Re: Pure not acting pure.
On 2011-06-15 23:29:46 -0400, Charles McAnany mcana...@rose-hulman.edu said: Ah, so does the compiler figure out which ones are strongly and weakly pure and then optimize as appropriate? Is there a way to indicate that a function is strongly pure? Because it would seem odd to call a function you thought was pure and wind up with a mutated argument. Just make sure all the parameters are either const or immutable or passed by copy and do not contain any pointer or reference. That'll make the function strongly pure, and the compiler will be able optimize. -- Michel Fortin michel.for...@michelf.com http://michelf.com/
Re: Pure not acting pure.
Michel Fortin:
Just make sure all the parameters are either const or immutable or
passed by copy and do not contain any pointer or reference. That'll
make the function strongly pure, and the compiler will be able optimize.
I have just done a little test on this, currently DMD calls sqr one time in
main() only if you add nothrow too:
pure nothrow int sqr(int x) { return x * x; }
int main() {
enum int n = 10;
immutable int r = sqr(n) + sqr(n);
return r;
}
Bye,
bearophile
D documentation
I'm working a small but reasonably interesting D project which I'd like to, ultimately, open source. To make it attractive I need to document my class. Is there anything remotely like Javadoc which works with D?
Re: D documentation
Lloyd Dupont: I'm working a small but reasonably interesting D project which I'd like to, ultimately, open source. To make it attractive I need to document my class. Is there anything remotely like Javadoc which works with D? http://www.digitalmars.com/d/2.0/ddoc.html I suggest you to take a general look at D documentation. Bye, bearophile
Anybody want to help this Haskell programmer out about type classes?
http://stackoverflow.com/questions/6328444/type-classes-in-d This guy seems to be very optimistic about D's functional possibilities. Unfortunately, the only answer there right now is a generic ramble concerning D.
Re: D documentation
Thanks bearophile! bearophile wrote in message news:itcs1k$1l8t$1...@digitalmars.com... Lloyd Dupont: I'm working a small but reasonably interesting D project which I'd like to, ultimately, open source. To make it attractive I need to document my class. Is there anything remotely like Javadoc which works with D? http://www.digitalmars.com/d/2.0/ddoc.html I suggest you to take a general look at D documentation. Bye, bearophile
__traits, compile time member type info
I'm trying to build an introspection system for a project I have.
I already have a working template to get members value
=== working members getter ===
Variant GETTER(T, string member)(Object target)
{
T tt = cast(T)target;
if (!tt)
throw new ReflectionException(target is null or not ~T.stringof
~ value: ~target.toString());
return Variant(__traits(getMember, tt, member));
}
===
Now I'm trying to implement a working setter. I have a problem for testing
and converts the value.
Ideally I'd like something like the code below
= not compiling setter
void SETTER(T, string member)(Object target, Variant value)
{
T tt = cast(T)target;
if (!tt)
throw new ReflectionException(target is null or not ~T.stringof
~ value: ~target.toString());
if(!value.convertsTo!( typeid(__traits(getMember, T, member))) )
throw new ReflectionException(Can't convert ~value.stringof ~ to
~typeid(__traits(getMember, T, member)).toString());
__traits(getMember, tt, member) = value.coerce(
typeid(__traits(getMember, T, member)) );
}
==
it doesn't compile because I don't know how to get the type, at compile
time, of member member of type T
how do I do that please?
Solved!
work with typeinfo!! :)
=
void SETTER(T, string member)(Object target, Variant value)
{
T tt = cast(T)target;
if (!tt)
throw new ReflectionException(target is null or not ~T.stringof
~ value: ~target.toString());
if(!value.convertsTo!( typeof(__traits(getMember, T, member)) ))
throw new ReflectionException(Can't convert ~value.stringof ~ to
~typeof(__traits(getMember, T, member)).stringof);
__traits(getMember, tt, member) = value.coerce!(
typeof(__traits(getMember, T, member)) );
}
==
Re: __traits, compile time member type info
works with typeinfo!
void SETTER(T, string member)(Object target, Variant value)
{
T tt = cast(T)target;
if (!tt)
throw new ReflectionException(target is null or not ~T.stringof
~ value: ~target.toString());
if(!value.convertsTo!( typeof(__traits(getMember, T, member)) ))
throw new ReflectionException(Can't convert ~value.stringof ~ to
~typeof(__traits(getMember, T, member)).stringof);
__traits(getMember, tt, member) = value.coerce!(
typeof(__traits(getMember, T, member)) );
}
Lloyd Dupont wrote in message news:itcthc$1onk$1...@digitalmars.com...
