[fusion] memory layout of fusion::vector vs boost::tuple/boost::array

older
Re: [Boost-users] [serialization]...

alfC

6 Oct 2010 6 Oct '10

7:53 p.m.

Hi, I am trying to convert a boost::array into a fusion::vector. reinterpreting the memory from boost::array. It works for tuples but not fusion::vector, why is that. Is there a way to make it work? ( I was hoping that the memory layout is the same to make conversion from one to the other very easy.) #include #include #include #include<iostream> #include using std::clog; using std::endl; int main(){ boost::array ba={{1.0, 2.1, 3.2}}; clog << "ba: " << ba[0] << ", " << ba[1] << ", " << ba[2] << endl; using namespace boost::fusion; vector bfv( reinterpret_cast&>(ba) ); //doesn't work as expected clog << "bfv: " << at_c<0>(bfv) << ", "<< at_c<1>(bfv) << ", " << at_c<2>(bfv) << endl; using boost::tuple; tuple btt(reinterpret_cast&>(ba) ); // works! clog << "btt: " << btt.get<0>() << ", "<< (btt).get<1>() << ", " << (btt).get<2>() << endl; return 0; } outputs: ba: 1, 2.1, 3.2 //original data bfv: -9.25597e+61, -2.35344e-185, 4.87507e-270 //not expected btt: 1, 2.1, 3.2 //works! Let me give some context, the reason I am doing this is that I need to interface fusion::vector of boost::quantities (all have the sizeof double) and adimensionalize the list of quantities into an array of doubles and vise-versa In reality, the final goal is to get this to work: boost::array ba={{1.0, 2.1, 3.2}}; vectorsi::length, quantitysi::time , quantitysi::mass > bfv( reinterpret_castsi::length, quantitysi::time , quantitysi::mass >&>(ba) ); (or other generic option to go from array to vector) but since I can't make it work with simple 'doubles' I am asking why this could be done. Thank you again, Alfredo

Show replies by date

Joel de Guzman

7 Oct 7 Oct

12:33 a.m.

On 10/7/2010 3:53 AM, alfC wrote:

...

Hi,

I am trying to convert a boost::array into a fusion::vector. reinterpreting the memory from boost::array. It works for tuples but not fusion::vector, why is that. Is there a way to make it work? ( I was hoping that the memory layout is the same to make conversion from one to the other very easy.)

Don't do that. It will *never* be guaranteed to work even if it works now. The memory layout of fusion::vector is an internal implementation detail and can change anytime. Regards, -- Joel de Guzman http://www.boostpro.com http://spirit.sf.net

alfC

4:54 a.m.

On Oct 6, 5:33 pm, Joel de Guzman wrote:

...

On 10/7/2010 3:53 AM, alfC wrote:

...
Hi,

...
I am trying to convert a boost::array into a fusion::vector. reinterpreting the memory from boost::array. It works for tuples but not fusion::vector, why is that. Is there a way to make it work? ( I was hoping that the memory layout is the same to make conversion from one to the other very easy.)

Don't do that. It will *never* be guaranteed to work even if it works now. The memory layout of fusion::vector is an internal implementation detail and can change anytime.

It is not only not guaranteed, it does not work at present. Any two cents why is that fusion::vector may (do?) implement a non obvious memory layout? I solved the problem of array to vector: I found the feature of the library that allows conversion from array to vector. Just by including #include #include boost::array ba={{1.0, 2.1, 3.2}}; vector bfv1(ba); //works. I still can the other way around. boost::array ba0(bfv1); //doesn't work ideas? great, I also managed to solve the original problem on how to convert from boost::array to fusion::vector thanks to your other posting suggesting using fusion::transform quantity<T> can be constructed from a double with quantity<T>::from_value(double) The line of code that does this is typedef boost::fusion::vectorsi::length, quantitysi::time, quantitysi::mass > vector_of_q; vector_of_q fvq( transform( ba, // boost::array // incld"/fusion/array.hpp"!! vector_of_q(), from_value() ) ); where "from_value" is defined by struct from_value{ // binary fusion transformation template <typename SignatureT> struct result; template <typename Q> struct result{ typedef Q type; }; template<typename Q> Q operator()(double const& value, Q const&) const{ return Q::from_value(value); } }; I had to use the *binary* version of transform because the type of the result depends on the destination fusion::vector. The other way around from vector to array should be easier (because the destination type is always double by using the quantity<T>::value() ) but I didn't managed because I didn't find the way to convert from vector to array. ideas? Thank you for making fusion, it is an incredible library once you realize what it is for. Alfredo

OvermindDL1

10:33 p.m.

On Wed, Oct 6, 2010 at 10:54 PM, alfC wrote:

...

On Oct 6, 5:33 pm, Joel de Guzman wrote:

...
On 10/7/2010 3:53 AM, alfC wrote:

...
Hi,

...
I am trying to convert a boost::array into a fusion::vector. reinterpreting the memory from boost::array. It works for tuples but not fusion::vector, why is that. Is there a way to make it work? ( I was hoping that the memory layout is the same to make conversion from one to the other very easy.)

Don't do that. It will *never* be guaranteed to work even if it works now. The memory layout of fusion::vector is an internal implementation detail and can change anytime.

It is not only not guaranteed, it does not work at present. Any two cents why is that fusion::vector may (do?) implement a non obvious memory layout?

I solved the problem of array to vector: I found the feature of the library that allows conversion from array to vector. Just by including #include #include

boost::array ba={{1.0, 2.1, 3.2}}; vector bfv1(ba); //works.

