Hi All,

I've been working with Stephen Kyne on getting the Visual Studio build
to run on Appveyor (a continuous integration system like Travis).
We're getting stuck because downloading and building Boost takes a
long time and thus the build times out.  As far as I can see, we seem
to have a dependency on Boost in just a few places:

1. interfaces/CsoundFile.cpp
2. CsoundAC
3. CsoundVST

I think for now our best bet to get a continuous Windows build with
Appveyor going is to remove the dependency on Boost from
interfaces/CsoundFile.cpp and just not build CsoundAC/CsoundVST.

For CsoundFile.cpp, there's just one use of boost:

 // TODO: Replace boost throughout interfaces.
  #if !defined(__ANDROID__)
  boost::split(args, buffer, boost::is_any_of(" \t\n\r"),
boost::token_compress_on);
  #endif

Could someone (Michael perhaps? I'm guessing this might be code of
yours?) take a look and perhaps find a way to get rid of using Boost
here?

Thanks!

I will take care of this. I would also like to remove the boost
dependency from CsoundAC etc, not sure how hard that is. The
dependencies regarding matrix arithmetic would be replaced by Eigen
(eigen3), can Appveyor deal with that?

Best,
Mike

-----------------------------------------------------
Michael Gogins
Irreducible Productions
http://michaelgogins.tumblr.com
Michael dot Gogins at gmail dot com


On Tue, Apr 25, 2017 at 5:09 PM, Steven Yi  wrote:
> Hi All,
>
> I've been working with Stephen Kyne on getting the Visual Studio build
> to run on Appveyor (a continuous integration system like Travis).
> We're getting stuck because downloading and building Boost takes a
> long time and thus the build times out.  As far as I can see, we seem
> to have a dependency on Boost in just a few places:
>
> 1. interfaces/CsoundFile.cpp
> 2. CsoundAC
> 3. CsoundVST
>
> I think for now our best bet to get a continuous Windows build with
> Appveyor going is to remove the dependency on Boost from
> interfaces/CsoundFile.cpp and just not build CsoundAC/CsoundVST.
>
> For CsoundFile.cpp, there's just one use of boost:
>
>  // TODO: Replace boost throughout interfaces.
>   #if !defined(__ANDROID__)
>   boost::split(args, buffer, boost::is_any_of(" \t\n\r"),
> boost::token_compress_on);
>   #endif
>
> Could someone (Michael perhaps? I'm guessing this might be code of
> yours?) take a look and perhaps find a way to get rid of using Boost
> here?
>
> Thanks!

Thanks Michael!

I do think eigen3 is fine as a dependency.  It works at least with
vcpkg as far as I can tell.  We'll have to test again once boost is
off the list of dependencies as it sort of just killed the whole build
and I'm not sure we have any data yet on how long it takes to build
without boost.  (Stephen may have a better idea of it all though.)

On Tue, Apr 25, 2017 at 6:35 PM, Michael Gogins
 wrote:
> I will take care of this. I would also like to remove the boost
> dependency from CsoundAC etc, not sure how hard that is. The
> dependencies regarding matrix arithmetic would be replaced by Eigen
> (eigen3), can Appveyor deal with that?
>
> Best,
> Mike
>
> -----------------------------------------------------
> Michael Gogins
> Irreducible Productions
> http://michaelgogins.tumblr.com
> Michael dot Gogins at gmail dot com
>
>
> On Tue, Apr 25, 2017 at 5:09 PM, Steven Yi  wrote:
>> Hi All,
>>
>> I've been working with Stephen Kyne on getting the Visual Studio build
>> to run on Appveyor (a continuous integration system like Travis).
>> We're getting stuck because downloading and building Boost takes a
>> long time and thus the build times out.  As far as I can see, we seem
>> to have a dependency on Boost in just a few places:
>>
>> 1. interfaces/CsoundFile.cpp
>> 2. CsoundAC
>> 3. CsoundVST
>>
>> I think for now our best bet to get a continuous Windows build with
>> Appveyor going is to remove the dependency on Boost from
>> interfaces/CsoundFile.cpp and just not build CsoundAC/CsoundVST.
>>
>> For CsoundFile.cpp, there's just one use of boost:
>>
>>  // TODO: Replace boost throughout interfaces.
>>   #if !defined(__ANDROID__)
>>   boost::split(args, buffer, boost::is_any_of(" \t\n\r"),
>> boost::token_compress_on);
>>   #endif
>>
>> Could someone (Michael perhaps? I'm guessing this might be code of
>> yours?) take a look and perhaps find a way to get rid of using Boost
>> here?
>>
>> Thanks!

I've created issues in GitHub for removing the Boost dependency from
the Csound interfaces, from CsoundVST, and from CsoundAC.

I have pushed the fixes for the Csound interfaces and for CsoundVST,
but not tested them, which I will do later today.

It turns out I had already replaced boost::numeric::ublas with eigen3
in CsoundAC. It remains to replace Boost formatting and random
distributions. I am working on that now and expect it to be
straightforward, but the C++ requirement will move to -std=c++11. It
looks like the Boost random distributions more or less went into
C++11.

Best,
Mike

-----------------------------------------------------
Michael Gogins
Irreducible Productions
http://michaelgogins.tumblr.com
Michael dot Gogins at gmail dot com