I'm trying to build an introspection system for a project I have.
I already have a working template to get members value
=== working members getter ===
Variant GETTER(T, string member)(Object target)
{
T tt = cast(T)target;
if (!tt)
throw new ReflectionException(target is null or not ~T.stringof
~ value: ~target.toString());
return Variant(__traits(getMember, tt, member));
}
===
Now I'm trying to implement a working setter. I have a problem for testing
and converts the value.
Ideally I'd like something like the code below
= not compiling setter
void SETTER(T, string member)(Object target, Variant value)
{
T tt = cast(T)target;
if (!tt)
throw new ReflectionException(target is null or not ~T.stringof
~ value: ~target.toString());
if(!value.convertsTo!( typeid(__traits(getMember, T, member))) )
throw new ReflectionException(Can't convert ~value.stringof ~ to
~typeid(__traits(getMember, T, member)).toString());
__traits(getMember, tt, member) = value.coerce(
typeid(__traits(getMember, T, member)) );
}
==
it doesn't compile because I don't know how to get the type, at compile
time, of member member of type T
how do I do that please?
templated overloaded operator problem.. :~
I have 2 overload of the opCall() in one of my class.
They cause a compile time error / conflict... any idea on how to solve it?
=
class MemberDesc
{
Variant opCall(Object target)
{
return getter(target);
}
void opCall(T)(Object target, T value)
{
setter(target, Variant(value));
}
const string name;
const TypeInfo type;
const Variant function(Object target) getter;
const void function(Object target, Variant value) setter;
this(
string name,
TypeInfo type,
Variant function(Object target) getter,
void function(Object target, Variant value) setter = null)
{
this.name = name;
this.type = type;
this.getter = getter;
this.setter = setter;
}
}
=
Re: templated overloaded operator problem.. :~
Add a pair of parentheses to the first overload to add an empty template
argument list – currently, template and non-template functions can't be
overloaded.
David
On 6/16/11 2:59 PM, Lloyd Dupont wrote:
I have 2 overload of the opCall() in one of my class.
They cause a compile time error / conflict... any idea on how to solve it?
=
class MemberDesc
{
Variant opCall(Object target)
{
return getter(target);
}
void opCall(T)(Object target, T value)
{
setter(target, Variant(value));
}
const string name;
const TypeInfo type;
const Variant function(Object target) getter;
const void function(Object target, Variant value) setter;
this(
string name, TypeInfo type,
Variant function(Object target) getter, void function(Object target,
Variant value) setter = null)
{
this.name = name;
this.type = type;
this.getter = getter;
this.setter = setter;
}
}
=
Re: templated overloaded operator problem.. :~
Ho thanks, even better than the work around I just found! :) David Nadlinger wrote in message news:itcvbj$1td2$1...@digitalmars.com... Add a pair of parentheses to the first overload to add an empty template argument list – currently, template and non-template functions can't be overloaded. David
Re: Advice on threading/fibers/?
On Wed, 15 Jun 2011 23:57:25 +, Justin Whear wrote: Consider the following: You have 10 million data points and you need to apply a multipass algorithm to them. Each pass is like a cellular automata: it can read from the previous pass but it doesn't know the current values. This makes the actual processing of each value trivially parallelizable. The actual operation for each value is fairly simple and cheap (essentially a multidimensional ancestor-child averaging operation). After each value has been operated on once, the pass is complete and the current and old buffers are switched (conceptually, the current buffer can only be written to, the old buffer can only be read--using __gshared here). The number of passes is not fixed; in the course of each value operation, an error is computed. When the worst individual error falls below a certain threshold, the algorithm is finished. Generally this will take between one thousand and ten thousand passes. How would you go about parallelizing this? My thought is to take the map/reduce approach within each pass: each thread/fiber takes a slice of the dataset, makes its modifications, then returns an error summary. These summaries are quickly combined and the algorithm loop decides whether to run again. Each pass shouldn't take more than a second or two, so I'm not sure whether introducing the overhead of spawning, say, 10 threads each pass is worthwhile (times 5000 passes). On the other hand, I have plenty of CPUs to throw at it (at least 16 cores, each with hyperthreading) and am in a situation where as fast as possible is important (while individual datasets may not grow, the number of them is). Any thoughts appreciated. I would recommend you take a look at the new std.parallelism module, which was introduced in the most recent DMD release (2.053): http://www.d-programming-language.org/phobos-prerelease/ std_parallelism.html -Lars
smarter reflection issue
I have a MemberDesc class which describe a class's members.