I still can the other way around. boost::array ba0(bfv1); //doesn't work ideas?

great, I also managed to solve the original problem on how to convert from boost::array to fusion::vector thanks to your other posting suggesting using fusion::transform

quantity<T> can be constructed from a double with quantity<T>::from_value(double)

The line of code that does this is

typedef boost::fusion::vectorsi::length, quantitysi::time, quantitysi::mass > vector_of_q; vector_of_q fvq( transform( ba, // boost::array // incld"/fusion/array.hpp"!! vector_of_q(), from_value() ) );

where "from_value" is defined by

struct from_value{ // binary fusion transformation template <typename SignatureT> struct result; template <typename Q> struct result{ typedef Q type; }; template<typename Q> Q operator()(double const& value, Q const&) const{ return Q::from_value(value); } };

I had to use the *binary* version of transform because the type of the result depends on the destination fusion::vector.

The other way around from vector to array should be easier (because the destination type is always double by using the quantity<T>::value() ) but I didn't managed because I didn't find the way to convert from vector to array. ideas?

Thank you for making fusion, it is an incredible library once you realize what it is for. Alfredo

You do know that based on this: http://www.boost.org/doc/libs/1_44_0/libs/fusion/doc/html/fusion/adapted/boo... boost::array is already adapted as a fusion container and can be used directly, as soon as you include the proper hpp that is?

Alfredo Correa

11:42 p.m.

On Thu, Oct 7, 2010 at 3:33 PM, OvermindDL1 wrote:

...

On Wed, Oct 6, 2010 at 10:54 PM, alfC wrote:

...
On Oct 6, 5:33 pm, Joel de Guzman wrote:

...
On 10/7/2010 3:53 AM, alfC wrote:

...
Hi,

...
I am trying to convert a boost::array into a fusion::vector. reinterpreting the memory from boost::array. It works for tuples but not fusion::vector, why is that. Is there a way to make it work? ( I was hoping that the memory layout is the same to make conversion from one to the other very easy.)

Don't do that. It will *never* be guaranteed to work even if it works now. The memory layout of fusion::vector is an internal implementation detail and can change anytime.

It is not only not guaranteed, it does not work at present. Any two cents why is that fusion::vector may (do?) implement a non obvious memory layout?

I solved the problem of array to vector: I found the feature of the library that allows conversion from array to vector. Just by including #include #include

boost::array ba={{1.0, 2.1, 3.2}}; vector bfv1(ba); //works.

I still can the other way around. boost::array ba0(bfv1); //doesn't work ideas?

great, I also managed to solve the original problem on how to convert from boost::array to fusion::vector thanks to your other posting suggesting using fusion::transform

quantity<T> can be constructed from a double with quantity<T>::from_value(double)

The line of code that does this is

typedef boost::fusion::vectorsi::length, quantitysi::time, quantitysi::mass > vector_of_q; vector_of_q fvq( transform( ba, // boost::array //

incld"/fusion/array.hpp"!!

...
vector_of_q(), from_value() ) );

where "from_value" is defined by

struct from_value{ // binary fusion transformation template <typename SignatureT> struct result; template <typename Q> struct result{ typedef Q type; }; template<typename Q> Q operator()(double const& value, Q const&) const{ return Q::from_value(value); } };

I had to use the *binary* version of transform because the type of the result depends on the destination fusion::vector.

The other way around from vector to array should be easier (because the destination type is always double by using the quantity<T>::value() ) but I didn't managed because I didn't find the way to convert from vector to array. ideas?

Thank you for making fusion, it is an incredible library once you realize what it is for. Alfredo

You do know that based on this:

http://www.boost.org/doc/libs/1_44_0/libs/fusion/doc/html/fusion/adapted/boo... boost::array is already adapted as a fusion container and can be used directly, as soon as you include the proper hpp that is?

yes, thank you, that what I wrote in my last post . I can make this work #include #include ... boost::array boostarr={{1.0, 2.1, 3.2}}; vector fusionvec(boostarr); //now works I still can't do the other way around: boost::array boostarr2(fusionvec); //doesn't work ideas? _______________________________________________

...

Boost-users mailing list Boost-users@lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

-- Sie haben diese Nachricht erhalten, da Sie der Google Groups-Gruppe Boost Users beigetreten sind. Wenn Sie Nachrichten in dieser Gruppe posten möchten, senden Sie eine E-Mail an boostusers@googlegroups.com. Wenn Sie aus dieser Gruppe austreten möchten, senden Sie eine E-Mail an boostusers+unsubscribe@googlegroups.com . Besuchen Sie die Gruppe unter http://groups.google.com/group/boostusers?hl=de, um weitere Optionen zu erhalten.

OvermindDL1

8 Oct 8 Oct

12:13 a.m.

On Thu, Oct 7, 2010 at 5:42 PM, Alfredo Correa wrote:

...

On Thu, Oct 7, 2010 at 3:33 PM, OvermindDL1 wrote:

...
You do know that based on this:

http://www.boost.org/doc/libs/1_44_0/libs/fusion/doc/html/fusion/adapted/boo... boost::array is already adapted as a fusion container and can be used directly, as soon as you include the proper hpp that is?

yes, thank you, that what I wrote in my last post . I can make this work

#include #include ... boost::array boostarr={{1.0, 2.1, 3.2}}; vector fusionvec(boostarr); //now works

Ah, heh, got lost in the noise and I missed it. On Thu, Oct 7, 2010 at 5:42 PM, Alfredo Correa wrote:

...

I still can't do the other way around:

boost::array boostarr2(fusionvec); //doesn't work ideas?