On Tue, Apr 25, 2017 at 7:52 PM, Steven Yi  wrote:
> Thanks Michael!
>
> I do think eigen3 is fine as a dependency.  It works at least with
> vcpkg as far as I can tell.  We'll have to test again once boost is
> off the list of dependencies as it sort of just killed the whole build
> and I'm not sure we have any data yet on how long it takes to build
> without boost.  (Stephen may have a better idea of it all though.)
>
> On Tue, Apr 25, 2017 at 6:35 PM, Michael Gogins
>  wrote:
>> I will take care of this. I would also like to remove the boost
>> dependency from CsoundAC etc, not sure how hard that is. The
>> dependencies regarding matrix arithmetic would be replaced by Eigen
>> (eigen3), can Appveyor deal with that?
>>
>> Best,
>> Mike
>>
>> -----------------------------------------------------
>> Michael Gogins
>> Irreducible Productions
>> http://michaelgogins.tumblr.com
>> Michael dot Gogins at gmail dot com
>>
>>
>> On Tue, Apr 25, 2017 at 5:09 PM, Steven Yi  wrote:
>>> Hi All,
>>>
>>> I've been working with Stephen Kyne on getting the Visual Studio build
>>> to run on Appveyor (a continuous integration system like Travis).
>>> We're getting stuck because downloading and building Boost takes a
>>> long time and thus the build times out.  As far as I can see, we seem
>>> to have a dependency on Boost in just a few places:
>>>
>>> 1. interfaces/CsoundFile.cpp
>>> 2. CsoundAC
>>> 3. CsoundVST
>>>
>>> I think for now our best bet to get a continuous Windows build with
>>> Appveyor going is to remove the dependency on Boost from
>>> interfaces/CsoundFile.cpp and just not build CsoundAC/CsoundVST.
>>>
>>> For CsoundFile.cpp, there's just one use of boost:
>>>
>>>  // TODO: Replace boost throughout interfaces.
>>>   #if !defined(__ANDROID__)
>>>   boost::split(args, buffer, boost::is_any_of(" \t\n\r"),
>>> boost::token_compress_on);
>>>   #endif
>>>
>>> Could someone (Michael perhaps? I'm guessing this might be code of
>>> yours?) take a look and perhaps find a way to get rid of using Boost
>>> here?
>>>
>>> Thanks!

I think it makes sense to move to c++11 for the whole codebase, after all it’s 2017. The only issue might be VST, 
which appears to have problems with it (I noticed when trying to compile the VST host opcodes).
========================
Prof. Victor Lazzarini
Dean of Arts, Celtic Studies, and Philosophy,
Maynooth University,
Maynooth, Co Kildare, Ireland
Tel: 00 353 7086936
Fax: 00 353 1 7086952 

> On 26 Apr 2017, at 17:08, Michael Gogins  wrote:
> 
> I've created issues in GitHub for removing the Boost dependency from
> the Csound interfaces, from CsoundVST, and from CsoundAC.
> 
> I have pushed the fixes for the Csound interfaces and for CsoundVST,
> but not tested them, which I will do later today.
> 
> It turns out I had already replaced boost::numeric::ublas with eigen3
> in CsoundAC. It remains to replace Boost formatting and random
> distributions. I am working on that now and expect it to be
> straightforward, but the C++ requirement will move to -std=c++11. It
> looks like the Boost random distributions more or less went into
> C++11.
> 
> Best,
> Mike
> 
> -----------------------------------------------------
> Michael Gogins
> Irreducible Productions
> http://michaelgogins.tumblr.com
> Michael dot Gogins at gmail dot com
> 
> 
> On Tue, Apr 25, 2017 at 7:52 PM, Steven Yi  wrote:
>> Thanks Michael!
>> 
>> I do think eigen3 is fine as a dependency.  It works at least with
>> vcpkg as far as I can tell.  We'll have to test again once boost is
>> off the list of dependencies as it sort of just killed the whole build
>> and I'm not sure we have any data yet on how long it takes to build
>> without boost.  (Stephen may have a better idea of it all though.)
>> 
>> On Tue, Apr 25, 2017 at 6:35 PM, Michael Gogins
>>  wrote:
>>> I will take care of this. I would also like to remove the boost
>>> dependency from CsoundAC etc, not sure how hard that is. The
>>> dependencies regarding matrix arithmetic would be replaced by Eigen
>>> (eigen3), can Appveyor deal with that?
>>> 
>>> Best,
>>> Mike
>>> 
>>> -----------------------------------------------------
>>> Michael Gogins
>>> Irreducible Productions
>>> http://michaelgogins.tumblr.com
>>> Michael dot Gogins at gmail dot com
>>> 
>>> 
>>> On Tue, Apr 25, 2017 at 5:09 PM, Steven Yi  wrote:
>>>> Hi All,
>>>> 
>>>> I've been working with Stephen Kyne on getting the Visual Studio build
>>>> to run on Appveyor (a continuous integration system like Travis).
>>>> We're getting stuck because downloading and building Boost takes a
>>>> long time and thus the build times out.  As far as I can see, we seem
>>>> to have a dependency on Boost in just a few places:
>>>> 
>>>> 1. interfaces/CsoundFile.cpp
>>>> 2. CsoundAC
>>>> 3. CsoundVST
>>>> 
>>>> I think for now our best bet to get a continuous Windows build with
>>>> Appveyor going is to remove the dependency on Boost from
>>>> interfaces/CsoundFile.cpp and just not build CsoundAC/CsoundVST.
>>>> 
>>>> For CsoundFile.cpp, there's just one use of boost:
>>>> 
>>>> // TODO: Replace boost throughout interfaces.
>>>>  #if !defined(__ANDROID__)
>>>>  boost::split(args, buffer, boost::is_any_of(" \t\n\r"),
>>>> boost::token_compress_on);
>>>>  #endif
>>>> 
>>>> Could someone (Michael perhaps? I'm guessing this might be code of
>>>> yours?) take a look and perhaps find a way to get rid of using Boost
>>>> he

Thanks Michael for the update!  I think C++11 makes sense. We'll just
have to be sure to test on all of the compilers we use. I think we'll
be fine for Visual Studio (I found this link for VS2015, but not one
yet for VS2017, but C++11 looks well covered even in VS2015:
https://msdn.microsoft.com/en-us/library/hh567368.aspx).  I think the
only other compilers that are used are GCC and Clang and they should
both be fine on all of the platforms.  (Assuming Emscripten/Clang is
fine.)  We'll just have to be sure to explicitly set C++11 for all of
our platforms.