I fill it with a template method like that (with GETTER and SETTER some
other templated method I wrote)
=
MemberDesc MEMBER(T, string memberName)()
{
TypeInfo ti = typeid( typeof(__traits(getMember, T, memberName)) );
Variant function(Object target) getter = null;
getter = GETTER!(T, memberName);
void function(Object target, Variant value) setter = null;
setter = SETTER!(T, memberName);
return new MemberDesc(memberName, ti , getter, setter);
}
=
And it works except that I don't do any check that the setter / getter
method exist!
I tried something like that
static if( __traits(compiles, __traits(getMember, T, member)) )
{
getter = GETTER!(T, memberName);
}
but this always fail... mm.. how could I check for the getter?
and i guess it's even harder for the setter!
how about something like that (how to fix it?)
=
static if( __traits(compiles, __traits(getMember, T, member) =
typeof(__traits(getMember, T, memberName)).init) )
{
setter = SETTER!(T, memberName);
}
=
Re: Advice on threading/fibers/?
On 06/16/2011 06:22 AM, Lars T. Kyllingstad wrote: On Wed, 15 Jun 2011 23:57:25 +, Justin Whear wrote: Consider the following: You have 10 million data points and you need to apply a multipass algorithm to them. Each pass is like a cellular automata: it can read from the previous pass but it doesn't know the current values. This makes the actual processing of each value trivially parallelizable. The actual operation for each value is fairly simple and cheap (essentially a multidimensional ancestor-child averaging operation). After each value has been operated on once, the pass is complete and the current and old buffers are switched (conceptually, the current buffer can only be written to, the old buffer can only be read--using __gshared here). The number of passes is not fixed; in the course of each value operation, an error is computed. When the worst individual error falls below a certain threshold, the algorithm is finished. Generally this will take between one thousand and ten thousand passes. How would you go about parallelizing this? My thought is to take the map/reduce approach within each pass: each thread/fiber takes a slice of the dataset, makes its modifications, then returns an error summary. These summaries are quickly combined and the algorithm loop decides whether to run again. Each pass shouldn't take more than a second or two, so I'm not sure whether introducing the overhead of spawning, say, 10 threads each pass is worthwhile (times 5000 passes). On the other hand, I have plenty of CPUs to throw at it (at least 16 cores, each with hyperthreading) and am in a situation where as fast as possible is important (while individual datasets may not grow, the number of them is). Any thoughts appreciated. I would recommend you take a look at the new std.parallelism module, which was introduced in the most recent DMD release (2.053): http://www.d-programming-language.org/phobos-prerelease/ std_parallelism.html -Lars I wrote an article on using std.parallelism for a bucket-sort algorithm. http://www.codestrokes.com/archives/116 I hope it helps.
Re: Pure not acting pure.