I know that would not work because that is still a 'custom' type as far as fusion is concerned, and only free functions can work on it, a constructor is obviously not a free function, so you need to write a free function to do it (probably called 'copy', probably using fusion's foreach to copy each element one to the other, should be pretty well optimized out thanks to the template-magic).

alfC

6:59 a.m.

On Oct 7, 5:13 pm, OvermindDL1 wrote:

...

On Thu, Oct 7, 2010 at 5:42 PM, Alfredo Correa wrote:

...
On Thu, Oct 7, 2010 at 3:33 PM, OvermindDL1 wrote:

...
You do know that based on this:

...
...
http://www.boost.org/doc/libs/1_44_0/libs/fusion/doc/html/fusion/adap... boost::array is already adapted as a fusion container and can be used directly, as soon as you include the proper hpp that is?

...
yes, thank you, that what I wrote in my last post . I can make this work

...
#include #include ... boost::array boostarr={{1.0, 2.1, 3.2}}; vector fusionvec(boostarr); //now works

Ah, heh, got lost in the noise and I missed it.

On Thu, Oct 7, 2010 at 5:42 PM, Alfredo Correa wrote:

...
I still can't do the other way around:

...
boost::array boostarr2(fusionvec); //doesn't work ideas?

I know that would not work because that is still a 'custom' type as far as fusion is concerned, and only free functions can work on it, a constructor is obviously not a free function, so you need to write a free function to do it (probably called 'copy', probably using fusion's foreach to copy each element one to the other, should be pretty well optimized out thanks to the template-magic).

no luck. I can't make it work. First I tried with "for_each" and didn't manage to put together the code because I need to copy one sequence from the other and the for_each doesn't keep track of the index. So I tried with "transform", the code compiles but does nothing to the array. boost::array ba0; boost::fusion::transform(ba0, bfv1, copy()); where struct copy{ typedef void result_type; void operator()(double& lhs, double const& rhs) const{ lhs = rhs; } }; ba0 is still undefined after running this. Obviously I am getting Boost.Fusion wrong. Thanks, Alfredo

...

_______________________________________________ Boost-users mailing list Boost-us...@lists.boost.orghttp://lists.boost.org/mailman/listinfo.cgi/boost-users

Thomas Heller

7:44 a.m.

alfC wrote:

...

On Oct 7, 5:13 pm, OvermindDL1 wrote:

...
On Thu, Oct 7, 2010 at 5:42 PM, Alfredo Correa wrote:

...
On Thu, Oct 7, 2010 at 3:33 PM, OvermindDL1 wrote:

...
You do know that based on this:

...
...

http://www.boost.org/doc/libs/1_44_0/libs/fusion/doc/html/fusion/adap...

...

...
...
...
boost::array is already adapted as a fusion container and can be used directly, as soon as you include the proper hpp that is?

...
yes, thank you, that what I wrote in my last post . I can make this work

...
#include #include ... boost::array boostarr={{1.0, 2.1, 3.2}}; vector fusionvec(boostarr); //now works

Ah, heh, got lost in the noise and I missed it.

On Thu, Oct 7, 2010 at 5:42 PM, Alfredo Correa wrote:

...
I still can't do the other way around:

...
boost::array boostarr2(fusionvec); //doesn't work ideas?

I know that would not work because that is still a 'custom' type as far as fusion is concerned, and only free functions can work on it, a constructor is obviously not a free function, so you need to write a free function to do it (probably called 'copy', probably using fusion's foreach to copy each element one to the other, should be pretty well optimized out thanks to the template-magic).

no luck. I can't make it work. First I tried with "for_each" and didn't manage to put together the code because I need to copy one sequence from the other and the for_each doesn't keep track of the index. So I tried with "transform", the code compiles but does nothing to the array.

boost::array ba0; boost::fusion::transform(ba0, bfv1, copy());

where struct copy{ typedef void result_type; void operator()(double& lhs, double const& rhs) const{ lhs = rhs; } };

ba0 is still undefined after running this.

Obviously I am getting Boost.Fusion wrong.

Thanks, Alfredo

You will find a version using fold attached. What it does, is that it increases the size of the boost::array passed in as state by one for every element in the fusion sequence and copies over the elements. There could be a more elegant solution to the problem ;) Cheers

Alfredo Correa

6:29 p.m.

2010/10/8 Thomas Heller

...

alfC wrote:

...
On Oct 7, 5:13 pm, OvermindDL1 wrote:

...
On Thu, Oct 7, 2010 at 5:42 PM, Alfredo Correa <

alfredo.cor...@gmail.com>

...
...
wrote:

...
On Thu, Oct 7, 2010 at 3:33 PM, OvermindDL1 wrote:

...
You do know that based on this:

...
...
http://www.boost.org/doc/libs/1_44_0/libs/fusion/doc/html/fusion/adap...

...
boost::array is already adapted as a fusion container and can be used directly, as soon as you include the proper hpp that is?

...
yes, thank you, that what I wrote in my last post . I can make this work

...
#include #include ... boost::array boostarr={{1.0, 2.1, 3.2}}; vector fusionvec(boostarr); //now works

Ah, heh, got lost in the noise and I missed it.

On Thu, Oct 7, 2010 at 5:42 PM, Alfredo Correa < alfredo.cor...@gmail.com> wrote:

...
I still can't do the other way around:

...
boost::array boostarr2(fusionvec); //doesn't work ideas?

