rjmcguire gm_no_ail.com> writes:
Rory McGuire wrote:

 On Wed, 21 Jul 2010 03:58:33 +0200, bearophile 

 wrote:
 
 Andrei Alexandrescu:

 emplace(), defined in std.conv, is relatively new. I haven't yet 



 emplace() for class objects, and this is as good an opportunity as 



 http://www.dsource.org/projects/phobos/changeset/1752

Thank you, I have used this, and later I have done few tests too. The "scope" for class instantiations can be deprecated once there


 acceptable alternative. You can't deprecate features before you 


 found a good enough alternative.

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

 A first problem is the syntax, to allocate an object on the stack 


 need something like:

 // is testbuf correctly aligned?
 ubyte[__traits(classInstanceSize, Test)] testbuf = void;
 Test t = emplace!(Test)(cast(void[])testbuf, arg1, arg2);


 That is too much worse looking, hairy and error prone than:
 scope Test t = new Test(arg1, arg2);


 I have tried to build a helper to improve the situation, like 


 that looks:
 Test t = StackAlloc!(Test, arg1, arg2);

 But failing that, my second try was this, not good enough:
 mixin(stackAlloc!(Test, Test)("t", "arg1, arg2"));

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

 A second problem is that this program compiles with no errors:

 import std.conv: emplace;

 final class Test {
     int x, y;
     this(int xx, int yy) {
         this.x = xx;
         this.y = yy;
     }
 }

 Test foo(int x, int y) {
     ubyte[__traits(classInstanceSize, Test)] testbuf = void;
     Test t = emplace!(Test)(cast(void[])testbuf, x, y);
     return t;
 }

 void main() {
     foo(1, 2);
 }



 While the following one gives:
 test.d(13): Error: escaping reference to scope local t


 import std.conv: emplace;

 final class Test {
     int x, y;
     this(int xx, int yy) {
         this.x = xx;
         this.y = yy;
     }
 }

 Test foo(int x, int y) {
     scope t = new Test(x, y);
     return t;
 }

 void main() {
     foo(1, 2);
 }


 So the compiler is aware that the scoped object can't escape, while
 using emplace things become more bug-prone. "scope" can cause other
 bugs, time ago I have filed a bug report about one problem, but it
 avoids the most common bug. (I am not sure the emplace solves that
 problem with scope, I think it shares the same problem, plus adds 


 ones).

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

 A third problem is that the ctor doesn't get called:


 import std.conv: emplace;
 import std.c.stdio: puts;

 final class Test {
     this() {
     }
     ~this() { puts("killed"); }
 }

 void main() {
     ubyte[__traits(classInstanceSize, Test)] testbuf = void;
     Test t = emplace!(Test)(cast(void[])testbuf);
 }


 That prints nothing. Using scope it gets called (even if it's not
 present!).

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

 This is not a problem of emplace(), it's a problem of the dmd 


 I have done few tests for the performance too. I have used this 


 pseudocode:

 while (i < Max)
 {
    create testObject(i, i, i, i, i, i)
    testObject.doSomething(i, i, i, i, i, i)
    testObject.doSomething(i, i, i, i, i, i)
    testObject.doSomething(i, i, i, i, i, i)
    testObject.doSomething(i, i, i, i, i, i)
    destroy testObject
    i++
 }


 Coming from here:
 http://www.drdobbs.com/java/184401976
 And its old timings:
 http://www.ddj.com/java/184401976?pgno=9


 The Java version of the code is simple:

 final class Obj {
     int i1, i2, i3, i4, i5, i6;

     Obj(int ii1, int ii2, int ii3, int ii4, int ii5, int ii6) {
         this.i1 = ii1;
         this.i2 = ii2;
         this.i3 = ii3;
         this.i4 = ii4;
         this.i5 = ii5;
         this.i6 = ii6;
     }

     void doSomething(int ii1, int ii2, int ii3, int ii4, int ii5, 


 ii6) {
     }
 }

 class Test {
     public static void main(String args[]) {
         final int N = 100_000_000;
         int i = 0;
         while (i < N) {
             Obj testObject = new Obj(i, i, i, i, i, i);
             testObject.doSomething(i, i, i, i, i, i);
             testObject.doSomething(i, i, i, i, i, i);
             testObject.doSomething(i, i, i, i, i, i);
             testObject.doSomething(i, i, i, i, i, i);
             // testObject = null; // makes no difference
             i++;
         }
     }
 }



 This is a D version that uses emplace() (if you don't use emplace 


 the performance of the D code is very bad compared to the Java 


 // program #1
 import std.conv: emplace;

 final class Test { // 32 bytes each instance
     int i1, i2, i3, i4, i5, i6;
     this(int ii1, int ii2, int ii3, int ii4, int ii5, int ii6) {
         this.i1 = ii1;
         this.i2 = ii2;
         this.i3 = ii3;
         this.i4 = ii4;
         this.i5 = ii5;
         this.i6 = ii6;
     }
     void doSomething(int ii1, int ii2, int ii3, int ii4, int ii5, 


 ii6) {
     }
 }

 void main() {
     enum int N = 100_000_000;

     int i;
     while (i < N) {
         ubyte[__traits(classInstanceSize, Test)] buf = void;
         Test testObject = emplace!(Test)(cast(void[])buf, i, i, i, 


 i);
         // Test testObject = new Test(i, i, i, i, i, i);
         // scope Test testObject = new Test(i, i, i, i, i, i);
         testObject.doSomething(i, i, i, i, i, i);
         testObject.doSomething(i, i, i, i, i, i);
         testObject.doSomething(i, i, i, i, i, i);
         testObject.doSomething(i, i, i, i, i, i);
         testObject = null;
         i++;
     }
 }


 The Java code (server) runs in about 0.25 seconds here.
 The D code (that doesn't do heap allocations at all) run in about 


 seconds.