On Wed, Apr 26, 2017 at 12:23 PM, Rory Walsh  wrote:
> I agree.
>
> On 26 April 2017 at 17:12, Victor Lazzarini 
> wrote:
>>
>> I think it makes sense to move to c++11 for the whole codebase, after all
>> it’s 2017. The only issue might be VST,
>> which appears to have problems with it (I noticed when trying to compile
>> the VST host opcodes).
>> ========================
>> Prof. Victor Lazzarini
>> Dean of Arts, Celtic Studies, and Philosophy,
>> Maynooth University,
>> Maynooth, Co Kildare, Ireland
>> Tel: 00 353 7086936
>> Fax: 00 353 1 7086952
>>
>> > On 26 Apr 2017, at 17:08, Michael Gogins 
>> > wrote:
>> >
>> > I've created issues in GitHub for removing the Boost dependency from
>> > the Csound interfaces, from CsoundVST, and from CsoundAC.
>> >
>> > I have pushed the fixes for the Csound interfaces and for CsoundVST,
>> > but not tested them, which I will do later today.
>> >
>> > It turns out I had already replaced boost::numeric::ublas with eigen3
>> > in CsoundAC. It remains to replace Boost formatting and random
>> > distributions. I am working on that now and expect it to be
>> > straightforward, but the C++ requirement will move to -std=c++11. It
>> > looks like the Boost random distributions more or less went into
>> > C++11.
>> >
>> > Best,
>> > Mike
>> >
>> > -----------------------------------------------------
>> > Michael Gogins
>> > Irreducible Productions
>> > http://michaelgogins.tumblr.com
>> > Michael dot Gogins at gmail dot com
>> >
>> >
>> > On Tue, Apr 25, 2017 at 7:52 PM, Steven Yi  wrote:
>> >> Thanks Michael!
>> >>
>> >> I do think eigen3 is fine as a dependency.  It works at least with
>> >> vcpkg as far as I can tell.  We'll have to test again once boost is
>> >> off the list of dependencies as it sort of just killed the whole build
>> >> and I'm not sure we have any data yet on how long it takes to build
>> >> without boost.  (Stephen may have a better idea of it all though.)
>> >>
>> >> On Tue, Apr 25, 2017 at 6:35 PM, Michael Gogins
>> >>  wrote:
>> >>> I will take care of this. I would also like to remove the boost
>> >>> dependency from CsoundAC etc, not sure how hard that is. The
>> >>> dependencies regarding matrix arithmetic would be replaced by Eigen
>> >>> (eigen3), can Appveyor deal with that?
>> >>>
>> >>> Best,
>> >>> Mike
>> >>>
>> >>> -----------------------------------------------------
>> >>> Michael Gogins
>> >>> Irreducible Productions
>> >>> http://michaelgogins.tumblr.com
>> >>> Michael dot Gogins at gmail dot com
>> >>>
>> >>>
>> >>> On Tue, Apr 25, 2017 at 5:09 PM, Steven Yi  wrote:
>> >>>> Hi All,
>> >>>>
>> >>>> I've been working with Stephen Kyne on getting the Visual Studio
>> >>>> build
>> >>>> to run on Appveyor (a continuous integration system like Travis).
>> >>>> We're getting stuck because downloading and building Boost takes a
>> >>>> long time and thus the build times out.  As far as I can see, we seem
>> >>>> to have a dependency on Boost in just a few places:
>> >>>>
>> >>>> 1. interfaces/CsoundFile.cpp
>> >>>> 2. CsoundAC
>> >>>> 3. CsoundVST
>> >>>>
>> >>>> I think for now our best bet to get a continuous Windows build with
>> >>>> Appveyor going is to remove the dependency on Boost from
>> >>>> interfaces/CsoundFile.cpp and just not build CsoundAC/CsoundVST.
>> >>>>
>> >>>> For CsoundFile.cpp, there's just one use of boost:
>> >>>>
>> >>>> // TODO: Replace boost throughout interfaces.
>> >>>>  #if !defined(__ANDROID__)
>> >>>>  boost::split(args, buffer, boost::is_any_of(" \t\n\r"),
>> >>>> boost::token_compress_on);
>> >>>>  #endif
>> >>>>
>> >>>> Could someone (Michael perhaps? I'm guessing this might be code of
>> >>>> yours?) take a look and perhaps find a way to get rid of using Boost
>> >>>> here?
>> >>>>
>> >>>> Thanks!
>> >>>> steven
>>

Some older GCCs (e.g. the one that comes in Debian Jessie) do not have full c++11 support, but I think most
linuxes now come with the newer gccs (> 6.0 I guess). In raspbian jessie, I had to do get a newer compiler 
package (can’t remember which exactly). It can be a pain sometimes, but I think we should roll out the code for
c++11.

========================
Prof. Victor Lazzarini
Dean of Arts, Celtic Studies, and Philosophy,
Maynooth University,
Maynooth, Co Kildare, Ireland
Tel: 00 353 7086936
Fax: 00 353 1 7086952 