I know that would not work because that is still a 'custom' type as far as fusion is concerned, and only free functions can work on it, a constructor is obviously not a free function, so you need to write a free function to do it (probably called 'copy', probably using fusion's foreach to copy each element one to the other, should be pretty well optimized out thanks to the template-magic).

no luck. I can't make it work. First I tried with "for_each" and didn't manage to put together the code because I need to copy one sequence from the other and the for_each doesn't keep track of the index. So I tried with "transform", the code compiles but does nothing to the array.

boost::array ba0; boost::fusion::transform(ba0, bfv1, copy());

where struct copy{ typedef void result_type; void operator()(double& lhs, double const& rhs) const{ lhs = rhs; } };

ba0 is still undefined after running this.

Obviously I am getting Boost.Fusion wrong.

Thanks, Alfredo

You will find a version using fold attached. What it does, is that it increases the size of the boost::array passed in as state by one for every element in the fusion sequence and copies over the elements. There could be a more elegant solution to the problem ;)

Thomas, thanks so much for the code. Unfortunately it doesn't work (maybe it is my compiler?) $ c++ -v gcc 4.4.3 $ c++ fusion_test.cpp $ ./a.out terminate called after throwing an instance of 'boost::exception_detail::clone_implstd::out_of_range

...

' what(): attempt to access element of an empty array Aborted

Thanks, Alfredo

...

Cheers

_______________________________________________ Boost-users mailing list Boost-users@lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users -- Sie haben diese Nachricht erhalten, da Sie der Google Groups-Gruppe Boost Users beigetreten sind. Wenn Sie Nachrichten in dieser Gruppe posten möchten, senden Sie eine E-Mail an boostusers@googlegroups.com. Wenn Sie aus dieser Gruppe austreten möchten, senden Sie eine E-Mail an boostusers+unsubscribe@googlegroups.com . Besuchen Sie die Gruppe unter http://groups.google.com/group/boostusers?hl=de, um weitere Optionen zu erhalten.

alfC

6:44 p.m.

On Oct 6, 5:33 pm, Joel de Guzman wrote:

...

On 10/7/2010 3:53 AM, alfC wrote:

...
Hi,

...
I am trying to convert a boost::array into a fusion::vector. reinterpreting the memory from boost::array. It works for tuples but not fusion::vector, why is that. Is there a way to make it work? ( I was hoping that the memory layout is the same to make conversion from one to the other very easy.)

Don't do that. It will *never* be guaranteed to work even if it works now. The memory layout of fusion::vector is an internal implementation detail and can change anytime.

The question is: is it guaranteed for boost::tuples? (it seems so, just by the PODness of tuples of PODs in its stated design.) If so, then it is a *feature* of boost::tuple. Fusion, which is supposed to be an extension of boost::tuple (or at least that was my understanding), lost such feature that I could exploit. The situation is the following: quantity because it claims that has no runtime cost, this is like saying that it can be interpreted as its underlying type T (e.g. double). All numerical routines that I am using take fixed arrays of double or basically boost::arrays. The analogy is perfect if I use tuples because I can reinterpret them as arrays back *and* forth without copying anything or adding any additional code. Since I can't do this with fusion::vector then I will have to use boost::tuples. This will be a pain in it self because then I will have to manually code MPL to tuples generation and loose the nice fusion::transforms. It seem I am out of luck, and just fallen in a crack between tuples and fusion. There is second possibility that is that I got the boost::fusion theory totally wrong in the first place. Thanks so much for your help, Alfredo.

...

Regards, -- Joel de Guzmanhttp://www.boostpro.comhttp://spirit.sf.net

_______________________________________________ Boost-users mailing list Boost-us...@lists.boost.orghttp://lists.boost.org/mailman/listinfo.cgi/boost-users

Joel de Guzman

9 Oct 9 Oct

12:57 a.m.

On 10/9/2010 2:44 AM, alfC wrote:

...

On Oct 6, 5:33 pm, Joel de Guzman wrote:

...
On 10/7/2010 3:53 AM, alfC wrote:

...
Hi,

...
I am trying to convert a boost::array into a fusion::vector. reinterpreting the memory from boost::array. It works for tuples but not fusion::vector, why is that. Is there a way to make it work? ( I was hoping that the memory layout is the same to make conversion from one to the other very easy.)

Don't do that. It will *never* be guaranteed to work even if it works now. The memory layout of fusion::vector is an internal implementation detail and can change anytime.

The question is: is it guaranteed for boost::tuples? (it seems so, just by the PODness of tuples of PODs in its stated design.) If so, then it is a *feature* of boost::tuple.

No it is not. 1) It is not documented 2) It is not a POD 3) Anything with a reinterpret_cast is not guaranteed to work (you can only rely on reinterpret_cast only when casting back from a lost type e.g. through type erasure). Disregarding any of that is playing with fire. Regards, -- Joel de Guzman http://www.boostpro.com http://spirit.sf.net

OvermindDL1

3 a.m.

On Fri, Oct 8, 2010 at 6:57 PM, Joel de Guzman wrote:

...

On 10/9/2010 2:44 AM, alfC wrote:

...
On Oct 6, 5:33 pm, Joel de Guzman wrote:

...
On 10/7/2010 3:53 AM, alfC wrote:

...
Hi,

...
I am trying to convert a boost::array into a fusion::vector. reinterpreting the memory from boost::array. It works for tuples but not fusion::vector, why is that. Is there a way to make it work? ( I was hoping that the memory layout is the same to make conversion from one to the other very easy.)

Don't do that. It will *never* be guaranteed to work even if it works now. The memory layout of fusion::vector is an internal implementation detail and can change anytime.

The question is: is it guaranteed for boost::tuples? (it seems so, just by the PODness of tuples of PODs in its stated design.) If so, then it is a *feature* of boost::tuple.