On Thu, 16 Jun 2011 06:52:45 -0400, Michel Fortin wrote: On 2011-06-15 23:29:46 -0400, Charles McAnany mcana...@rose-hulman.edu said: Ah, so does the compiler figure out which ones are strongly and weakly pure and then optimize as appropriate? Is there a way to indicate that a function is strongly pure? Because it would seem odd to call a function you thought was pure and wind up with a mutated argument. Just make sure all the parameters are either const or immutable or passed by copy and do not contain any pointer or reference. That'll make the function strongly pure, and the compiler will be able optimize. If you want a strongly pure function, the parameters need to be immutable or implicitly convertible to immutable. const references may be mutated elsewhere. -Lars
Re: Is it reasonable to learn D
On 2011-06-07 22:45, Timon Gehr wrote: Fabian wrote: - There are no maintained GUI libraries I don't know about this but I think QtD and DWT are still being maintained? DWT is still maintained, although very slowly. -- /Jacob Carlborg
Re: Is it reasonable to learn D
On 2011-06-07 23:02, Fabian wrote: The community does not grow if people stay away because it is small. Thank you for your answer. - You've got a big point! I don't know about this but I think QtD and DWT are still being maintained? I can't see any changes on this web page: http://www.dsource.org/projects/dwt/wiki There's currently no reason to update the wiki, please see http://hg.dsource.org/projects/dwt2 for changelog. -- /Jacob Carlborg
Re: Is it reasonable to learn D
On 2011-06-07 23:33, Jonathan M Davis wrote: On 2011-06-07 14:02, Fabian wrote: The community does not grow if people stay away because it is small. Thank you for your answer. - You've got a big point! I don't know about this but I think QtD and DWT are still being maintained? I can't see any changes on this web page: http://www.dsource.org/projects/dwt/wiki That's too bad :-( http://hg.dsource.org/projects/dwt2 DWT was recently ported to D2. Also, QtD is definitely under active development. But I'm sure that there's still plenty of work to be done. - Jonathan M Davis There is a lot of work to be done (for DWT) : * Fixing bugs * Finish porting the Mac version * Update to later versions of SWT * Port 64bit versions -- /Jacob Carlborg
Re: Install DWT2 using DMD and Tango
On 2011-06-08 22:59, Andrew Wiley wrote: On Wed, Jun 8, 2011 at 11:54 AM, Fabian contact-...@freenet.de mailto:contact-...@freenet.de wrote: Hi I'm trying to install DWT2 to create GUI applications with D. I have downloaded DWT2 with TortoiseHg already and I've installed Ruby and Rake. But when I try to build the packages I get the following error message: http://imageshack.us/photo/my-images/62/dwt2error.png/ Is there anybody who can help me? I believe it's possible to use DWT2 and Tango - isn't it? It's called DWT2 because it's a port of the existing DWT code to D2. Tango is a replacement standard library for D1, so no, they can't really be used together. I'm not sure about the state of DWT1, but it should be fairly stable, even if it hasn't been worked on recently. As an overview, D1 originally had Phobos as the standard library, but it wasn't that great, so an independent team built Tango as a replacement. Tango hasn't ever officially shipped with DMD, the reference compiler, but it's pretty much become the standard for D1 (with a few exceptions and attempts to reconcile it with Phobos). Nowadays, we have D2 and Phobos2 (which is generally just called Phobos). There's an ongoing project to port Tango to D2 as well, and Druntime, the runtime library, has been separated from Phobos2 to hopefully make Tango and Phobos2 coexist if/when that D2 port is completed (back in D1, Phobos and Tango both included their own runtimes, so you couldn't use both at the same time). In the meantime, Phobos2 far better than Phobos1, and Andrei has been leading its development. Learn to Tango with D obviously refers to D1/Tango The D Programming Language refers to D2/Phobos2 Hopefully that makes everything a bit clearer. DWT2 is supposed to work with D2/Phobos and D1/Tango, anything other is a bug. I haven't compiled DWT with D1/Tango for a while now so it might not compile. -- /Jacob Carlborg
Re: Install DWT2 using DMD and Tango
On 2011-06-08 20:54, Fabian wrote: Hi I'm trying to install DWT2 to create GUI applications with D. I have downloaded DWT2 with TortoiseHg already and I've installed Ruby and Rake. But when I try to build the packages I get the following error message: http://imageshack.us/photo/my-images/62/dwt2error.png/ Is there anybody who can help me? I believe it's possible to use DWT2 and Tango - isn't it? Greetings Fabian I haven't tested to compile DWT with D1/Tango for a while so it might not be compiling, I'll look in to it. -- /Jacob Carlborg
Re: Serialization libs?
On 2011-06-10 14:07, Robert Clipsham wrote: On 10/06/2011 08:30, Nick Sabalausky wrote: Are there any good D2 serialization libs out there that utilize introspecition (ie, don't have to manually specify all the member of each type), handle cyclic graphs and have flexible output? I've never used it, and I don't know if it's any good or even does what you need, but: http://dsource.org/projects/orange/ Don't know if it works with the current DMD. I'm also in the middle of a complete rewrite of the library. But version 0.0.1 should at least work with some version of DMD. -- /Jacob Carlborg
Re: Pure not acting pure.
