220 29394 <215ef486-c593-4a7d-bb89-d134b18a9fab@isocpp.org> article Path: news.gmane.org!.POSTED!not-for-mail From: "'Michael Davies' via ISO C++ Standard - Future Proposals" Newsgroups: gmane.comp.lang.c++.isocpp.proposals Subject: Re: why not formally support meta programming. Date: Thu, 10 Nov 2016 02:35:02 -0800 (PST) Lines: 558 Approved: news@gmane.org Message-ID: <215ef486-c593-4a7d-bb89-d134b18a9fab@isocpp.org> References: <668e0a04-c988-4354-ae32-30f626f75403@isocpp.org> Reply-To: std-proposals@isocpp.org NNTP-Posting-Host: blaine.gmane.org Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="----=_Part_363_184363760.1478774102699" X-Trace: blaine.gmane.org 1478774142 4527 195.159.176.226 (10 Nov 2016 10:35:42 GMT) X-Complaints-To: usenet@blaine.gmane.org NNTP-Posting-Date: Thu, 10 Nov 2016 10:35:42 +0000 (UTC) Cc: billconan@gmail.com To: ISO C++ Standard - Future Proposals Original-X-From: std-proposals+bncBCS45AVH5YKBBV42SHAQKGQE32AQSOY@isocpp.org Thu Nov 10 11:35:38 2016 Return-path: Envelope-to: gclcip-std-proposals@m.gmane.org Original-Received: from mail-yw0-f198.google.com ([209.85.161.198]) by blaine.gmane.org with esmtp (Exim 4.84_2) (envelope-from ) id 1c4mhB-0004ao-F2 for gclcip-std-proposals@m.gmane.org; Thu, 10 Nov 2016 11:35:01 +0100 Original-Received: by mail-yw0-f198.google.com with SMTP id t11sf269293139ywe.3 for ; Thu, 10 Nov 2016 02:35:04 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=isocpp-org.20150623.gappssmtp.com; s=20150623; h=date:from:to:cc:message-id:in-reply-to:references:subject :mime-version:x-original-sender:reply-to:precedence:mailing-list :list-id:x-spam-checked-in-group:list-post:list-help:list-archive :list-subscribe:list-unsubscribe; bh=V2kBa8I6r8yMlME1mQAhWlqGy8PAVxnbPswjO0GYthA=; b=zPLg+t7mVvgAgYCwXD3j1CYWsn5s0FjFVKM/EYmLZKC8f1lobpyx6f6gnqMkouhXII bF9LlXlaMFjog34qjtDR8jyTQKmIan4pJdPPa9CiBV4tnzKSFQdNj7Bu4mCRpXDB2Ri0 vniUaHHjEQCy8PyK3LYm2C9+lJep6n5Htm2jHKITQX+0uW7VpsSdPy9KLK0FoodUMYIw yPQpQ3AK5/YzeCHWPj9iHbx1FOSgeHPCET8T12cSPr79K/oo1AUPsqEthAHx7BXj98qL J9NZMeWvWggSAdKOBwsIYo51seqZdF7ssmmoQtcF6xh7UXpPm2DlcEb5wXMqvwjp5o5f Vu9g== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20130820; h=x-gm-message-state:date:from:to:cc:message-id:in-reply-to :references:subject:mime-version:x-original-sender:reply-to :precedence:mailing-list:list-id:x-spam-checked-in-group:list-post :list-help:list-archive:list-subscribe:list-unsubscribe; bh=V2kBa8I6r8yMlME1mQAhWlqGy8PAVxnbPswjO0GYthA=; b=IbhSjluoOP1nfVIn6FhUBMl9O2tPcbj4TIrvhkm/2HSnKUcUCTq+uTveRxnvnQyh10 5YPiBDzRjcGApL22xqxBgWSLcWWofEGFGtfEPeHlX7VCB2PTkgmmrWRbxMfMBRElv2E9 8OfSgGr/taEp8DRSwSE4FTPQ9REyDodmwkKL6NMK5dyDVmV/MAkCQJ+glXrnJpxxMtiG SLppI9nu37dUu4kwCh6h7Y/zzTx2bZ5fb48wA2wXowIaaxiZINkqxjlh8qI9LqTwBr2C 13FJqZLTXU4XSEN7EL85Wh7AkapT/s10T7EFbyqOwWaozktxc25kr4eTO3UxexMGFE4D 7x6w== X-Gm-Message-State: ABUngvepRcmFrYWzGA59lTRneH1qmA1eK1+gPVPuHf915v0GIVttB6Ui010fDPnyi1pSQA== X-Received: by 10.13.227.67 with SMTP id m64mr908942ywe.36.1478774104442; Thu, 10 Nov 2016 02:35:04 -0800 (PST) X-BeenThere: std-proposals@isocpp.org