No it is not. 1) It is not documented 2) It is not a POD 3) Anything with a reinterpret_cast is not guaranteed to work (you can only rely on reinterpret_cast only when casting back from a lost type e.g. through type erasure).

Disregarding any of that is playing with fire.

The previous posts a few posts ago seems overly complicated, I would just do this (suitably wrapped into a function to handle it, maybe named copy): // Proper includes here... struct set1stTo2nd { template<typename T> void operator()(T& t) const { using namespace boost::fusion; deref(begin(t)) = deref(next(begin(t))); } }; typedef boost::array arrType; typedef boost::fusion::vector vecType; typedef boost::fusion::vector zipSeqType; arrType arr(1,1,1); vecType vec(5,5,5); for_each(zipSeqType(arr,vec), set1stTo2nd()); assert(arr == make_vector(5,5,5)); Nice and simple; there might be a better way, but these kind of things suit me perfectly, and who knows, there might be a Phoenix way to do set1stTo2nd instead of the functor.

alfC

9:24 a.m.

On Oct 8, 8:00 pm, OvermindDL1 wrote:

...

On Fri, Oct 8, 2010 at 6:57 PM, Joel de Guzman

wrote:

...
On 10/9/2010 2:44 AM, alfC wrote:

...
...
On Oct 6, 5:33 pm, Joel de Guzman wrote:

...
...
...
On 10/7/2010 3:53 AM, alfC wrote:

...
...
...
...
Hi,

...
...
...
...
I am trying to convert a boost::array into a fusion::vector. reinterpreting the memory from boost::array. It works for tuples but not fusion::vector, why is that. Is there a way to make it work? ( I was hoping that the memory layout is the same to make conversion from one to the other very easy.)

...
...
...
Don't do that. It will *never* be guaranteed to work even if it works now. The memory layout of fusion::vector is an internal implementation detail and can change anytime.

...
...
The question is: is it guaranteed for boost::tuples? (it seems so, just by the PODness of tuples of PODs in its stated design.) If so, then it is a *feature* of boost::tuple.

...
No it is not. 1) It is not documented 2) It is not a POD 3) Anything with a reinterpret_cast is not guaranteed to work (you can only rely on reinterpret_cast only when casting back from a lost type e.g. through type erasure).

...
Disregarding any of that is playing with fire.

The previous posts a few posts ago seems overly complicated, I would just do this (suitably wrapped into a function to handle it, maybe named copy):

// Proper includes here...

struct set1stTo2nd { template<typename T> void operator()(T& t) const { using namespace boost::fusion; deref(begin(t)) = deref(next(begin(t))); }};

typedef boost::array arrType; typedef boost::fusion::vector vecType; typedef boost::fusion::vector zipSeqType;

arrType arr(1,1,1); vecType vec(5,5,5);

for_each(zipSeqType(arr,vec), set1stTo2nd()); assert(arr == make_vector(5,5,5));

Thank you very much for the code. Unfortunately it doesn't work, compiles but it doesn't copy the vector into the array. Below is the full program, it leaves the array intact. It doesn't make sense either, the for_each argument has only two elements, the array and the vector. The for_each will just first "do" the array and then "do" the vector, but it just seems to assign the first element of each to the second but not copying values from the vector to the array. (I still remember the time when I could copy one an array of doubles to the other.) Below is the full program if you want to give it a try: #include #include #include #include #include #include #include #include<iostream> using std::clog; using std::endl; using namespace boost::fusion; struct set1stTo2nd { template<typename T> void operator()(T& t) const { deref(boost::fusion::begin(t)) = deref(boost::fusion::next(boost::fusion::begin(t))); } }; typedef boost::array arrType; typedef boost::fusion::vector vecType; typedef boost::fusion::vector zipSeqType; int main(){ boost::array arr1={{1.0, 2.1, 3.2}}; clog << "arr1: " << arr1[0] << ", " << arr1[1] << ", " << arr1[2] << endl; boost::fusion::vector vec1(arr1); clog << "vec1: " << at_c<0>(vec1) << ", "<< at_c<1>(vec1) << ", " << at_c<2>(vec1) << endl; boost::array arr2; for_each(zipSeqType(arr2,vec1), set1stTo2nd()); clog << "arr2: " << at_c<0>(arr2) << ", "<< at_c<1>(arr2) << ", " << at_c<2>(arr2) << endl; return 0; } prints: arr1: 1, 2.1, 3.2 vec1: 1, 2.1, 3.2 arr2: 4.85589e-270, 4.85589e-270, 4.90963e-270

Christopher Schmidt

9:46 a.m.

Am 09.10.2010 11:24, schrieb alfC:

...

On Oct 8, 8:00 pm, OvermindDL1 wrote:

[snip]

...

...
The previous posts a few posts ago seems overly complicated, I would just do this (suitably wrapped into a function to handle it, maybe named copy):

// Proper includes here...

struct set1stTo2nd { template<typename T> void operator()(T& t) const { using namespace boost::fusion; deref(begin(t)) = deref(next(begin(t))); }};

typedef boost::array arrType; typedef boost::fusion::vector vecType; typedef boost::fusion::vector zipSeqType;

arrType arr(1,1,1); vecType vec(5,5,5);

for_each(zipSeqType(arr,vec), set1stTo2nd()); assert(arr == make_vector(5,5,5));

Thank you very much for the code. Unfortunately it doesn't work, compiles but it doesn't copy the vector into the array. Below is the full program, it leaves the array intact.