 With a bit of experiments I have seen that emplace() doesn't get
 inlined, and the cause is it contains enforce(). enforce contains a
 throw, and it seems dmd doesn't inline functions that can throw, 


 test it with a little test program like this:


 import std.c.stdlib: atoi;
 void foo(int b) {
     if (b)
         throw new Throwable(null);
 }
 void main() {
     int b = atoi("0");
     foo(b);
 }


 So if you comment out the two enforce() inside emplace() dmd 


 emplace() and the running time becomes about 2.30 seconds, less 


 times slower than Java.

 If emplace() doesn't contain calls to enforce() then the loop in 


 becomes (dmd 2.047, optmized build):


 L1A:		push	dword ptr 02Ch[ESP]
 mov	EDX,_D10test6_good4Test7__ClassZ[0Ch]
 mov	EAX,_D10test6_good4Test7__ClassZ[08h]
 push	EDX
 push	ESI
 call	near ptr _memcpy
 mov	ECX,03Ch[ESP]
 mov	8[ECX],EBX
 mov	0Ch[ECX],EBX
 mov	010h[ECX],EBX
 mov	014h[ECX],EBX
 mov	018h[ECX],EBX
 mov	01Ch[ECX],EBX
 inc	EBX
 add	ESP,0Ch
 cmp	EBX,05F5E100h
 jb	L1A


 (The memcpy is done by emplace to initialize the object before 


 its ctor. You must perform the initialization because it needs the
 pointer to the virtual table and monitor. The monitor here was 


 think a future LDC2 can optimize away more stuff in that loop, so 


 not so bad).


 If you use this in program #1:
 scope Test testObject = new Test(i, i, i, i, i, i);
 It runs in about 6 seconds (also because the ctor is called even 


 missing).

 If in program #1 you use just new, without scope, the runtime is 


 27.2 seconds, about 110 times slower than Java.

 Bye,
 bearophile

Takes 18m27.720s in PHP :)

Takes 5m26.776s in Python. Takes 0m1.008s in Java. can't test D version I don't have emplace and dsource is ignoring me.
Jul 21 2010
parent reply bearophile <bearophileHUGS lycos.com> writes:
Rory Mcguire:
 Takes 5m26.776s in Python.
 Takes 0m1.008s in Java.

(I suggest you to round away milliseconds, they are never significant in such benchmarks.) Python 2.7 uses its GC a bit better, so it can be a bit faster. Your Java code has run four times slower than my slow PC, that's a lot. In Java have you used the -server switch? Bye, bearophile
Jul 21 2010
parent reply Rory Mcguire <rjmcguire gm_no_ail.com> writes:
bearophile wrote:

 Rory Mcguire:
 Takes 5m26.776s in Python.
 Takes 0m1.008s in Java.

(I suggest you to round away milliseconds, they are never significant in such benchmarks.) Python 2.7 uses its GC a bit better, so it can be a bit faster. Your Java code has run four times slower than my slow PC, that's a lot. In Java have you used the -server switch? Bye, bearophile

On ubuntu 10.04 64 I'm using `time` to get the timing. I wan't using -server, with it I get 0m1.047s. D version gets 0m8.162s using a 32bit chroot environment. Processor is a core i7 (1.6Ghz * 8). 6GB ram. Interesting thing about the python one is it used 3GB of ram most of the time.
Jul 21 2010
parent Rory Mcguire <rjmcguire gm_no_ail.com> writes:
Rory Mcguire wrote:

 bearophile wrote:
 
 Rory Mcguire:
 Takes 5m26.776s in Python.
 Takes 0m1.008s in Java.

(I suggest you to round away milliseconds, they are never significant in such benchmarks.) Python 2.7 uses its GC a bit better, so it can be a bit faster. Your Java code has run four times slower than my slow PC, that's a lot. In Java have you used the -server switch? Bye, bearophile

On ubuntu 10.04 64 I'm using `time` to get the timing. I wan't using -server, with it I get 0m1.047s. D version gets 0m8.162s using a 32bit chroot environment. Processor is a core i7 (1.6Ghz * 8). 6GB ram. Interesting thing about the python one is it used 3GB of ram most of the time.

Perhaps the slow times are because I'm reporting the real timings not the user/sys time.
Jul 21 2010
prev sibling parent bearophile <bearophileHUGS lycos.com> writes:
Rory McGuire:
 Takes 18m27.720s in PHP :)

You have lot of patience :-)
 can't test D version I don't have emplace and dsource is ignoring me.

This was Andrei's code before dsource went down (ddoc and unittest removed): T emplace(T, Args...)(void[] chunk, Args args) if (is(T == class)) { enforce(chunk.length >= __traits(classInstanceSize, T)); auto a = cast(size_t) chunk.ptr; enforce(a % real.alignof == 0); auto result = cast(typeof(return)) chunk.ptr; // Initialize the object in its pre-ctor state (cast(byte[]) chunk)[] = typeid(T).init[]; // Call the ctor if any static if (is(typeof(result.__ctor(args)))) { // T defines a genuine constructor accepting args // Go the classic route: write .init first, then call ctor result.__ctor(args); } else { static assert(args.length == 0 && !is(typeof(&T.__ctor)), "Don't know how to initialize an object of type " ~ T.stringof ~ " with arguments " ~ Args.stringof); } return result; } Bye, bearophile
Jul 21 2010
prev sibling parent reply Dmitry Olshansky <dmitry.olsh gmail.com> writes:
On 21.07.2010 5:58, bearophile wrote:
 Andrei Alexandrescu:

    
 emplace(), defined in std.conv, is relatively new. I haven't yet added
 emplace() for class objects, and this is as good an opportunity as any:
 http://www.dsource.org/projects/phobos/changeset/1752
      