Original-Received: by 10.157.43.24 with SMTP id o24ls6444145otb.41.gmail; Thu, 10 Nov 2016 02:35:03 -0800 (PST) X-Received: by 10.157.11.120 with SMTP id p53mr501826otd.19.1478774103541; Thu, 10 Nov 2016 02:35:03 -0800 (PST) In-Reply-To: <668e0a04-c988-4354-ae32-30f626f75403@isocpp.org> X-Original-Sender: MichaelDavies32@googlemail.com X-Original-From: Michael Davies Precedence: list Mailing-list: list std-proposals@isocpp.org; contact std-proposals+owners@isocpp.org List-ID: X-Google-Group-Id: 399137483710 List-Post: , List-Help: , List-Archive: List-Subscribe: , List-Unsubscribe: , Xref: news.gmane.org gmane.comp.lang.c++.isocpp.proposals:29394 Archived-At: ------=_Part_363_184363760.1478774102699 Content-Type: multipart/alternative; boundary="----=_Part_364_2086495820.1478774102700" ------=_Part_364_2086495820.1478774102700 Content-Type: text/plain; charset=UTF-8 I think that alternatively the other way to do this would be to use ADL(Argument Dependent Lookup) based namespace composure. You should look here: https://svn.boost.org/trac/boost/wiki/BestPracticeHandbook Essentially you would have something like this: namespace CPU { struct Device {}; //tag template void function_execution(DataType data, tag) {...} } namespace GPU { struct Device {}; //tag template void function_execution(DataType data, tag) {...} } template class BlahData { void function(DataType data) { function_execution(data, DeviceType()); } } template class Blah : public BlahData { DataType shared_member; void shared_function() { function(); } }; Obviously constexpr if is the way to go. But this, I think, is a possible alternative if you aren't ready to move on to the next standard or you are required to remain where you are. On Wednesday, November 9, 2016 at 7:32:54 PM UTC, bill...@gmail.com wrote: > > I'm trying to design a library for different hardware (CPU and GPU) > > For users, I want the interface looks like this: > > > Blah b(); > > or > > Blah c(); > > > ideally inside the class of Blah, I want to do this to avoid duplication > and reduce workload: > > template > class Blah > { > DataType shared_member; > > > #Meta_if(deviceUsed == GPU) //some kind of meta programming syntax > to enable code based on template > #{ > DataType member_only_used_by_gpu; > void function_on_gpu(); > #} > > > #Meta_if(deviceUsed == CPU) > #{ > DataType member_only_used_by_cpu; > void function_on_cpu(); > #} > > > void shared_function() > { > .... > #Meta_if(deviceUsed = GPU) { > .... > #} > } > }; > > > This is the actual code I have right now. > > as you can see, I have to use std::enable_if > > I really don't like it, because it is a trick, a hacky solution, and only > works for functions, not for members or function internals. > > and to make the following code to work, I had to try many options, some > don't compile, some don't do what I want. And It's really hard to > understand this syntax: > > template std::enable_if<(T & devices) != 0> > > > > template devices) != 0>::type > > > > template != 0>::type* = 0 > > > And for class