> On 26 Apr 2017, at 18:11, Steven Yi  wrote:
> 
> Thanks Michael for the update!  I think C++11 makes sense. We'll just
> have to be sure to test on all of the compilers we use. I think we'll
> be fine for Visual Studio (I found this link for VS2015, but not one
> yet for VS2017, but C++11 looks well covered even in VS2015:
> https://msdn.microsoft.com/en-us/library/hh567368.aspx).  I think the
> only other compilers that are used are GCC and Clang and they should
> both be fine on all of the platforms.  (Assuming Emscripten/Clang is
> fine.)  We'll just have to be sure to explicitly set C++11 for all of
> our platforms.
> 
> On Wed, Apr 26, 2017 at 12:23 PM, Rory Walsh  wrote:
>> I agree.
>> 
>> On 26 April 2017 at 17:12, Victor Lazzarini 
>> wrote:
>>> 
>>> I think it makes sense to move to c++11 for the whole codebase, after all
>>> it’s 2017. The only issue might be VST,
>>> which appears to have problems with it (I noticed when trying to compile
>>> the VST host opcodes).
>>> ========================
>>> Prof. Victor Lazzarini
>>> Dean of Arts, Celtic Studies, and Philosophy,
>>> Maynooth University,
>>> Maynooth, Co Kildare, Ireland
>>> Tel: 00 353 7086936
>>> Fax: 00 353 1 7086952
>>> 
>>>> On 26 Apr 2017, at 17:08, Michael Gogins 
>>>> wrote:
>>>> 
>>>> I've created issues in GitHub for removing the Boost dependency from
>>>> the Csound interfaces, from CsoundVST, and from CsoundAC.
>>>> 
>>>> I have pushed the fixes for the Csound interfaces and for CsoundVST,
>>>> but not tested them, which I will do later today.
>>>> 
>>>> It turns out I had already replaced boost::numeric::ublas with eigen3
>>>> in CsoundAC. It remains to replace Boost formatting and random
>>>> distributions. I am working on that now and expect it to be
>>>> straightforward, but the C++ requirement will move to -std=c++11. It
>>>> looks like the Boost random distributions more or less went into
>>>> C++11.
>>>> 
>>>> Best,
>>>> Mike
>>>> 
>>>> -----------------------------------------------------
>>>> Michael Gogins
>>>> Irreducible Productions
>>>> http://michaelgogins.tumblr.com
>>>> Michael dot Gogins at gmail dot com
>>>> 
>>>> 
>>>> On Tue, Apr 25, 2017 at 7:52 PM, Steven Yi  wrote:
>>>>> Thanks Michael!
>>>>> 
>>>>> I do think eigen3 is fine as a dependency.  It works at least with
>>>>> vcpkg as far as I can tell.  We'll have to test again once boost is
>>>>> off the list of dependencies as it sort of just killed the whole build
>>>>> and I'm not sure we have any data yet on how long it takes to build
>>>>> without boost.  (Stephen may have a better idea of it all though.)
>>>>> 
>>>>> On Tue, Apr 25, 2017 at 6:35 PM, Michael Gogins
>>>>>  wrote:
>>>>>> I will take care of this. I would also like to remove the boost
>>>>>> dependency from CsoundAC etc, not sure how hard that is. The
>>>>>> dependencies regarding matrix arithmetic would be replaced by Eigen
>>>>>> (eigen3), can Appveyor deal with that?
>>>>>> 
>>>>>> Best,
>>>>>> Mike
>>>>>> 
>>>>>> -----------------------------------------------------
>>>>>> Michael Gogins
>>>>>> Irreducible Productions
>>>>>> http://michaelgogins.tumblr.com
>>>>>> Michael dot Gogins at gmail dot com
>>>>>> 
>>>>>> 
>>>>>> On Tue, Apr 25, 2017 at 5:09 PM, Steven Yi  wrote:
>>>>>>> Hi All,
>>>>>>> 
>>>>>>> I've been working with Stephen Kyne on getting the Visual Studio
>>>>>>> build
>>>>>>> to run on Appveyor (a continuous integration system like Travis).
>>>>>>> We're getting stuck because downloading and building Boost takes a
>>>>>>> long time and thus the build times out.  As far as I can see, we seem
>>>>>>> to have a dependency on Boost in just a few places:
>>>>>>> 
>>>>>>> 1. interfaces/CsoundFile.cpp
>>>>>>> 2. CsoundAC
>>>>>>> 3. CsoundVST
>>>>>>> 
>>>>>>> I think for now our best bet to get a continuous Windows build with
>>>>>>> Appveyor going is to remove the dependency on Boost from
>>>>>>> interfaces/CsoundFile.cpp and just not build CsoundAC/CsoundVST.
>>>>>>> 
>>>>>>> For CsoundFile.cpp, there's just one use of boost:
>>>>>>> 
>>>>>>> // TODO: Replace boost throughout interfaces.
>>>>>>> #if !defined(__ANDROID__)
>>>>>>> boost::split(args, buffer, boost::is_any_of(" \t\n\r"),
>>>>>>> boost::token_compress_on);
>>>>>>> #endif
>>>>>>> 
>>>>>>> Could someone (Michael perhaps? I'm guessing this might be code of
>>>>>>> yours?) take a look and perhaps find a way to get rid of using Boost
>>>>>>> here?
>>>>>>> 
>>>>>>> T

I agree that we should move to uniformly requiring at least C++11. If
the VST host opcodes are a problem, I will fix them.

Best,
Mike

-----------------------------------------------------
Michael Gogins
Irreducible Productions
http://michaelgogins.tumblr.com
Michael dot Gogins at gmail dot com