Ok, I think I get it. That cleared it up. =). So, if you have a functioned labelled pure, it's your job to not pass it mutable arguments, but the compiler's job to make sure it doesn't mutate anything not in the arguments. And that's why a strongly pure function can call a weakly pure one - only the first function's internal state can be mutated by a weakly pure function. Thanks!
Re: Pure not acting pure.
On Thu, 16 Jun 2011 17:38:27 +, Charles McAnany wrote: Ok, I think I get it. That cleared it up. =). So, if you have a functioned labelled pure, it's your job to not pass it mutable arguments, but the compiler's job to make sure it doesn't mutate anything not in the arguments. And that's why a strongly pure function can call a weakly pure one - only the first function's internal state can be mutated by a weakly pure function. Thanks! Exactly. :) -Lars
Re: Pure not acting pure.
On 2011-06-16 10:38, Charles McAnany wrote: Ok, I think I get it. That cleared it up. =). So, if you have a functioned labelled pure, it's your job to not pass it mutable arguments, but the compiler's job to make sure it doesn't mutate anything not in the arguments. And that's why a strongly pure function can call a weakly pure one - only the first function's internal state can be mutated by a weakly pure function. Thanks! Well, essentially. But it's a question of parameters, not arguments. It doesn't matter whether you pass the function mutable arguments or not. What matters is whether its parameters are immutable or implicitly immutable. If they are, then you'll be forced to pass it arguments which are immutable or implicitly immutable. If they aren't, then the function is weakly pure and no optimizations can take place (but the function still can't access mutable global variables and can still be called from other pure functions), regardless of how mutable the arguments are. - Jonathan M Davis
Re: Pure not acting pure.
On Thu, 16 Jun 2011 14:33:17 -0400, Jonathan M Davis jmdavisp...@gmx.com wrote: On 2011-06-16 10:38, Charles McAnany wrote: Ok, I think I get it. That cleared it up. =). So, if you have a functioned labelled pure, it's your job to not pass it mutable arguments, but the compiler's job to make sure it doesn't mutate anything not in the arguments. And that's why a strongly pure function can call a weakly pure one - only the first function's internal state can be mutated by a weakly pure function. Thanks! Well, essentially. But it's a question of parameters, not arguments. It doesn't matter whether you pass the function mutable arguments or not. Actually, it can matter. For instance, a pure function like this: pure int foo(const(int)* m); can be strong pure if you pass it a pointer to immutable data. -Steve
sorting a BigInt[] with algorithm.sort()
Hi, all. This is a strange one.
import std.bigint;
import std.algorithm;
void main(){
BigInt[] ar = [BigInt(2), BigInt(1), BigInt(3)];
sort(ar);
}
object.Exception@src\rt\arraycat.d(40): overlapping array copy
Is this a library bug? Or is sort() not supposed to work on this data?
Oh, additionally,
bool res = ar[0] ar[1]; compiles and executes just fine, as does the normal
swapping algorithm:
BigInt t = ar[0];
ar[0] = ar[1];
ar[1] = t;
Thanks,
Charles.
Re: Pure not acting pure.
On 2011-06-16 11:59, Steven Schveighoffer wrote: On Thu, 16 Jun 2011 14:33:17 -0400, Jonathan M Davis jmdavisp...@gmx.com wrote: On 2011-06-16 10:38, Charles McAnany wrote: Ok, I think I get it. That cleared it up. =). So, if you have a functioned labelled pure, it's your job to not pass it mutable arguments, but the compiler's job to make sure it doesn't mutate anything not in the arguments. And that's why a strongly pure function can call a weakly pure one - only the first function's internal state can be mutated by a weakly pure function. Thanks! Well, essentially. But it's a question of parameters, not arguments. It doesn't matter whether you pass the function mutable arguments or not. Actually, it can matter. For instance, a pure function like this: pure int foo(const(int)* m); can be strong pure if you pass it a pointer to immutable data. I believe that every time that Don has discussed it, he has made it clear that it's entirely a matter of the function's signature and that all parameters must be immutable or implicitly convertible to immutable for a function to be strongly pure. So, that's what I'm going by. However, I can see how passing immutable values to a pure function with const parameters could be optimized out just like if the parameters were actually immutable. So, it may be that things were changed so that that counts as well. Everything that I've seen discussed on it though has been about the parameters all having to be immutable or implicitly convertible to immutable. - Jonathan M Davis
Re: sorting a BigInt[] with algorithm.sort()
Charles McAnany: object.Exception@src\rt\arraycat.d(40): overlapping array copy Is this a library bug? Or is sort() not supposed to work on this data? It's a known bug: http://d.puremagic.com/issues/show_bug.cgi?id=5705 (Is Bugzilla down?) Bye, bearophile
Re: Pure not acting pure.