members, since std::enable_if can't be used, I have to put > shared stuff into a base class. And this is only manageable to do when I > have few template variables. > > > typedef enum > > > { > > > CPU_NAIVE = 0x1, > > > CPU = 0x2, > > > CPU_SIMD = 0x2, > > > GPU = 0x4, > > > GPU_CUDA = 0x4 > > > } DeviceType; > > > #define DEVICE_SPECIFIC(devices) \ > > > template std::enable_if<(T & devices) != 0> > > > > > template > > > class Blob_base > > > { > > > protected: > > > const std::string m_name; > > > const unsigned int m_batchSize; > > > const unsigned int m_depth; > > > const unsigned int m_height; > > > const unsigned int m_width; > > > DataType *m_data; > > > }; > > > template > > > class Blob : public Blob_base > > > { > > > public: > > > Blob(const unsigned int batchSize, > > > const unsigned int depth, > > > const unsigned int height, > > > const unsigned int width, > > > const std::string &name = "no_name"); > > > bool init(); > > > DEVICE_SPECIFIC(CPU | GPU) > > > void test(); > > > }; > > > template > > > class Blob : public Blob_base > > > { > > > public: > > > bool init(); > > > Blob(const unsigned int batchSize, > > > const unsigned int depth, > > > const unsigned int height, > > > const unsigned int width, > > > const std::string &name = "no_name"); > > > }; > > > And because the limitation of partial specification of c++ (you have to > partially specify a class first before specifying its member functions), I > had to create many versions of the same class, with different template<> > variables. > > I have to duplicate lots of code. > > Also, I really want the flexibility to do this: > > template > class > { > #meta_if(device == CPU || device == CPU_SIMD) > int a,b; > void function() > {... }; > #meta_end > } > > > apparently using the base class still doesn't allow me to do this > > > I think many people share the same pain with me, if you google those > questions about std::enable_if and how to specify a template member > function. > > So if meta programming is so useful, why don't you consider adding it > formally to C++ by introducing the above syntax, rather than standardizing > the partially working ugly hack std::enable_if? > > > PHP for example, can be mixed with html. > > why can't c++ define something similar: > > > > such that what's inside will be executed by the preprocessor > and generate code? > > -- You received this message because you are subscribed to the Google Groups "ISO C++ Standard - Future Proposals" group. To unsubscribe from this group and stop receiving emails from it, send an email to std-proposals+unsubscribe@isocpp.org. To post to this group, send email to std-proposals@isocpp.org. To view this discussion on the web visit https://groups.google.com/a/isocpp.org/d/msgid/std-proposals/215ef486-c593-4a7d-bb89-d134b18a9fab%40isocpp.org. ------=_Part_364_2086495820.1478774102700 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable
I think that alternatively the other way to do this would = be to use ADL(Argument Dependent Lookup) based namespace composure.