It doesn't make sense either, the for_each argument has only two elements, the array and the vector. The for_each will just first "do" the array and then "do" the vector, but it just seems to assign the first element of each to the second but not copying values from the vector to the array.

(I still remember the time when I could copy one an array of doubles to the other.) Below is the full program if you want to give it a try:

#include #include #include #include #include #include #include #include<iostream> using std::clog; using std::endl;

using namespace boost::fusion; struct set1stTo2nd { template<typename T> void operator()(T& t) const { deref(boost::fusion::begin(t)) = deref(boost::fusion::next(boost::fusion::begin(t))); } };

typedef boost::array arrType; typedef boost::fusion::vector vecType; typedef boost::fusion::vector zipSeqType;

int main(){ boost::array arr1={{1.0, 2.1, 3.2}}; clog << "arr1: " << arr1[0] << ", " << arr1[1] << ", " << arr1[2] << endl;

boost::fusion::vector vec1(arr1); clog << "vec1: " << at_c<0>(vec1) << ", "<< at_c<1>(vec1) << ", " << at_c<2>(vec1) << endl;

boost::array arr2; for_each(zipSeqType(arr2,vec1), set1stTo2nd());

fusion::zip(arr2,vec1) should be used here, or fusion::zip_view<zipSeqType>(zipSeqType(arr2,vec1)) . http://www.boost.org/doc/libs/1_44_0/libs/fusion/doc/html/fusion/algorithm/t... http://www.boost.org/doc/libs/1_44_0/libs/fusion/doc/html/fusion/view/zip_vi...

...

clog << "arr2: " << at_c<0>(arr2) << ", "<< at_c<1>(arr2) << ", " << at_c<2>(arr2) << endl; return 0; }

prints: arr1: 1, 2.1, 3.2 vec1: 1, 2.1, 3.2 arr2: 4.85589e-270, 4.85589e-270, 4.90963e-270

-Christopher

alfC

10:11 a.m.

On Oct 9, 2:24 am, alfC wrote:

...

On Oct 8, 8:00 pm, OvermindDL1 wrote:

...
On Fri, Oct 8, 2010 at 6:57 PM, Joel de Guzman

...
wrote:

...
On 10/9/2010 2:44 AM, alfC wrote:

...
...
...
On Oct 6, 5:33 pm, Joel de Guzman wrote:

...
...
...
...
On 10/7/2010 3:53 AM, alfC wrote:

...
...
...
...
...
Hi,

...
...
...
...
...
I am trying to convert a boost::array into a fusion::vector. reinterpreting the memory from boost::array. It works for tuples but not fusion::vector, why is that. Is there a way to make it work? ( I was hoping that the memory layout is the same to make conversion from one to the other very easy.)

...
...
...
...
Don't do that. It will *never* be guaranteed to work even if it works now. The memory layout of fusion::vector is an internal implementation detail and can change anytime.

...
...
...
The question is: is it guaranteed for boost::tuples? (it seems so, just by the PODness of tuples of PODs in its stated design.) If so, then it is a *feature* of boost::tuple.

...
...
No it is not. 1) It is not documented 2) It is not a POD 3) Anything with a reinterpret_cast is not guaranteed to work (you can only rely on reinterpret_cast only when casting back from a lost type e.g. through type erasure).

...
...
Disregarding any of that is playing with fire.

...
The previous posts a few posts ago seems overly complicated, I would just do this (suitably wrapped into a function to handle it, maybe named copy):

...
// Proper includes here...

...
struct set1stTo2nd { template<typename T> void operator()(T& t) const { using namespace boost::fusion; deref(begin(t)) = deref(next(begin(t))); }};

...
typedef boost::array arrType; typedef boost::fusion::vector vecType; typedef boost::fusion::vector zipSeqType;

...
arrType arr(1,1,1); vecType vec(5,5,5);

...
for_each(zipSeqType(arr,vec), set1stTo2nd()); assert(arr == make_vector(5,5,5));

FINALLY! I got it. I feel better now. It turns out that the correct line of code is for_each( boost::fusion::zip_view<zipSeqType>(zipSeqType(arr2,vec1)), set1stTo2nd() ); you wrote the correct typedef but then didn't apply the zip_view function. (Although, related to the original question, I am still a bit worried about too much unnecessary copy which seems unavoidable since the memory layout of array and fusion::vector is different.) below is the full program that *works*: #include #include #include #include #include #include #include #include #include #include #include #include<iostream> using std::clog; using std::endl; using namespace boost::fusion; struct set1stTo2nd { template<typename T> void operator()(T t) const { deref(boost::fusion::begin(t)) = deref(boost::fusion::next(boost::fusion::begin(t))); } }; typedef boost::array arrType; typedef boost::fusion::vector vecType; typedef boost::fusion::vector zipSeqType; int main(){ boost::array arr1={{1.0, 2.1, 3.2}}; clog << "arr1: " << arr1[0] << ", " << arr1[1] << ", " << arr1[2] << endl; boost::fusion::vector vec1(arr1); clog << "vec1: " << at_c<0>(vec1) << ", "<< at_c<1>(vec1) << ", " << at_c<2>(vec1) << endl; boost::array arr2; for_each( boost::fusion::zip_view<zipSeqType>(zipSeqType(arr2,vec1)), set1stTo2nd() ); clog << "arr2: " << at_c<0>(arr2) << ", "<< at_c<1>(arr2) << ", " << at_c<2>(arr2) << endl; return 0; } prints: arr1: 1, 2.1, 3.2 vec1: 1, 2.1, 3.2 arr2: 1, 2.1, 3.2 Have a good weekend, Alfredo

Joel de Guzman

2:42 p.m.