The "scope" for class instantiations can be deprecated once there is an acceptable alternative. You can't deprecate features before you have found a good enough alternative. --------------------- A first problem is the syntax, to allocate an object on the stack you need something like: // is testbuf correctly aligned? ubyte[__traits(classInstanceSize, Test)] testbuf = void; Test t = emplace!(Test)(cast(void[])testbuf, arg1, arg2); That is too much worse looking, hairy and error prone than: scope Test t = new Test(arg1, arg2); I have tried to build a helper to improve the situation, like something that looks: Test t = StackAlloc!(Test, arg1, arg2);

T* create(T, Args...)(Args args) if ( !is(T == class) ){ return emplace!T(malloc(T.sizeof)[0..T.sizeof], args); } void destroy(T)(T* ptr) if ( !is(T == class) ){ assert(ptr); clear(ptr); free(ptr); } //then auto a = create!T(params); I guess one could easily patch it for classes.
 But failing that, my second try was this, not good enough:
 mixin(stackAlloc!(Test, Test)("t", "arg1, arg2"));

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

 A second problem is that this program compiles with no errors:

 import std.conv: emplace;

 final class Test {
      int x, y;
      this(int xx, int yy) {
          this.x = xx;
          this.y = yy;
      }
 }

 Test foo(int x, int y) {
      ubyte[__traits(classInstanceSize, Test)] testbuf = void;
      Test t = emplace!(Test)(cast(void[])testbuf, x, y);
      return t;
 }

 void main() {
      foo(1, 2);
 }
    

about custom allocation, not scoped variables.
 While the following one gives:
 test.d(13): Error: escaping reference to scope local t


 import std.conv: emplace;

 final class Test {
      int x, y;
      this(int xx, int yy) {
          this.x = xx;
          this.y = yy;
      }
 }

 Test foo(int x, int y) {
      scope t = new Test(x, y);
      return t;
 }

 void main() {
      foo(1, 2);
 }


 So the compiler is aware that the scoped object can't escape, while using
emplace things become more bug-prone. "scope" can cause other bugs, time ago I
have filed a bug report about one problem, but it avoids the most common bug.
(I am not sure the emplace solves that problem with scope, I think it shares
the same problem, plus adds new ones).

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

 A third problem is that the ctor doesn't get called:


 import std.conv: emplace;
 import std.c.stdio: puts;

 final class Test {
      this() {
      }
      ~this() { puts("killed"); }
 }

 void main() {
      ubyte[__traits(classInstanceSize, Test)] testbuf = void;
      Test t = emplace!(Test)(cast(void[])testbuf);
 }
    

replacement for placement new( no pun). Sure enough with manual memory management you need to call clear(t) at exit.
 That prints nothing. Using scope it gets called (even if it's not present!).

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

 This is not a problem of emplace(), it's a problem of the dmd optimizer.
 I have done few tests for the performance too. I have used this basic
pseudocode:

 while (i<  Max)
 {
     create testObject(i, i, i, i, i, i)
     testObject.doSomething(i, i, i, i, i, i)
     testObject.doSomething(i, i, i, i, i, i)
     testObject.doSomething(i, i, i, i, i, i)
     testObject.doSomething(i, i, i, i, i, i)
     destroy testObject
     i++
 }


 Coming from here:
 http://www.drdobbs.com/java/184401976
 And its old timings:
 http://www.ddj.com/java/184401976?pgno=9


 The Java version of the code is simple:

 final class Obj {
      int i1, i2, i3, i4, i5, i6;

      Obj(int ii1, int ii2, int ii3, int ii4, int ii5, int ii6) {
          this.i1 = ii1;
          this.i2 = ii2;
          this.i3 = ii3;
          this.i4 = ii4;
          this.i5 = ii5;
          this.i6 = ii6;
      }

      void doSomething(int ii1, int ii2, int ii3, int ii4, int ii5, int ii6) {
      }
 }

 class Test {
      public static void main(String args[]) {
          final int N = 100_000_000;
          int i = 0;
          while (i<  N) {
              Obj testObject = new Obj(i, i, i, i, i, i);
              testObject.doSomething(i, i, i, i, i, i);
              testObject.doSomething(i, i, i, i, i, i);
              testObject.doSomething(i, i, i, i, i, i);
              testObject.doSomething(i, i, i, i, i, i);
              // testObject = null; // makes no difference
              i++;
          }
      }
 }



 This is a D version that uses emplace() (if you don't use emplace here the
performance of the D code is very bad compared to the Java one):

 // program #1
 import std.conv: emplace;

 final class Test { // 32 bytes each instance
      int i1, i2, i3, i4, i5, i6;
      this(int ii1, int ii2, int ii3, int ii4, int ii5, int ii6) {
          this.i1 = ii1;
          this.i2 = ii2;
          this.i3 = ii3;
          this.i4 = ii4;
          this.i5 = ii5;
          this.i6 = ii6;
      }
      void doSomething(int ii1, int ii2, int ii3, int ii4, int ii5, int ii6) {
      }
 }

 void main() {
      enum int N = 100_000_000;

      int i;
      while (i<  N) {
          ubyte[__traits(classInstanceSize, Test)] buf = void;
          Test testObject = emplace!(Test)(cast(void[])buf, i, i, i, i, i, i);
          // Test testObject = new Test(i, i, i, i, i, i);
          // scope Test testObject = new Test(i, i, i, i, i, i);
          testObject.doSomething(i, i, i, i, i, i);
          testObject.doSomething(i, i, i, i, i, i);
          testObject.doSomething(i, i, i, i, i, i);
          testObject.doSomething(i, i, i, i, i, i);
          testObject = null;
          i++;
      }
 }


 The Java code (server) runs in about 0.25 seconds here.
 The D code (that doesn't do heap allocations at all) run in about 3.60 seconds.