= You should look here: https://svn.boost.org/trac/boost/wiki/BestPracticeHandboo= k

Essentially you would have something like this:
namespa= ce CPU
{
struct Device {}; //tag
template<typename DataTyp= e, typename <= code>tag= > vo= id function_execution(DataType <= /code>= data= , = t= ag) {...}=
}

namespace GPU
{
=
struct<= /span> Device = {}; //tag
templa= te<typename DataType, typename tag> void function_execution(= DataType data, tag) {...}
}

template<typename DeviceType, typename DataType>
class BlahData
{
=C2=A0= void f= unction(DataTyp= e data) { <= code>function_execution(<= code>data, DeviceType(= )); }
}

template<typename DeviceType, type= name Da= taType>
class Blah : <= /span>public <= /span>BlahData= <DeviceType, DataType>
{
=C2=A0 DataType shared_member;

=C2=A0 v= oid shared_function()
=C2=A0 {
=C2=A0 =C2=A0 function();<= /span>
=C2=A0 }
};

Obviously constexpr if is the way t= o go. But this, I think, is a possible alternative if you aren't ready = to move on to the next standard or you are required to remain where you are= ..

On Wednesday, November 9, 2016 at 7:32:54 PM UTC, bill...@gmail.co= m wrote:
I'm trying to desi= gn a library for different hardware (CPU and GPU)

For users, I want the i= nterface looks like this:


Blah<GPU, float> b();

or

Blah<CPU, double> c();


ideally inside the class of Bla= h, I want to do this to avoid duplication
and reduce workload:

template<DeviceType deviceUsed, typename DataType>
class Blah=
{
=C2=A0 =C2=A0 =C2=A0 DataType shared_member;


=C2=A0 =C2= =A0 =C2=A0 #Meta_if(deviceUsed =3D=3D GPU) =C2=A0 //some kind of meta progr= amming syntax
to enable code based on template
=C2=A0 =C2=A0 =C2=A0 = #{
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0DataType member_only_used_by= _gpu;
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0void function_on_gpu();=C2=A0 =C2=A0 =C2=A0 #}


=C2=A0 =C2=A0 =C2=A0 #Meta_if(deviceUse= d =3D=3D CPU)
=C2=A0 =C2=A0 =C2=A0 #{
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2= =A0 DataType member_only_used_by_cpu;
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0= void function_on_cpu();
=C2=A0 =C2=A0 =C2=A0 #}


=C2=A0= =C2=A0 =C2=A0 void shared_function()
=C2=A0 =C2=A0 =C2=A0 {
=C2=A0 = =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 ....
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2= =A0 =C2=A0 #Meta_if(deviceUsed =3D GPU) {
=C2=A0 =C2=A0 =C2=A0 =C2=A0 = =C2=A0 =C2=A0....
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 #}
=C2=A0= =C2=A0 =C2=A0 }
};


This is the actual code I have righ= t now.

as you can see, I have to use std::enable_if

I really do= n't like it, because it is a trick, a hacky solution, and only
work= s for functions, not for members or function internals.

and to make t= he following code to work, I had to try many options, some
don't co= mpile, some don't do what I want. And It's really hard to
under= stand this syntax:

template<DeviceType T =3D DeviceUsed, typename = Enabled =3D typename std::enable_if<(T & devices) !=3D 0> >


template<DeviceType T =3D DeviceUsed, typename =3D std::enable_i= f<(T & devices) !=3D 0>::type >


template<DeviceTy= pe T =3D DeviceUsed, typename std::enable_if<(T & devices) !=3D 0>= ;::type* =3D 0 >

And for class members, since std::enable_i= f can't be used, I have to put shared stuff into a base class. And this= is only manageable to do when I have few template variables.


=C2=A0 =C2=A0 =C2=A0 =C2=A0 typedef enum


=C2=A0 =C2= =A0 =C2=A0 =C2=A0 {


=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2= =A0 =C2=A0 CPU_NAIVE =C2=A0 =C2=A0 =C2=A0=3D 0x1,


=C2=A0 =C2=A0 = =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 CPU =C2=A0 =C2=A0 =C2=A0 =C2=A0 = =C2=A0 =C2=A0=3D 0x2,


=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 = =C2=A0 =C2=A0 CPU_SIMD =C2=A0 =C2=A0 =C2=A0 =3D 0x2,


=C2=A0 =C2= =A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 GPU =C2=A0 =C2=A0 =C2=A0 =C2= =A0 =C2=A0 =C2=A0=3D 0x4,


=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2= =A0 =C2=A0 =C2=A0 GPU_CUDA =C2=A0 =C2=A0 =C2=A0 =3D 0x4


=C2=A0 = =C2=A0 =C2=A0 =C2=A0 } DeviceType;


#define DEVICE_SPECIFIC= (devices) \


=C2=A0 =C2=A0 template<DeviceType T =3D DeviceUsed= , typename Enabled =3D typename std::enable_if<(T & devices) !=3D 0&= gt; >


=C2=A0
=C2=A0 =C2=A0template<typenam= e DataType =3D float>


=C2=A0 =C2=A0 class Blob_base


= =C2=A0 =C2=A0 {


=C2=A0 =C2=A0 protected:


=C2=A0 =C2=A0 = =C2=A0 =C2=A0 const std::string m_name;


=C2=A0 =C2=A0 =C2=A0 =C2= =A0 const unsigned int m_batchSize;


=C2=A0 =C2=A0 =C2=A0 =C2=A0 c= onst unsigned int m_depth;