On Wed, Apr 26, 2017 at 1:25 PM, Victor Lazzarini
 wrote:
> Some older GCCs (e.g. the one that comes in Debian Jessie) do not have full c++11 support, but I think most
> linuxes now come with the newer gccs (> 6.0 I guess). In raspbian jessie, I had to do get a newer compiler
> package (can’t remember which exactly). It can be a pain sometimes, but I think we should roll out the code for
> c++11.
>
> ========================
> Prof. Victor Lazzarini
> Dean of Arts, Celtic Studies, and Philosophy,
> Maynooth University,
> Maynooth, Co Kildare, Ireland
> Tel: 00 353 7086936
> Fax: 00 353 1 7086952
>
>> On 26 Apr 2017, at 18:11, Steven Yi  wrote:
>>
>> Thanks Michael for the update!  I think C++11 makes sense. We'll just
>> have to be sure to test on all of the compilers we use. I think we'll
>> be fine for Visual Studio (I found this link for VS2015, but not one
>> yet for VS2017, but C++11 looks well covered even in VS2015:
>> https://msdn.microsoft.com/en-us/library/hh567368.aspx).  I think the
>> only other compilers that are used are GCC and Clang and they should
>> both be fine on all of the platforms.  (Assuming Emscripten/Clang is
>> fine.)  We'll just have to be sure to explicitly set C++11 for all of
>> our platforms.
>>
>> On Wed, Apr 26, 2017 at 12:23 PM, Rory Walsh  wrote:
>>> I agree.
>>>
>>> On 26 April 2017 at 17:12, Victor Lazzarini 
>>> wrote:
>>>>
>>>> I think it makes sense to move to c++11 for the whole codebase, after all
>>>> it’s 2017. The only issue might be VST,
>>>> which appears to have problems with it (I noticed when trying to compile
>>>> the VST host opcodes).
>>>> ========================
>>>> Prof. Victor Lazzarini
>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>> Maynooth University,
>>>> Maynooth, Co Kildare, Ireland
>>>> Tel: 00 353 7086936
>>>> Fax: 00 353 1 7086952
>>>>
>>>>> On 26 Apr 2017, at 17:08, Michael Gogins 
>>>>> wrote:
>>>>>
>>>>> I've created issues in GitHub for removing the Boost dependency from
>>>>> the Csound interfaces, from CsoundVST, and from CsoundAC.
>>>>>
>>>>> I have pushed the fixes for the Csound interfaces and for CsoundVST,
>>>>> but not tested them, which I will do later today.
>>>>>
>>>>> It turns out I had already replaced boost::numeric::ublas with eigen3
>>>>> in CsoundAC. It remains to replace Boost formatting and random
>>>>> distributions. I am working on that now and expect it to be
>>>>> straightforward, but the C++ requirement will move to -std=c++11. It
>>>>> looks like the Boost random distributions more or less went into
>>>>> C++11.
>>>>>
>>>>> Best,
>>>>> Mike
>>>>>
>>>>> -----------------------------------------------------
>>>>> Michael Gogins
>>>>> Irreducible Productions
>>>>> http://michaelgogins.tumblr.com
>>>>> Michael dot Gogins at gmail dot com
>>>>>
>>>>>
>>>>> On Tue, Apr 25, 2017 at 7:52 PM, Steven Yi  wrote:
>>>>>> Thanks Michael!
>>>>>>
>>>>>> I do think eigen3 is fine as a dependency.  It works at least with
>>>>>> vcpkg as far as I can tell.  We'll have to test again once boost is
>>>>>> off the list of dependencies as it sort of just killed the whole build
>>>>>> and I'm not sure we have any data yet on how long it takes to build
>>>>>> without boost.  (Stephen may have a better idea of it all though.)
>>>>>>
>>>>>> On Tue, Apr 25, 2017 at 6:35 PM, Michael Gogins
>>>>>>  wrote:
>>>>>>> I will take care of this. I would also like to remove the boost
>>>>>>> dependency from CsoundAC etc, not sure how hard that is. The
>>>>>>> dependencies regarding matrix arithmetic would be replaced by Eigen
>>>>>>> (eigen3), can Appveyor deal with that?
>>>>>>>
>>>>>>> Best,
>>>>>>> Mike
>>>>>>>
>>>>>>> -----------------------------------------------------
>>>>>>> Michael Gogins
>>>>>>> Irreducible Productions
>>>>>>> http://michaelgogins.tumblr.com
>>>>>>> Michael dot Gogins at gmail dot com
>>>>>>>
>>>>>>>
>>>>>>> On Tue, Apr 25, 2017 at 5:09 PM, Steven Yi  wrote:
>>>>>>>> Hi All,
>>>>>>>>
>>>>>>>> I've been working with Stephen Kyne on getting the Visual Studio
>>>>>>>> build
>>>>>>>> to run on Appveyor (a continuous integration system like Travis).
>>>>>>>> We're getting stuck because downloading and building Boost takes a
>>>>>>>> long time and thus the build times out.  As far as I can see, we seem
>>>>>>>> to have a dependency on Boost in just a few places:
>>>>>>>>
>>>>>>>> 1. interfaces/CsoundFile.cpp
>>>>>>>> 2. CsoundAC
>>>>>>>> 3. CsoundVST
>>>>>>>>
>>>>>>>> I think for now our best bet to get a continuous Windows build with
>>>>>>>> Appveyor going is to remove the dependency on Boost from
>>>>>>>> interfaces/CsoundFile.cpp and just not build CsoundAC/CsoundVST.
>>>>>>>>
>>>>>>>> For CsoundFile.cpp, there's just one use of boost:
>>>>>>>>
>>>>>>>> // TODO: Replace boost throughout interfaces.
>>>>>>>> #if !defined(__ANDROID__)
>>>>>>>> boost::split(args, buffer, boost::is_any_of(" \t\n\r"),
>>>>>>>> boost::token_compress_on);
>>>>>>>> #endif
>>>>>>>>
>>>>>>>> Could someone (Michael perhaps? I'm guessing this might be code of
>>>>>>>> yours?) take a look and perhaps find a way to get rid of using Boost
>>>>>>>> here?
>>>>>>>>
>>>>>>>> Thanks!
>>>>>>>> steven
>>>>
>>>

The problem is not with our code, but with the SDK. 
========================
Prof. Victor Lazzarini
Dean of Arts, Celtic Studies, and Philosophy,
Maynooth University,
Maynooth, Co Kildare, Ireland
Tel: 00 353 7086936
Fax: 00 353 1 7086952 