 With a bit of experiments I have seen that emplace() doesn't get inlined, and
the cause is it contains enforce(). enforce contains a throw, and it seems dmd
doesn't inline functions that can throw, you can test it with a little test
program like this:


 import std.c.stdlib: atoi;
 void foo(int b) {
      if (b)
          throw new Throwable(null);
 }
 void main() {
      int b = atoi("0");
      foo(b);
 }


 So if you comment out the two enforce() inside emplace() dmd inlines emplace()
and the running time becomes about 2.30 seconds, less than ten times slower
than Java.

 If emplace() doesn't contain calls to enforce() then the loop in main()
becomes (dmd 2.047, optmized build):


 L1A:		push	dword ptr 02Ch[ESP]
 		mov	EDX,_D10test6_good4Test7__ClassZ[0Ch]
 		mov	EAX,_D10test6_good4Test7__ClassZ[08h]
 		push	EDX
 		push	ESI
 		call	near ptr _memcpy
 		mov	ECX,03Ch[ESP]
 		mov	8[ECX],EBX
 		mov	0Ch[ECX],EBX
 		mov	010h[ECX],EBX
 		mov	014h[ECX],EBX
 		mov	018h[ECX],EBX
 		mov	01Ch[ECX],EBX
 		inc	EBX
 		add	ESP,0Ch
 		cmp	EBX,05F5E100h
 		jb	L1A


 (The memcpy is done by emplace to initialize the object before calling its
ctor. You must perform the initialization because it needs the pointer to the
virtual table and monitor. The monitor here was null. I think a future LDC2 can
optimize away more stuff in that loop, so it's not so bad).


 If you use this in program #1:
 scope Test testObject = new Test(i, i, i, i, i, i);
 It runs in about 6 seconds (also because the ctor is called even if's missing).

 If in program #1 you use just new, without scope, the runtime is about 27.2
seconds, about 110 times slower than Java.

 Bye,
 bearophile
    

-- Dmitry Olshansky
Jul 21 2010
next sibling parent reply bearophile <bearophileHUGS lycos.com> writes:
Dmitry Olshansky:
 Well, I'm using this for structs, very straightforward:
 
 T* create(T, Args...)(Args args)
 if ( !is(T == class) ){
      return emplace!T(malloc(T.sizeof)[0..T.sizeof], args);
 }

That's not good enough, you are allocating on the (C) heap. If you use that in the D benchmark I have shown you probably can get bad timing results.
 This is dtor not get called, and it's because emplace is a library 
 replacement for placement new( no pun).
 Sure enough with manual memory management you need to call clear(t) at exit.

If the class is allocated on the stack it's much better if the destructor is called when the class gets out of scope. Otherwise it's like C programming. (I suggest to edit your post, to remove useless parts of the original post.) Bye, bearophile
Jul 21 2010
parent reply Dmitry Olshansky <dmitry.olsh gmail.com> writes:
On 21.07.2010 14:20, bearophile wrote:
 Dmitry Olshansky:
    
 Well, I'm using this for structs, very straightforward:

 T* create(T, Args...)(Args args)
 if ( !is(T == class) ){
       return emplace!T(malloc(T.sizeof)[0..T.sizeof], args);
 }
      


allocation is risky business to say least. Some kind of memory pool should be handy.
 This is dtor not get called, and it's because emplace is a library
 replacement for placement new( no pun).
 Sure enough with manual memory management you need to call clear(t) at exit.
      

(I suggest to edit your post, to remove useless parts of the original post.)

classes and scope. That's, of course, losing the inheritance and such. The problem is designing such classes and then documenting: "you should always use it as 'scope' ", is awkward. Moreover, the function which you pass the stack allocated class instance are unaware of that clever trick. Slight modification of your benchmark: import std.conv: emplace; import std.contracts; import std.stdio; final class Test { // 32 bytes each instance int i1, i2, i3, i4, i5, i6; this(int ii1, int ii2, int ii3, int ii4, int ii5, int ii6) { this.i1 = ii1; this.i2 = ii2; this.i3 = ii3; this.i4 = ii4; this.i5 = ii5; this.i6 = ii6; } void doSomething(int ii1, int ii2, int ii3, int ii4, int ii5, int ii6) { } } Test hidden; void fun(Test t){ hidden = t; } void bench(){ enum int N = 10_000_000; int i; while (i < N) { scope Test testObject = new Test(i, i, i, i, i, i); fun(testObject); testObject.doSomething(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); i++; } } void main() { int a,b,c;// bench(); //what's the hidden now? writefln("%d %d", hidden.i1,hidden.i2); writefln("%d %d", hidden.i1,hidden.i2); } The second writefln prints garbage. I guess it's because of pointer to the long gone stackframe, which is ovewritten by the first writeln. -- Dmitry Olshansky
Jul 21 2010
parent reply bearophile <bearophileHUGS lycos.com> writes:
Dmitry Olshansky:
 The problem is designing such classes and then documenting: "you should 
 always use it as 'scope' ", is awkward.