On 10/9/2010 6:11 PM, alfC wrote:

...

...
...
The previous posts a few posts ago seems overly complicated, I would just do this (suitably wrapped into a function to handle it, maybe named copy):

Yep. We need fusion to have two algorithms: copy and copy_if. Any takers? Should be easy. Regards, -- Joel de Guzman http://www.boostpro.com http://spirit.sf.net

Steven Watanabe

4 p.m.

New subject: [fusion] memory layout of fusion::vector vs boost::tuple/boost::array

AMDG On 10/9/2010 7:42 AM, Joel de Guzman wrote:

...

On 10/9/2010 6:11 PM, alfC wrote:

...
...
...
The previous posts a few posts ago seems overly complicated, I would just do this (suitably wrapped into a function to handle it, maybe named copy):

Yep. We need fusion to have two algorithms: copy and copy_if. Any takers? Should be easy.

I have an implementation of copy somewhere, if I can find it. In Christ, Steven Watanabe

Joel de Guzman

10:58 p.m.

On 10/10/2010 12:00 AM, Steven Watanabe wrote:

...

AMDG

On 10/9/2010 7:42 AM, Joel de Guzman wrote:

...
On 10/9/2010 6:11 PM, alfC wrote:

...
...
...
The previous posts a few posts ago seems overly complicated, I would just do this (suitably wrapped into a function to handle it, maybe named copy):

Yep. We need fusion to have two algorithms: copy and copy_if. Any takers? Should be easy.

I have an implementation of copy somewhere, if I can find it.

Cool. I'd appreciate that, Steven. Regards, -- Joel de Guzman http://www.boostpro.com http://spirit.sf.net

Nat Goodspeed

14 Oct 14 Oct

3:55 a.m.

On Oct 9, 2010, at 12:00 PM, Steven Watanabe wrote:

...

AMDG

On 10/9/2010 7:42 AM, Joel de Guzman wrote:

...
Yep. We need fusion to have two algorithms: copy and copy_if. Any takers? Should be easy.

I have an implementation of copy somewhere, if I can find it.

Hence, copy...if. ;-)

...

OvermindDL1

9 Oct 9 Oct

2:53 p.m.

On Sat, Oct 9, 2010 at 4:11 AM, alfC wrote:

...

On Oct 9, 2:24 am, alfC wrote:

...
On Oct 8, 8:00 pm, OvermindDL1 wrote:

...
On Fri, Oct 8, 2010 at 6:57 PM, Joel de Guzman

...
wrote:

...
On 10/9/2010 2:44 AM, alfC wrote:

...
...
...
On Oct 6, 5:33 pm, Joel de Guzman wrote:

...
...
...
...
On 10/7/2010 3:53 AM, alfC wrote:

...
...
...
...
> Hi,

...
...
...
...
> I am trying to convert a boost::array into a fusion::vector. > reinterpreting the memory from boost::array. It works for tuples but > not fusion::vector, why is that. Is there a way to make it work? ( I > was hoping that the memory layout is the same to make conversion from > one to the other very easy.)

...
...
...
...
Don't do that. It will *never* be guaranteed to work even if it works now. The memory layout of fusion::vector is an internal implementation detail and can change anytime.

...
...
...
The question is: is it guaranteed for boost::tuples? (it seems so, just by the PODness of tuples of PODs in its stated design.) If so, then it is a *feature* of boost::tuple.

...
...
No it is not. 1) It is not documented 2) It is not a POD 3) Anything with a reinterpret_cast is not guaranteed to work (you can only rely on reinterpret_cast only when casting back from a lost type e.g. through type erasure).

...
...
Disregarding any of that is playing with fire.

...
The previous posts a few posts ago seems overly complicated, I would just do this (suitably wrapped into a function to handle it, maybe named copy):

...
// Proper includes here...

...
struct set1stTo2nd { template<typename T> void operator()(T& t) const { using namespace boost::fusion; deref(begin(t)) = deref(next(begin(t))); }};

...
typedef boost::array arrType; typedef boost::fusion::vector vecType; typedef boost::fusion::vector zipSeqType;

...
arrType arr(1,1,1); vecType vec(5,5,5);

...
for_each(zipSeqType(arr,vec), set1stTo2nd()); assert(arr == make_vector(5,5,5));

FINALLY! I got it. I feel better now.

It turns out that the correct line of code is

for_each( boost::fusion::zip_view<zipSeqType>(zipSeqType(arr2,vec1)), set1stTo2nd() );

you wrote the correct typedef but then didn't apply the zip_view function.

(Although, related to the original question, I am still a bit worried about too much unnecessary copy which seems unavoidable since the memory layout of array and fusion::vector is different.)

below is the full program that *works*:

#include #include #include #include #include #include #include #include #include #include #include

#include<iostream> using std::clog; using std::endl;

using namespace boost::fusion; struct set1stTo2nd { template<typename T> void operator()(T t) const { deref(boost::fusion::begin(t)) = deref(boost::fusion::next(boost::fusion::begin(t))); } };

typedef boost::array arrType; typedef boost::fusion::vector vecType; typedef boost::fusion::vector zipSeqType;

int main(){ boost::array arr1={{1.0, 2.1, 3.2}}; clog << "arr1: " << arr1[0] << ", " << arr1[1] << ", " << arr1[2] << endl;

boost::fusion::vector vec1(arr1); clog << "vec1: " << at_c<0>(vec1) << ", "<< at_c<1>(vec1) << ", " << at_c<2>(vec1) << endl;

boost::array arr2; for_each( boost::fusion::zip_view<zipSeqType>(zipSeqType(arr2,vec1)), set1stTo2nd() ); clog << "arr2: " << at_c<0>(arr2) << ", "<< at_c<1>(arr2) << ", " << at_c<2>(arr2) << endl; return 0; }

prints: arr1: 1, 2.1, 3.2 vec1: 1, 2.1, 3.2 arr2: 1, 2.1, 3.2

Have a good weekend, Alfredo

Yeah, my mistake, busy at work and not paying attention, I meant to have the fusion::zip there. :)

Alfredo Correa

6:14 p.m.