On Thu, 16 Jun 2011 16:36:01 -0400, Jonathan M Davis jmdavisp...@gmx.com wrote: On 2011-06-16 11:59, Steven Schveighoffer wrote: On Thu, 16 Jun 2011 14:33:17 -0400, Jonathan M Davis jmdavisp...@gmx.com wrote: On 2011-06-16 10:38, Charles McAnany wrote: Ok, I think I get it. That cleared it up. =). So, if you have a functioned labelled pure, it's your job to not pass it mutable arguments, but the compiler's job to make sure it doesn't mutate anything not in the arguments. And that's why a strongly pure function can call a weakly pure one - only the first function's internal state can be mutated by a weakly pure function. Thanks! Well, essentially. But it's a question of parameters, not arguments. It doesn't matter whether you pass the function mutable arguments or not. Actually, it can matter. For instance, a pure function like this: pure int foo(const(int)* m); can be strong pure if you pass it a pointer to immutable data. I believe that every time that Don has discussed it, he has made it clear that it's entirely a matter of the function's signature and that all parameters must be immutable or implicitly convertible to immutable for a function to be strongly pure. So, that's what I'm going by. However, I can see how passing immutable values to a pure function with const parameters could be optimized out just like if the parameters were actually immutable. So, it may be that things were changed so that that counts as well. Everything that I've seen discussed on it though has been about the parameters all having to be immutable or implicitly convertible to immutable. Don's the one who brought it to my attention (have no idea what thread it was in...) that the optimizations could be done based on this. AFAIK, this optimization is not implemented. -Steve
Re: smarter reflection issue
MemberDesc MEMBER(T, string memberName)() if(std.traits.hasMember!(T,
memberName))
{
...
}
Lloyd Dupont Wrote:
I have a MemberDesc class which describe a class's members.
I fill it with a template method like that (with GETTER and SETTER some
other templated method I wrote)
=
MemberDesc MEMBER(T, string memberName)()
{
TypeInfo ti = typeid( typeof(__traits(getMember, T, memberName)) );
Variant function(Object target) getter = null;
getter = GETTER!(T, memberName);
void function(Object target, Variant value) setter = null;
setter = SETTER!(T, memberName);
return new MemberDesc(memberName, ti , getter, setter);
}
=
And it works except that I don't do any check that the setter / getter
method exist!
I tried something like that
static if( __traits(compiles, __traits(getMember, T, member)) )
{
getter = GETTER!(T, memberName);
}
but this always fail... mm.. how could I check for the getter?
and i guess it's even harder for the setter!
how about something like that (how to fix it?)
=
static if( __traits(compiles, __traits(getMember, T, member) =
typeof(__traits(getMember, T, memberName)).init) )
{
setter = SETTER!(T, memberName);
}
=
Re: Pure not acting pure.