If you really want a class to be used as scope only you can do this, see the error message: scope class Foo {} void main() { Foo f = new Foo; }
 The second writefln prints garbage. I guess it's because of pointer to 
 the long gone stackframe, which is ovewritten by the first writeln.

Yes scope has this and other problems (and I think two of them can be fixed), but I don't think emplace() is a big improvement. Bye, bearophile
Jul 21 2010
parent reply Dmitry Olshansky <dmitry.olsh gmail.com> writes:
On 21.07.2010 16:26, bearophile wrote:
 Dmitry Olshansky:
    
 The problem is designing such classes and then documenting: "you should
 always use it as 'scope' ", is awkward.
      

scope class Foo {} void main() { Foo f = new Foo; }
 The second writefln prints garbage. I guess it's because of pointer to
 the long gone stackframe, which is ovewritten by the first writeln.
      

Bye, bearophile

feature, I made a (somewhat) successful try at implementing scoped classes: struct Scoped(T){ ubyte[__traits(classInstanceSize, Test)] _payload; T getPayload(){ return cast(T)(_payload.ptr); } alias getPayload this; static Scoped opCall(Args...)(Args args) if ( is(typeof(T.init.__ctor(args))) ){// TODO: should also provide decent error message Scoped!T s; emplace!T(cast(void[])s._payload,args); return s; } ~this(){ clear(getPayload); } } now replace the orignal while loop with this: while (i < N) { auto testObject = Scoped!Test(i, i, i, i, i, i); //assuming we have aforementioned evil function func(Test t), that keeps global reference to t. //fun(testObject); //uncoment to get an compile error - type mismatch testObject.doSomething(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); i++; } and all works just the same as with deprecated scope storage class. Even better it disallows passing the variable to functions expecting vanilla Test, it's limiting but for a good reason. There are still issues that should be solved (name clash for one, plus the ability to define default construct Scoped!T) but overall it's OK to me. -- Dmitry Olshansky
Jul 21 2010
next sibling parent Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
Dmitry Olshansky wrote:
 On 21.07.2010 16:26, bearophile wrote:
 Dmitry Olshansky:
   
 The problem is designing such classes and then documenting: "you should
 always use it as 'scope' ", is awkward.
      

see the error message: scope class Foo {} void main() { Foo f = new Foo; }
 The second writefln prints garbage. I guess it's because of pointer to
 the long gone stackframe, which is ovewritten by the first writeln.
      

fixed), but I don't think emplace() is a big improvement. Bye, bearophile

feature, I made a (somewhat) successful try at implementing scoped classes: struct Scoped(T){ ubyte[__traits(classInstanceSize, Test)] _payload; T getPayload(){ return cast(T)(_payload.ptr); } alias getPayload this; static Scoped opCall(Args...)(Args args) if ( is(typeof(T.init.__ctor(args))) ){// TODO: should also provide decent error message Scoped!T s; emplace!T(cast(void[])s._payload,args); return s; } ~this(){ clear(getPayload); } } now replace the orignal while loop with this: while (i < N) { auto testObject = Scoped!Test(i, i, i, i, i, i); //assuming we have aforementioned evil function func(Test t), that keeps global reference to t. //fun(testObject); //uncoment to get an compile error - type mismatch testObject.doSomething(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); i++; } and all works just the same as with deprecated scope storage class. Even better it disallows passing the variable to functions expecting vanilla Test, it's limiting but for a good reason. There are still issues that should be solved (name clash for one, plus the ability to define default construct Scoped!T) but overall it's OK to me.

Nice work. To avoid name clashes with alias this, you may want to use a trick invented by Shin Fujishiro: struct ExWhyZee { template ExWhyZee() { // implementation goes here } alias ExWhyZee!().whatever this; } This way you never have a confusion between the symbols defined by the wrapped type and your own type. I use a variant of the same trick in RefCounted. Andrei
Jul 21 2010
prev sibling next sibling parent Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
Dmitry Olshansky wrote:
 On 21.07.2010 16:26, bearophile wrote:
 Dmitry Olshansky:
   
 The problem is designing such classes and then documenting: "you should
 always use it as 'scope' ", is awkward.
      

see the error message: scope class Foo {} void main() { Foo f = new Foo; }
 The second writefln prints garbage. I guess it's because of pointer to
 the long gone stackframe, which is ovewritten by the first writeln.
      

fixed), but I don't think emplace() is a big improvement. Bye, bearophile

feature, I made a (somewhat) successful try at implementing scoped classes:

I salute this approach.
 struct Scoped(T){
     ubyte[__traits(classInstanceSize, Test)] _payload;

s/Test/T/ I suppose.
     T getPayload(){
         return cast(T)(_payload.ptr);
     }
     alias getPayload this;
     static Scoped opCall(Args...)(Args args) if ( 
 is(typeof(T.init.__ctor(args))) ){// TODO: should also provide decent 
 error message
         Scoped!T s;
         emplace!T(cast(void[])s._payload,args);
         return s;
     }
     ~this(){
         clear(getPayload);
     }
 }
 
 now replace the orignal while loop with this:
 while (i < N) {
         auto testObject = Scoped!Test(i, i, i, i, i, i);
         //assuming we have aforementioned evil function func(Test t), 
 that keeps global reference to t.
         //fun(testObject); //uncoment to get an compile error - type 
 mismatch
         testObject.doSomething(i, i, i, i, i, i);
         testObject.doSomething(i, i, i, i, i, i);
         testObject.doSomething(i, i, i, i, i, i);
         testObject.doSomething(i, i, i, i, i, i);
         i++;
     }
 
 and all works just the same as with deprecated scope storage class.
 Even better it disallows passing the variable to functions expecting 
 vanilla Test, it's limiting but for a good reason.
 There are still issues that should be solved (name clash for one, plus 
 the ability to define default construct Scoped!T) but overall it's OK to 
 me.