> On 26 Apr 2017, at 18:34, Michael Gogins  wrote:
> 
> I agree that we should move to uniformly requiring at least C++11. If
> the VST host opcodes are a problem, I will fix them.
> 
> Best,
> Mike
> 
> -----------------------------------------------------
> Michael Gogins
> Irreducible Productions
> http://michaelgogins.tumblr.com
> Michael dot Gogins at gmail dot com
> 
> 
> On Wed, Apr 26, 2017 at 1:25 PM, Victor Lazzarini
>  wrote:
>> Some older GCCs (e.g. the one that comes in Debian Jessie) do not have full c++11 support, but I think most
>> linuxes now come with the newer gccs (> 6.0 I guess). In raspbian jessie, I had to do get a newer compiler
>> package (can’t remember which exactly). It can be a pain sometimes, but I think we should roll out the code for
>> c++11.
>> 
>> ========================
>> Prof. Victor Lazzarini
>> Dean of Arts, Celtic Studies, and Philosophy,
>> Maynooth University,
>> Maynooth, Co Kildare, Ireland
>> Tel: 00 353 7086936
>> Fax: 00 353 1 7086952
>> 
>>> On 26 Apr 2017, at 18:11, Steven Yi  wrote:
>>> 
>>> Thanks Michael for the update!  I think C++11 makes sense. We'll just
>>> have to be sure to test on all of the compilers we use. I think we'll
>>> be fine for Visual Studio (I found this link for VS2015, but not one
>>> yet for VS2017, but C++11 looks well covered even in VS2015:
>>> https://msdn.microsoft.com/en-us/library/hh567368.aspx).  I think the
>>> only other compilers that are used are GCC and Clang and they should
>>> both be fine on all of the platforms.  (Assuming Emscripten/Clang is
>>> fine.)  We'll just have to be sure to explicitly set C++11 for all of
>>> our platforms.
>>> 
>>> On Wed, Apr 26, 2017 at 12:23 PM, Rory Walsh  wrote:
>>>> I agree.
>>>> 
>>>> On 26 April 2017 at 17:12, Victor Lazzarini 
>>>> wrote:
>>>>> 
>>>>> I think it makes sense to move to c++11 for the whole codebase, after all
>>>>> it’s 2017. The only issue might be VST,
>>>>> which appears to have problems with it (I noticed when trying to compile
>>>>> the VST host opcodes).
>>>>> ========================
>>>>> Prof. Victor Lazzarini
>>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>>> Maynooth University,
>>>>> Maynooth, Co Kildare, Ireland
>>>>> Tel: 00 353 7086936
>>>>> Fax: 00 353 1 7086952
>>>>> 
>>>>>> On 26 Apr 2017, at 17:08, Michael Gogins 
>>>>>> wrote:
>>>>>> 
>>>>>> I've created issues in GitHub for removing the Boost dependency from
>>>>>> the Csound interfaces, from CsoundVST, and from CsoundAC.
>>>>>> 
>>>>>> I have pushed the fixes for the Csound interfaces and for CsoundVST,
>>>>>> but not tested them, which I will do later today.
>>>>>> 
>>>>>> It turns out I had already replaced boost::numeric::ublas with eigen3
>>>>>> in CsoundAC. It remains to replace Boost formatting and random
>>>>>> distributions. I am working on that now and expect it to be
>>>>>> straightforward, but the C++ requirement will move to -std=c++11. It
>>>>>> looks like the Boost random distributions more or less went into
>>>>>> C++11.
>>>>>> 
>>>>>> Best,
>>>>>> Mike
>>>>>> 
>>>>>> -----------------------------------------------------
>>>>>> Michael Gogins
>>>>>> Irreducible Productions
>>>>>> http://michaelgogins.tumblr.com
>>>>>> Michael dot Gogins at gmail dot com
>>>>>> 
>>>>>> 
>>>>>> On Tue, Apr 25, 2017 at 7:52 PM, Steven Yi  wrote:
>>>>>>> Thanks Michael!
>>>>>>> 
>>>>>>> I do think eigen3 is fine as a dependency.  It works at least with
>>>>>>> vcpkg as far as I can tell.  We'll have to test again once boost is
>>>>>>> off the list of dependencies as it sort of just killed the whole build
>>>>>>> and I'm not sure we have any data yet on how long it takes to build
>>>>>>> without boost.  (Stephen may have a better idea of it all though.)
>>>>>>> 
>>>>>>> On Tue, Apr 25, 2017 at 6:35 PM, Michael Gogins
>>>>>>>  wrote:
>>>>>>>> I will take care of this. I would also like to remove the boost
>>>>>>>> dependency from CsoundAC etc, not sure how hard that is. The
>>>>>>>> dependencies regarding matrix arithmetic would be replaced by Eigen
>>>>>>>> (eigen3), can Appveyor deal with that?
>>>>>>>> 
>>>>>>>> Best,
>>>>>>>> Mike
>>>>>>>> 
>>>>>>>> -----------------------------------------------------
>>>>>>>> Michael Gogins
>>>>>>>> Irreducible Productions
>>>>>>>> http://michaelgogins.tumblr.com
>>>>>>>> Michael dot Gogins at gmail dot com
>>>>>>>> 
>>>>>>>> 
>>>>>>>> On Tue, Apr 25, 2017 at 5:09 PM, Steven Yi  wrote:
>>>>>>>>> Hi All,
>>>>>>>>> 
>>>>>>>>> I've been working with Stephen Kyne on getting the Visual Studio
>>>>>>>>> build
>>>>>>>>> to run on Appveyor (a continuous integration system like Travis).
>>>>>>>>> We're getting stuck because downloading and building Boost takes a
>>>>>>>>> long time and thus the build times out.  As far as I can see, we seem
>>>>>>>>> to have a dependency on Boost in just a few places:
>>>>>>>>> 
>>>>>>>>> 1. interfaces/CsoundFile.cpp
>>>>>>>>> 2. CsoundAC
>>>>>>>>> 3. CsoundVST
>>>>>>>>> 
>>>>>>>>> I think for now our best bet to get a continuous Windows build with
>>>>>>>>> Appveyor going is to remove the dependency on Boost from
>>>>>>>>> interfaces/CsoundFile.cpp and just not build CsoundAC/CsoundVST.
>>>>>>>>> 
>>>>>>>>> For CsoundFile.cpp, there's just one use of boost:
>>>>>>>>> 
>>>>>>>>> // TODO: Replace boost throughout interfaces.
>>>>>>>>> #if !defined(__ANDROID__)
>>>>>>>>> boost::split(args, buffer, boost::is_any_of(" \t\n\r"),
>>>>>>>>> boost::token_compress_on);
>>>>>>>>> #endif
>>>>>>>>> 
>>>>>>>>> Could someone (Michael perhaps? I'm guessing this might be code of
>>>>>>>>> yours?) take a look and perhaps find a way to get rid of using Boost
>>>>>>>>> here?
>>>>>>>>> 
>>>>>>>>> Tha

Well, I might be wrong, but I remember adding -std=c++11 to cmake and making VST4cs fail to build with
a complaint about an SDK header.