On 2011-06-16 13:47, Steven Schveighoffer wrote: On Thu, 16 Jun 2011 16:36:01 -0400, Jonathan M Davis jmdavisp...@gmx.com wrote: On 2011-06-16 11:59, Steven Schveighoffer wrote: On Thu, 16 Jun 2011 14:33:17 -0400, Jonathan M Davis jmdavisp...@gmx.com wrote: On 2011-06-16 10:38, Charles McAnany wrote: Ok, I think I get it. That cleared it up. =). So, if you have a functioned labelled pure, it's your job to not pass it mutable arguments, but the compiler's job to make sure it doesn't mutate anything not in the arguments. And that's why a strongly pure function can call a weakly pure one - only the first function's internal state can be mutated by a weakly pure function. Thanks! Well, essentially. But it's a question of parameters, not arguments. It doesn't matter whether you pass the function mutable arguments or not. Actually, it can matter. For instance, a pure function like this: pure int foo(const(int)* m); can be strong pure if you pass it a pointer to immutable data. I believe that every time that Don has discussed it, he has made it clear that it's entirely a matter of the function's signature and that all parameters must be immutable or implicitly convertible to immutable for a function to be strongly pure. So, that's what I'm going by. However, I can see how passing immutable values to a pure function with const parameters could be optimized out just like if the parameters were actually immutable. So, it may be that things were changed so that that counts as well. Everything that I've seen discussed on it though has been about the parameters all having to be immutable or implicitly convertible to immutable. Don's the one who brought it to my attention (have no idea what thread it was in...) that the optimizations could be done based on this. AFAIK, this optimization is not implemented. Well, conceptually, it seems solid. And ultimately, it's up to the compiler what it decides is actually strongly pure, so it'll probably be implemented at some point. I think that it's safe to say that in the long run at least, we'll optimize pure functions as much as we can reasonably get away with. But regardless, I'm _very_ glad that Don came up with the weakly pure idea, or pure would have been almost unusable. - Jonathan M Davis
Re: smarter reflection issue
Hi Jesse, this won't work!
It's my fault in not explaining my problem well though...
I forget to mention something...
I'm using property syntax and try to call the method
Say I have
=
private int _foo;
@property public int Foo() { return _foo; }
@property public void Foo(int value) { _foo = value; }
=
if I call MEMBER(MyClass, Foo)
I'd like to do some static test for int Foo() and void Foo(int)
So... how will I go on solving that?
Jesse Phillips wrote in message news:itdqs4$mv0$1...@digitalmars.com...
MemberDesc MEMBER(T, string memberName)() if(std.traits.hasMember!(T,
memberName))
{
...
}
Lloyd Dupont Wrote:
I have a MemberDesc class which describe a class's members.
I fill it with a template method like that (with GETTER and SETTER some
other templated method I wrote)
=
MemberDesc MEMBER(T, string memberName)()
{
TypeInfo ti = typeid( typeof(__traits(getMember, T, memberName)) );
Variant function(Object target) getter = null;
getter = GETTER!(T, memberName);
void function(Object target, Variant value) setter = null;
setter = SETTER!(T, memberName);
return new MemberDesc(memberName, ti , getter, setter);
}
=
And it works except that I don't do any check that the setter / getter
method exist!
I tried something like that
static if( __traits(compiles, __traits(getMember, T, member)) )
{
getter = GETTER!(T, memberName);
}
but this always fail... mm.. how could I check for the getter?
and i guess it's even harder for the setter!
how about something like that (how to fix it?)
=
static if( __traits(compiles, __traits(getMember, T, member) =
typeof(__traits(getMember, T, memberName)).init) )
{
setter = SETTER!(T, memberName);
}
=
allocating gobs of memory to my program
Hi, all. I'm back! I've got an enormous array that I need to store, preferably in RAM. (It's iterated a bunch.) I have 16 Gb on my machine, and at any time, about 12 Gb is free. I'd like to be able to use about 10 Gb for this program. But when I try to use more than about 800 Mb, I get Memory allocation failed. (I'm using new long[], not malloc(), but I do free() variables because the garbage collector was having trouble keeping up with me.) Is there a switch I can mark to say that the runtime might have to deal with lots of memory? Incidentally, Win7 x64, Intel I7 @4.4 GHz, compiling with dmd -O -release -inline. Thanks, Charles
Re: allocating gobs of memory to my program
On 6/17/11 12:32 AM, Charles McAnany wrote: Win7 x64, Intel I7 @4.4 GHz, compiling with dmd -O -release -inline. Regardless whether you are running on x86 or x86_64, DMD is only able to create 32 bit binaries on Windows. David
Type qualifiers - inout error
Hi,
I'm new to D and trying to follow Alexandrescu's TDPL code examples I came
across an error on the code below:
class A
{
inout(int) val() inout
{
return _val; // error - see below
}
private int _val;
};
The compiler (dmd v2.052) complains on the marked line with the message
Error: inout on return means inout must be on a parameter as well for inout
inout(int)().
What am I doing wrong? TDPL has a very similar example...