I agree Scope has a rightful place in the standard library. Andrei
Jul 21 2010
prev sibling parent reply Rory Mcguire <rjmcguire gm_no_ail.com> writes:
Dmitry Olshansky wrote:

 now replace the orignal while loop with this:
 while (i < N) {
 auto testObject = Scoped!Test(i, i, i, i, i, i);
 //assuming we have aforementioned evil function func(Test t),
 that keeps global reference to t.
 //fun(testObject); //uncoment to get an compile error - type
 mismatch
 testObject.doSomething(i, i, i, i, i, i);
 testObject.doSomething(i, i, i, i, i, i);
 testObject.doSomething(i, i, i, i, i, i);
 testObject.doSomething(i, i, i, i, i, i);
 i++;
 }

With your code I `time` reports the below timings on my machine: real 0m19.658s user 0m19.590s sys 0m0.010s compared to: real 0m9.122s user 0m9.090s sys 0m0.000s with bearofiles original version. With -O -release its about 4 seconds faster for each.
Jul 21 2010
parent reply Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
On 07/21/2010 01:59 PM, Rory Mcguire wrote:
 Dmitry Olshansky wrote:

 now replace the orignal while loop with this:
 while (i<  N) {
 auto testObject = Scoped!Test(i, i, i, i, i, i);
 //assuming we have aforementioned evil function func(Test t),
 that keeps global reference to t.
 //fun(testObject); //uncoment to get an compile error - type
 mismatch
 testObject.doSomething(i, i, i, i, i, i);
 testObject.doSomething(i, i, i, i, i, i);
 testObject.doSomething(i, i, i, i, i, i);
 testObject.doSomething(i, i, i, i, i, i);
 i++;
 }

With your code I `time` reports the below timings on my machine: real 0m19.658s user 0m19.590s sys 0m0.010s compared to: real 0m9.122s user 0m9.090s sys 0m0.000s with bearofiles original version. With -O -release its about 4 seconds faster for each.

I compiled and ran the tests myself with -O -release -inline and got 1.95s for Dmitry's implementation and 1.55s for bearophile's. I optimized Dmitry's implementation in two ways: I replaced the call to clear() with a straight call to the destructor and added = void in two places to avoid double initialization. I got 1.11s, which significantly undercuts the implementation using scope. Here's the code I used for testing: struct Scoped(T) { ubyte[__traits(classInstanceSize, Test)] _payload = void; T getPayload(){ return cast(T)(_payload.ptr); } alias getPayload this; static Scoped opCall(Args...)(Args args) if (is(typeof(T.init.__ctor(args)))) { // TODO: should also provide decent error message Scoped!T s = void; emplace!T(cast(void[])s._payload,args); return s; } ~this() { static if (is(typeof(getPayload.__dtor()))) { getPayload.__dtor(); } } } final class Test { // 32 bytes each instance int i1, i2, i3, i4, i5, i6; this(int ii1, int ii2, int ii3, int ii4, int ii5, int ii6) { this.i1 = ii1; this.i2 = ii2; this.i3 = ii3; this.i4 = ii4; this.i5 = ii5; this.i6 = ii6; } void doSomething(int ii1, int ii2, int ii3, int ii4, int ii5, int ii6) { } } void main(string[] args) { enum int N = 10_000_000; int i; while (i < N) { auto testObject = Scoped!Test(i, i, i, i, i, i); //scope testObject = new Test(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); i++; } } Andrei
Jul 21 2010
next sibling parent reply Rory Mcguire <rjmcguire gm_no_ail.com> writes:
Andrei Alexandrescu wrote:

 On 07/21/2010 01:59 PM, Rory Mcguire wrote:
 Dmitry Olshansky wrote:

 now replace the orignal while loop with this:
 while (i<  N) {
 auto testObject = Scoped!Test(i, i, i, i, i, i);
 //assuming we have aforementioned evil function func(Test t),
 that keeps global reference to t.
 //fun(testObject); //uncoment to get an compile error - type
 mismatch
 testObject.doSomething(i, i, i, i, i, i);
 testObject.doSomething(i, i, i, i, i, i);
 testObject.doSomething(i, i, i, i, i, i);
 testObject.doSomething(i, i, i, i, i, i);
 i++;
 }

With your code I `time` reports the below timings on my machine: real 0m19.658s user 0m19.590s sys 0m0.010s compared to: real 0m9.122s user 0m9.090s sys 0m0.000s with bearofiles original version. With -O -release its about 4 seconds faster for each.

I compiled and ran the tests myself with -O -release -inline and got 1.95s for Dmitry's implementation and 1.55s for bearophile's. I optimized Dmitry's implementation in two ways: I replaced the call to clear() with a straight call to the destructor and added = void in two places to avoid double initialization. I got 1.11s, which significantly undercuts the implementation using scope. Here's the code I used for testing: struct Scoped(T) { ubyte[__traits(classInstanceSize, Test)] _payload = void; T getPayload(){ return cast(T)(_payload.ptr); } alias getPayload this; static Scoped opCall(Args...)(Args args) if (is(typeof(T.init.__ctor(args)))) { // TODO: should also provide decent error message Scoped!T s = void; emplace!T(cast(void[])s._payload,args); return s; } ~this() { static if (is(typeof(getPayload.__dtor()))) { getPayload.__dtor(); } } } final class Test { // 32 bytes each instance int i1, i2, i3, i4, i5, i6; this(int ii1, int ii2, int ii3, int ii4, int ii5, int ii6) { this.i1 = ii1; this.i2 = ii2; this.i3 = ii3; this.i4 = ii4; this.i5 = ii5; this.i6 = ii6; } void doSomething(int ii1, int ii2, int ii3, int ii4, int ii5, int ii6) { } } void main(string[] args) { enum int N = 10_000_000; int i; while (i < N) { auto testObject = Scoped!Test(i, i, i, i, i, i); //scope testObject = new Test(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); testObject.doSomething(i, i, i, i, i, i); i++; } } Andrei

Thanks Andrei!!!
Jul 21 2010
parent Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
Rory Mcguire wrote:
 Andrei Alexandrescu wrote:

 Thanks Andrei!!!