On Sat, Oct 9, 2010 at 7:53 AM, OvermindDL1 wrote:

...

On Sat, Oct 9, 2010 at 4:11 AM, alfC wrote:

...
On Oct 9, 2:24 am, alfC wrote:

...
On Oct 8, 8:00 pm, OvermindDL1 wrote:

...
On Fri, Oct 8, 2010 at 6:57 PM, Joel de Guzman

...
wrote:

...
On 10/9/2010 2:44 AM, alfC wrote:

...
...
...
On Oct 6, 5:33 pm, Joel de Guzman wrote:

...
...
...
> On 10/7/2010 3:53 AM, alfC wrote:

...
...
...
>> Hi,

...
...
...
>> I am trying to convert a boost::array into a fusion::vector. >> reinterpreting the memory from boost::array. It works for tuples but >> not fusion::vector, why is that. Is there a way to make it work? ( I >> was hoping that the memory layout is the same to make conversion from >> one to the other very easy.)

...
...
...
> Don't do that. It will *never* be guaranteed to work even if it works > now. The memory layout of fusion::vector is an internal implementation > detail and can change anytime.

...
...
...
The question is: is it guaranteed for boost::tuples? (it seems so, just by the PODness of tuples of PODs in its stated design.) If so, then it is a *feature* of boost::tuple.

...
...
No it is not. 1) It is not documented 2) It is not a POD 3) Anything with a reinterpret_cast is not guaranteed to work (you can only rely on reinterpret_cast only when casting back from a lost type e.g. through type erasure).

...
...
Disregarding any of that is playing with fire.

...
The previous posts a few posts ago seems overly complicated, I would just do this (suitably wrapped into a function to handle it, maybe named copy):

...
// Proper includes here...

...
struct set1stTo2nd { template<typename T> void operator()(T& t) const { using namespace boost::fusion; deref(begin(t)) = deref(next(begin(t))); }};

...
typedef boost::array arrType; typedef boost::fusion::vector vecType; typedef boost::fusion::vector zipSeqType;

...
arrType arr(1,1,1); vecType vec(5,5,5);

...
for_each(zipSeqType(arr,vec), set1stTo2nd()); assert(arr == make_vector(5,5,5));

FINALLY! I got it. I feel better now.

It turns out that the correct line of code is

for_each( boost::fusion::zip_view<zipSeqType>(zipSeqType(arr2,vec1)), set1stTo2nd() );

you wrote the correct typedef but then didn't apply the zip_view function.

(Although, related to the original question, I am still a bit worried about too much unnecessary copy which seems unavoidable since the memory layout of array and fusion::vector is different.)

below is the full program that *works*:

#include #include #include #include #include #include #include #include #include #include #include

#include<iostream> using std::clog; using std::endl;

using namespace boost::fusion; struct set1stTo2nd { template<typename T> void operator()(T t) const { deref(boost::fusion::begin(t)) = deref(boost::fusion::next(boost::fusion::begin(t))); } };

typedef boost::array arrType; typedef boost::fusion::vector vecType; typedef boost::fusion::vector zipSeqType;

int main(){ boost::array arr1={{1.0, 2.1, 3.2}}; clog << "arr1: " << arr1[0] << ", " << arr1[1] << ", " << arr1[2] << endl;

boost::fusion::vector vec1(arr1); clog << "vec1: " << at_c<0>(vec1) << ", "<< at_c<1>(vec1) << ", " << at_c<2>(vec1) << endl;

boost::array arr2; for_each( boost::fusion::zip_view<zipSeqType>(zipSeqType(arr2,vec1)), set1stTo2nd() ); clog << "arr2: " << at_c<0>(arr2) << ", "<< at_c<1>(arr2) << ", " << at_c<2>(arr2) << endl; return 0; }

prints: arr1: 1, 2.1, 3.2 vec1: 1, 2.1, 3.2 arr2: 1, 2.1, 3.2

Have a good weekend, Alfredo

Yeah, my mistake, busy at work and not paying attention, I meant to have the fusion::zip there. :)

another subtlety is that the correct signature is struct set1stTo2nd { template<typename T> void operator()(T t) const // not T& t) const ...

...

_______________________________________________ Boost-users mailing list Boost-users@lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

-- Sie haben diese Nachricht erhalten, da Sie der Google Groups-Gruppe Boost Users beigetreten sind. Wenn Sie Nachrichten in dieser Gruppe posten möchten, senden Sie eine E-Mail an boostusers@googlegroups.com. Wenn Sie aus dieser Gruppe austreten möchten, senden Sie eine E-Mail an boostusers+unsubscribe@googlegroups.com. Besuchen Sie die Gruppe unter http://groups.google.com/group/boostusers?hl=de, um weitere Optionen zu erhalten.

5145

Age (days ago)

5153

Last active (days ago)

List overview

Download

20 comments

9 participants

participants (9)

alfC
Alfredo Correa
Alfredo Correa
Christopher Schmidt
Joel de Guzman
Nat Goodspeed
OvermindDL1
Steven Watanabe
Thomas Heller