Regards,
Claudiu
P.S. Is there a netiquette (a la C++ FAQ lite) about posting on this forum
that I should be aware of?
Instantiating a Tuple with immutable fields.
Hi,
I am trying to instantiate a Tuple that contains an immutable field.
Is there a way to do this with the current implementation? The
compiler gives me this error:
std/typecons.d(383): Error: can only initialize const member
_field_field_0 inside constructor
tuple_test.d(5): Error: template instance
std.typecons.Tuple!(immutable(int)).Tuple.__ctor!(int) error
instantiating
tuple_test.d(3): Error: function D main has no return statement, but
is expected to return a value of type int
here is the code:
import std.typecons;
int main()
{
Tuple!(immutable(int)) var = Tuple!(immutable(int))(5);
assert(var[0] == 5);
}
Thanks!
Re: allocating gobs of memory to my program
Hm. I'm not too good on architecture - does that mean it's impossible for an x32 program to have access to more memory? Is there, maybe, an x64 C library that I could use to abstract the memory out (Just a huge array wrapper, basically)? Or, that failing, does GCC automatically generate x64 code on an x64 machine? I could probably write the procedure in C... but yuck. David Nadlinger Wrote: On 6/17/11 12:32 AM, Charles McAnany wrote: Win7 x64, Intel I7 @4.4 GHz, compiling with dmd -O -release -inline. Regardless whether you are running on x86 or x86_64, DMD is only able to create 32 bit binaries on Windows. David
Re: allocating gobs of memory to my program
On Thu, Jun 16, 2011 at 10:20 PM, Charles McAnany mcana...@rose-hulman.edu wrote: Hm. I'm not too good on architecture - does that mean it's impossible for an x32 program to have access to more memory? Is there, maybe, an x64 C library that I could use to abstract the memory out (Just a huge array wrapper, basically)? Or, that failing, does GCC automatically generate x64 code on an x64 machine? I could probably write the procedure in C... but yuck. David Nadlinger Wrote: On 6/17/11 12:32 AM, Charles McAnany wrote: Win7 x64, Intel I7 @4.4 GHz, compiling with dmd -O -release -inline. Regardless whether you are running on x86 or x86_64, DMD is only able to create 32 bit binaries on Windows. David I don't know why you can't allocate more than 800mb but if you want to get around this by caching in other processes' memory then I can recommend memcache: http://memcached.org/
Re: Type qualifiers - inout error
Claudiu Verdes: What am I doing wrong? TDPL has a very similar example... I think this TDPL example is not good. And generally inout is not fully implemented in D yet. P.S. Is there a netiquette (a la C++ FAQ lite) about posting on this forum that I should be aware of? I think there isn't one. Anything goes! ;-) Bye, bearophile
Re: Instantiating a Tuple with immutable fields.
Jose Armando Garcia: I am trying to instantiate a Tuple that contains an immutable field. Is there a way to do this with the current implementation? I think this is not yet possible. I am not sure, but maybe even small changes to conts are needed to fix this problem. Bye, bearophile
Re: Pure not acting pure.
Jonathan M Davis: I'm _very_ glad that Don came up with the weakly pure idea, or pure would have been almost unusable. I think Don has said this wasn't fully his idea. The D community is good at creating ideas in group :-) Bye, bearophile
Strange behavior when concatenating array
It looks like the rt is not calling the postblit constructor when
concatenating arrays. For example, the following code:
import std.stdio;
struct Test
{
this(this) { writeln(copy done); }
void opAssign(Test rhs) { writeln(assignment done); }
~this() { writeln(destructor called); }
}
void main()
{
Test[] tests = new Test[1];
{
Test test;
tests ~= test;
}
writeln(done);
}
Gives the following output:
destructor called
done
The dtr for 'Test test;' is getting call after the scope exits but the
postblit ctr for 'tests[0]' is never called. I believe the output of
this code should either be:
done
or:
copy done
destructor called
done
Thanks!
-Jose
Re: Strange behavior when concatenating array
On Fri, Jun 17, 2011 at 1:05 AM, Jose Armando Garcia jsan...@gmail.com wrote: tests ~= test; Btw, if you replace this with 'test[0] = test;' it works as expected. The postblit ctor and the assignment operator get called and the dtor is called twice.