========================
Prof. Victor Lazzarini
Dean of Arts, Celtic Studies, and Philosophy,
Maynooth University,
Maynooth, Co Kildare, Ireland
Tel: 00 353 7086936
Fax: 00 353 1 7086952 

> On 26 Apr 2017, at 18:52, Rory Walsh  wrote:
> 
> That's odd, I've been building Cabbage with c++11 for quite some time without any problems, and it use the VST 2.4 SDK?  
> 
> On 26 April 2017 at 18:40, Victor Lazzarini  wrote:
> The problem is not with our code, but with the SDK.
> ========================
> Prof. Victor Lazzarini
> Dean of Arts, Celtic Studies, and Philosophy,
> Maynooth University,
> Maynooth, Co Kildare, Ireland
> Tel: 00 353 7086936
> Fax: 00 353 1 7086952
> 
> > On 26 Apr 2017, at 18:34, Michael Gogins  wrote:
> >
> > I agree that we should move to uniformly requiring at least C++11. If
> > the VST host opcodes are a problem, I will fix them.
> >
> > Best,
> > Mike
> >
> > -----------------------------------------------------
> > Michael Gogins
> > Irreducible Productions
> > http://michaelgogins.tumblr.com
> > Michael dot Gogins at gmail dot com
> >
> >
> > On Wed, Apr 26, 2017 at 1:25 PM, Victor Lazzarini
> >  wrote:
> >> Some older GCCs (e.g. the one that comes in Debian Jessie) do not have full c++11 support, but I think most
> >> linuxes now come with the newer gccs (> 6.0 I guess). In raspbian jessie, I had to do get a newer compiler
> >> package (can’t remember which exactly). It can be a pain sometimes, but I think we should roll out the code for
> >> c++11.
> >>
> >> ========================
> >> Prof. Victor Lazzarini
> >> Dean of Arts, Celtic Studies, and Philosophy,
> >> Maynooth University,
> >> Maynooth, Co Kildare, Ireland
> >> Tel: 00 353 7086936
> >> Fax: 00 353 1 7086952
> >>
> >>> On 26 Apr 2017, at 18:11, Steven Yi  wrote:
> >>>
> >>> Thanks Michael for the update!  I think C++11 makes sense. We'll just
> >>> have to be sure to test on all of the compilers we use. I think we'll
> >>> be fine for Visual Studio (I found this link for VS2015, but not one
> >>> yet for VS2017, but C++11 looks well covered even in VS2015:
> >>> https://msdn.microsoft.com/en-us/library/hh567368.aspx).  I think the
> >>> only other compilers that are used are GCC and Clang and they should
> >>> both be fine on all of the platforms.  (Assuming Emscripten/Clang is
> >>> fine.)  We'll just have to be sure to explicitly set C++11 for all of
> >>> our platforms.
> >>>
> >>> On Wed, Apr 26, 2017 at 12:23 PM, Rory Walsh  wrote:
> >>>> I agree.
> >>>>
> >>>> On 26 April 2017 at 17:12, Victor Lazzarini 
> >>>> wrote:
> >>>>>
> >>>>> I think it makes sense to move to c++11 for the whole codebase, after all
> >>>>> it’s 2017. The only issue might be VST,
> >>>>> which appears to have problems with it (I noticed when trying to compile
> >>>>> the VST host opcodes).
> >>>>> ========================
> >>>>> Prof. Victor Lazzarini
> >>>>> Dean of Arts, Celtic Studies, and Philosophy,
> >>>>> Maynooth University,
> >>>>> Maynooth, Co Kildare, Ireland
> >>>>> Tel: 00 353 7086936
> >>>>> Fax: 00 353 1 7086952
> >>>>>
> >>>>>> On 26 Apr 2017, at 17:08, Michael Gogins 
> >>>>>> wrote:
> >>>>>>
> >>>>>> I've created issues in GitHub for removing the Boost dependency from
> >>>>>> the Csound interfaces, from CsoundVST, and from CsoundAC.
> >>>>>>
> >>>>>> I have pushed the fixes for the Csound interfaces and for CsoundVST,
> >>>>>> but not tested them, which I will do later today.
> >>>>>>
> >>>>>> It turns out I had already replaced boost::numeric::ublas with eigen3
> >>>>>> in CsoundAC. It remains to replace Boost formatting and random
> >>>>>> distributions. I am working on that now and expect it to be
> >>>>>> straightforward, but the C++ requirement will move to -std=c++11. It
> >>>>>> looks like the Boost random distributions more or less went into
> >>>>>> C++11.
> >>>>>>
> >>>>>> Best,
> >>>>>> Mike
> >>>>>>
> >>>>>> -----------------------------------------------------
> >>>>>> Michael Gogins
> >>>>>> Irreducible Productions
> >>>>>> http://michaelgogins.tumblr.com
> >>>>>> Michael dot Gogins at gmail dot com
> >>>>>>
> >>>>>>
> >>>>>> On Tue, Apr 25, 2017 at 7:52 PM, Steven Yi  wrote:
> >>>>>>> Thanks Michael!
> >>>>>>>
> >>>>>>> I do think eigen3 is fine as a dependency.  It works at least with
> >>>>>>> vcpkg as far as I can tell.  We'll have to test again once boost is
> >>>>>>> off the list of dependencies as it sort of just killed the whole build
> >>>>>>> and I'm not sure we have any data yet on how long it takes to build
> >>>>>>> without boost.  (Stephen may have a better idea of it all though.)
> >>>>>>>
> >>>>>>> On Tue, Apr 25, 2017 at 6:35 PM, Michael Gogins
> >>>>>>>  wrote:
> >>>>>>>> I will take care of this. I would also like to remove the boost
> >>>>>>>> dependency from CsoundAC etc, not sure how hard that is. The
> >>>>>>>> dependencies regarding matrix arithmetic would be replaced by Eigen
> >>>>>>>> (eigen3), can Appveyor deal with that?
> >>>>>>>>
> >>>>>>>> Best,
> >>>>>>>> Mike
> >>>>>>>>
> >>>>>>>> -----------------------------------------------------
> >>>>>>>> Michael Gogins
> >>>>>>>> Irreducible Productions
> >>>>>>>> http://michaelgogins.tumblr.com
> >>>>>>>> Michael dot Gogins at gmail dot com
> >>>>>>>>
> >>>>>>>>
> >>>>>>>> On Tue, Apr 25, 2017 at 5:09 PM, Steven Yi  wrote:
> >>>>>>>>> Hi All,
> >>>>>>>>>
> >>>>>>>>> I've been working with Stephen Kyne on getting the Visual Studio
> >>>>>>>>> build
> >>>>>>>>> to run on Appveyor (a continuous integration system like Travis).
> >>>>>>>>> We're getting stuck because downloading and building Boost takes a
> >>>>>>>>> long time and thus the build times out.  As far as I can see, we seem
> >>>>>>>>> to have a dependency on Boost in just a few places:
> >>>>>>>>>
> >>>>>>>>> 1. interfaces/CsoundFile.cpp
> >>>>>>>>> 2. CsoundAC
> >>>>>>>>> 3. CsoundVST
> >>>>>>>>>
> >>>>>>>>> I think for now our best bet to get a continuous Windows build with
> >>>>>>>>> Appveyor going is to remove the dependency on Boost from
> >>>>>>>>> interfaces/CsoundFile.cpp and just not build CsoundAC/CsoundVST.
> >>>>>>>>>
> >>>>>>>>> For CsoundFile.cpp, there's just one use of boost:
> >>>>>>>>>
> >>>>>>>>> // TODO: Replace boost throughout interfaces.
> >>>>>>>>> #if !defined(__ANDROID__)
> >>>>>>>>> boost::split(args, buffer, boost::is_any_of(" \t\n\r"),
> >>>>>>>>> boost::token_compress_on);
> >>>>>>>>> #endif
> >>>>>>>>>
> >>>>>>>>> Could someone (Michael perhaps? I'm guessing this might be code of
> >>>>>>>>> yours?) take a look and perhaps find a way to get rid of using Boost
> >>>>>>>>> here?
> >>>>>>>>>
> >>>>>>>>> Thanks!
> >>>>>>>>> steven