Don't mention it. Thanks for not over-quoting in the future :o). Andrei
Jul 21 2010
prev sibling next sibling parent Rory Mcguire <rjmcguire gm_no_ail.com> writes:
Andrei Alexandrescu wrote:

 Here's the code I used for testing:
 

My timings with your code, I included the one without -inline because I havn't been using -inline in my other tests: $ dmd -O -release -inline program1_2.d $ time ./program1_2 real 0m0.526s user 0m0.520s sys 0m0.000s $ dmd -O -release program1_2.d $ time ./program1_2 real 0m0.820s user 0m0.810s sys 0m0.000s and bearofiles with -inline: $ time ./program1 real 0m2.267s user 0m2.260s sys 0m0.000s Nice improvement...so is this in phobos or druntime yet? :D
Jul 21 2010
prev sibling parent reply Dmitry Olshansky <dmitry.olsh gmail.com> writes:
On 21.07.2010 23:16, Andrei Alexandrescu wrote:
 I compiled and ran the tests myself with -O -release -inline and got 
 1.95s for Dmitry's implementation and 1.55s for bearophile's.

 I optimized Dmitry's implementation in two ways: I replaced the call 
 to clear() with a straight call to the destructor and added = void in 
 two places to avoid double initialization. I got 1.11s, which 
 significantly undercuts the implementation using scope.

 Here's the code I used for testing:

 struct Scoped(T) {
     ubyte[__traits(classInstanceSize, Test)] _payload = void;
     T getPayload(){
         return cast(T)(_payload.ptr);
     }
     alias getPayload this;

     static Scoped opCall(Args...)(Args args)
     if (is(typeof(T.init.__ctor(args)))) {
         // TODO: should also provide decent error message
         Scoped!T s = void;
         emplace!T(cast(void[])s._payload,args);
         return s;
     }
     ~this() {
         static if (is(typeof(getPayload.__dtor()))) {
             getPayload.__dtor();
         }
     }
 }

this implementation still has issues with it: it calls dtor twice. Not a good trait for RAII technique ! :) Since it's now considered useful I feel myself obliged to enhance and correct it. Sadly enough I still haven't managed to apply an inner template trick. Here's the end result along with a simple unittest: struct Scoped(T){ ubyte[__traits(classInstanceSize, T)] _scopedPayload = void; T getScopedPayload(){ return cast(T)(_scopedPayload.ptr); } alias getScopedPayload this; this(Args...)(Args args){ static if (!is(typeof(getScopedPayload.__ctor(args)))) { static assert(false,"Scoped: wrong arguments passed to ctor"); }else { emplace!T(cast(void[])_scopedPayload,args); } } ~this() { static if (is(typeof(getScopedPayload.__dtor()))) { getScopedPayload.__dtor(); } } } //also track down destruction/construction class A{ this(int a){ writeln("A with ",a); } this(real r){ writeln("A with ",r); } ~this(){ writeln("A destroyed"); } } unittest{ { auto a = Scoped!A(42); } { auto b = Scoped!A(5.5); } } -- Dmitry Olshansky
Jul 21 2010
parent reply Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
On 07/21/2010 03:35 PM, Dmitry Olshansky wrote:
 Thanks for kind feedback (and showing some optimization tricks). Also
 this implementation still has issues with it: it calls dtor twice. Not a
 good trait for RAII technique ! :)
 Since it's now considered useful I feel myself obliged to enhance and
 correct it. Sadly enough I still haven't managed to apply an inner
 template trick.
 Here's the end result along with a simple unittest:

The double destructor hits a bug in the compiler implementation. Anyway, here's the committed code: http://www.dsource.org/projects/phobos/changeset/1774 It uses a new idiom that is enabled by auto returns - defines a struct inside the function and returns it. That's a veritable existential type! (http://stackoverflow.com/questions/292274/what-is-an-existential-type) I expect more of that idiom in the upcoming commits. Andrei
Jul 21 2010
next sibling parent reply Dmitry Olshansky <dmitry.olsh gmail.com> writes:
On 22.07.2010 2:27, Andrei Alexandrescu wrote:
 On 07/21/2010 03:35 PM, Dmitry Olshansky wrote:
 Thanks for kind feedback (and showing some optimization tricks). Also
 this implementation still has issues with it: it calls dtor twice. Not a
 good trait for RAII technique ! :)
 Since it's now considered useful I feel myself obliged to enhance and
 correct it. Sadly enough I still haven't managed to apply an inner
 template trick.
 Here's the end result along with a simple unittest:

The double destructor hits a bug in the compiler implementation. Anyway, here's the committed code: http://www.dsource.org/projects/phobos/changeset/1774 It uses a new idiom that is enabled by auto returns - defines a struct inside the function and returns it. That's a veritable existential type! (http://stackoverflow.com/questions/292274/what-is-an-existential-type) I expect more of that idiom in the upcoming commits.