=C2=A0 =C2=A0 =C2=A0 =C2=A0 const unsi= gned int m_height;


=C2=A0 =C2=A0 =C2=A0 =C2=A0 const unsigned int= m_width;


=C2=A0 =C2=A0 =C2=A0 =C2=A0 DataType *m_data;

=C2=A0 =C2=A0 };


=C2=A0 =C2=A0 =C2=A0 =C2=A0 template<= DeviceType DeviceUsed =3D CPU, typename DataType =3D float>


= =C2=A0 =C2=A0 class Blob : public Blob_base<DataType>


=C2= =A0 =C2=A0 =C2=A0 =C2=A0 {


=C2=A0 =C2=A0 public:

=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 Blob(const unsigne= d int batchSize,


=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2= =A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0const unsigned int depth,

<= p>
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0= =C2=A0 =C2=A0 =C2=A0const unsigned int height,


=C2=A0 =C2=A0 =C2= =A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0co= nst unsigned int width,


=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0= =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0const std::string &nam= e =3D "no_name");


=C2=A0 =C2=A0 =C2=A0 =C2=A0 bo= ol init();


=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2= =A0 =C2=A0 DEVICE_SPECIFIC(CPU | GPU)


=C2=A0 =C2=A0 =C2=A0 =C2=A0= =C2=A0 =C2=A0 =C2=A0 =C2=A0 void test();


=C2=A0 =C2=A0 =C2=A0 = =C2=A0 };


=C2=A0 =C2=A0 template<typename DataType><= /p>


=C2=A0 =C2=A0 class Blob<GPU, DataType> : public Blob_base&= lt;DataType>


=C2=A0 =C2=A0 {


=C2=A0 =C2=A0 public:


=C2=A0 =C2=A0 =C2=A0 =C2=A0 bool init();


=C2=A0 = =C2=A0 =C2=A0 =C2=A0 Blob(const unsigned int batchSize,


=C2=A0 = =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0const unsigned int depth,

=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0const unsigned int height= ,


=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0const unsigned = int width,


=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0const = std::string &name =3D "no_name");


=C2=A0 =C2=A0 };<= /p>


And because the limitation of partial specification of c++= (you have to
partially specify a class first before specifying its mem= ber functions), I
had to create many versions of the same class, with d= ifferent template<>
variables.

I have to duplicate lots of = code.

Also, I really want the flexibility to do this:

template&= lt;DeviceType device>
class
{
=C2=A0 =C2=A0 =C2=A0 #meta_if(dev= ice =3D=3D CPU || device =3D=3D CPU_SIMD)
=C2=A0 =C2=A0 =C2=A0 =C2=A0int= a,b;
=C2=A0 =C2=A0 =C2=A0 =C2=A0void function()
=C2=A0 =C2=A0 =C2=A0= =C2=A0{... };
=C2=A0 =C2=A0 =C2=A0 =C2=A0#meta_end
}


appar= ently using the base class still doesn't allow me to do this

I think many people share the same pain with me, if you google those
q= uestions about std::enable_if and how to specify a template member
func= tion.

So if meta programming is so useful, why don't you consider= adding it
formally to C++ by introducing the above syntax, rather than= standardizing
the partially working ugly hack std::enable_if?

PHP for example, can be mixed with html.

why can't c++ define s= omething similar:

<?c++ echo 'While this is going to be parsed= ..'; ?>

such that what's inside <?c++ ... ?> will be = executed by the preprocessor
and generate code?

<= /p>

=

=

<= p>

<= /p>

--
You received this message because you are subscribed to the Google Groups &= quot;ISO C++ Standard - Future Proposals" group.
To unsubscribe from this group and stop receiving emails from it, send an e= mail to std-proposa= ls+unsubscribe@isocpp.org.
To post to this group, send email to std-proposals@isocpp.org.
To view this discussion on the web visit https://groups.google.= com/a/isocpp.org/d/msgid/std-proposals/215ef486-c593-4a7d-bb89-d134b18a9fab= %40isocpp.org.
------=_Part_364_2086495820.1478774102700-- ------=_Part_363_184363760.1478774102699-- .