I guess I should clarify the double destruction problem: it was caused by static opCall, the current implementation in Phobos does not suffers from it. -- Dmitry Olshansky
Jul 22 2010
parent Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
Dmitry Olshansky wrote:
 On 22.07.2010 2:27, Andrei Alexandrescu wrote:
 On 07/21/2010 03:35 PM, Dmitry Olshansky wrote:
 Thanks for kind feedback (and showing some optimization tricks). Also
 this implementation still has issues with it: it calls dtor twice. Not a
 good trait for RAII technique ! :)
 Since it's now considered useful I feel myself obliged to enhance and
 correct it. Sadly enough I still haven't managed to apply an inner
 template trick.
 Here's the end result along with a simple unittest:

The double destructor hits a bug in the compiler implementation. Anyway, here's the committed code: http://www.dsource.org/projects/phobos/changeset/1774 It uses a new idiom that is enabled by auto returns - defines a struct inside the function and returns it. That's a veritable existential type! (http://stackoverflow.com/questions/292274/what-is-an-existential-type) I expect more of that idiom in the upcoming commits.

I guess I should clarify the double destruction problem: it was caused by static opCall, the current implementation in Phobos does not suffers from it.

It doesn't, but just because it uses some heavy-handed tricks such as casting from an untyped buffer to Scoped. I found two distinct bugs by working on Scoped, which I'll submit soon. Andrei
Jul 22 2010
prev sibling parent Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
Philippe Sigaud wrote:
 On Thu, Jul 22, 2010 at 00:27, Andrei Alexandrescu 
 <SeeWebsiteForEmail erdani.org <mailto:SeeWebsiteForEmail erdani.org>> 
 wrote:
 
     The double destructor hits a bug in the compiler implementation.
     Anyway, here's the committed code:
 
     http://www.dsource.org/projects/phobos/changeset/1774
 
     It uses a new idiom that is enabled by auto returns - defines a
     struct inside the function and returns it. That's a veritable
     existential type!
     (http://stackoverflow.com/questions/292274/what-is-an-existential-type)
     I expect more of that idiom in the upcoming commits..
 
 
 Hmm, so the struct Scoped is implicitly parametrized by T and Args.... Cool.
 
 Why do you put a second layer of (Args...) in Scoped constructor? Why 
 not just
 
 this(Args args) if (etc) {...}
 
 And, in your case, if you used Args inside Scoped (which you don't do), 
 would that be the ctor's Args which'd be used?
 
 
 Philippe

Great point. I simplified the implementation: http://www.dsource.org/projects/phobos/changeset/1776
 PS: too bad that, if bug 2581 is not squashed, your scoped won't show in 
 the docs :(

I'll fix that later with a version(ddoc). Andrei
Jul 22 2010
prev sibling parent Philippe Sigaud <philippe.sigaud gmail.com> writes:
--00032555e4421a5758048bfdfa3b
Content-Type: text/plain; charset=ISO-8859-1

On Thu, Jul 22, 2010 at 00:27, Andrei Alexandrescu <
SeeWebsiteForEmail erdani.org> wrote:

 The double destructor hits a bug in the compiler implementation. Anyway,
 here's the committed code:

 http://www.dsource.org/projects/phobos/changeset/1774

 It uses a new idiom that is enabled by auto returns - defines a struct
 inside the function and returns it. That's a veritable existential type! (
 http://stackoverflow.com/questions/292274/what-is-an-existential-type) I
 expect more of that idiom in the upcoming commits.

Hmm, so the struct Scoped is implicitly parametrized by T and Args... Cool. Why do you put a second layer of (Args...) in Scoped constructor? Why not just this(Args args) if (etc) {...} And, in your case, if you used Args inside Scoped (which you don't do), would that be the ctor's Args which'd be used? Philippe PS: too bad that, if bug 2581 is not squashed, your scoped won't show in the docs :( --00032555e4421a5758048bfdfa3b Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable <div class=3D"gmail_quote">On Thu, Jul 22, 2010 at 00:27, Andrei Alexandres= cu <span dir=3D"ltr">&lt;<a href=3D"mailto:SeeWebsiteForEmail erdani.org">S= eeWebsiteForEmail erdani.org</a>&gt;</span> wrote:<br><blockquote class=3D"= gmail_quote" style=3D"margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb= (204, 204, 204); padding-left: 1ex;"> The double destructor hits a bug in the compiler implementation. Anyway, he= re&#39;s the committed code:<br> <br> <a href=3D"http://www.dsource.org/projects/phobos/changeset/1774" target=3D= "_blank">http://www.dsource.org/projects/phobos/changeset/1774</a><br> <br> It uses a new idiom that is enabled by auto returns - defines a struct insi= de the function and returns it. That&#39;s a veritable existential type! (<= a href=3D"http://stackoverflow.com/questions/292274/what-is-an-existential-= type" target=3D"_blank">http://stackoverflow.com/questions/292274/what-is-a= n-existential-type</a>) I expect more of that idiom in the upcoming commits= .<font color=3D"#888888"></font></blockquote> <div><br>Hmm, so the struct Scoped is implicitly parametrized by T and Args= ... Cool.<br><br>Why do you put a second layer of (Args...) in Scoped const= ructor? Why not just<br><br>this(Args args) if (etc) {...}<br><br>And, in y= our case, if you used Args inside Scoped (which you don&#39;t do), would th= at be the ctor&#39;s Args which&#39;d be used?<br> <br><br>Philippe<br><br>PS: too bad that, if bug 2581 is not squashed, your= scoped won&#39;t show in the docs :(<br></div></div> --00032555e4421a5758048bfdfa3b--
Jul 22 2010