yep, I tried it again:

/usr/local/lib/vstsdk2.4/public.sdk/source/vst2.x/audioeffect.cpp:512:20: error: non-constant-expression cannot
      be narrowed from type 'int' to 'char' in initializer list [-Wc++11-narrowing]
                        char temp[2] = {'0' + (char)digit, '\0'};
                                        ^~~~~~~~~~~~~~~~~
/usr/local/lib/vstsdk2.4/public.sdk/source/vst2.x/audioeffect.cpp:512:20: note: insert an explicit cast to
      silence this issue
                        char temp[2] = {'0' + (char)digit, '\0'};
                                        ^~~~~~~~~~~~~~~~~
                                        static_cast( )
1 error generated.


it needs some SDK surgery. Or maybe a compiler switch.
========================
Prof. Victor Lazzarini
Dean of Arts, Celtic Studies, and Philosophy,
Maynooth University,
Maynooth, Co Kildare, Ireland
Tel: 00 353 7086936
Fax: 00 353 1 7086952 

> On 26 Apr 2017, at 19:02, Victor Lazzarini  wrote:
> 
> Well, I might be wrong, but I remember adding -std=c++11 to cmake and making VST4cs fail to build with
> a complaint about an SDK header.
> 
> ========================
> Prof. Victor Lazzarini
> Dean of Arts, Celtic Studies, and Philosophy,
> Maynooth University,
> Maynooth, Co Kildare, Ireland
> Tel: 00 353 7086936
> Fax: 00 353 1 7086952 
> 
>> On 26 Apr 2017, at 18:52, Rory Walsh  wrote:
>> 
>> That's odd, I've been building Cabbage with c++11 for quite some time without any problems, and it use the VST 2.4 SDK?  
>> 
>> On 26 April 2017 at 18:40, Victor Lazzarini  wrote:
>> The problem is not with our code, but with the SDK.
>> ========================
>> Prof. Victor Lazzarini
>> Dean of Arts, Celtic Studies, and Philosophy,
>> Maynooth University,
>> Maynooth, Co Kildare, Ireland
>> Tel: 00 353 7086936
>> Fax: 00 353 1 7086952
>> 
>>> On 26 Apr 2017, at 18:34, Michael Gogins  wrote:
>>> 
>>> I agree that we should move to uniformly requiring at least C++11. If
>>> the VST host opcodes are a problem, I will fix them.
>>> 
>>> Best,
>>> Mike
>>> 
>>> -----------------------------------------------------
>>> Michael Gogins
>>> Irreducible Productions
>>> http://michaelgogins.tumblr.com
>>> Michael dot Gogins at gmail dot com
>>> 
>>> 
>>> On Wed, Apr 26, 2017 at 1:25 PM, Victor Lazzarini
>>>  wrote:
>>>> Some older GCCs (e.g. the one that comes in Debian Jessie) do not have full c++11 support, but I think most
>>>> linuxes now come with the newer gccs (> 6.0 I guess). In raspbian jessie, I had to do get a newer compiler
>>>> package (can’t remember which exactly). It can be a pain sometimes, but I think we should roll out the code for
>>>> c++11.
>>>> 
>>>> ========================
>>>> Prof. Victor Lazzarini
>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>> Maynooth University,
>>>> Maynooth, Co Kildare, Ireland
>>>> Tel: 00 353 7086936
>>>> Fax: 00 353 1 7086952
>>>> 
>>>>> On 26 Apr 2017, at 18:11, Steven Yi  wrote:
>>>>> 
>>>>> Thanks Michael for the update!  I think C++11 makes sense. We'll just
>>>>> have to be sure to test on all of the compilers we use. I think we'll
>>>>> be fine for Visual Studio (I found this link for VS2015, but not one
>>>>> yet for VS2017, but C++11 looks well covered even in VS2015:
>>>>> https://msdn.microsoft.com/en-us/library/hh567368.aspx).  I think the
>>>>> only other compilers that are used are GCC and Clang and they should
>>>>> both be fine on all of the platforms.  (Assuming Emscripten/Clang is
>>>>> fine.)  We'll just have to be sure to explicitly set C++11 for all of
>>>>> our platforms.
>>>>> 
>>>>> On Wed, Apr 26, 2017 at 12:23 PM, Rory Walsh  wrote:
>>>>>> I agree.
>>>>>> 
>>>>>> On 26 April 2017 at 17:12, Victor Lazzarini 
>>>>>> wrote:
>>>>>>> 
>>>>>>> I think it makes sense to move to c++11 for the whole codebase, after all
>>>>>>> it’s 2017. The only issue might be VST,
>>>>>>> which appears to have problems with it (I noticed when trying to compile
>>>>>>> the VST host opcodes).
>>>>>>> ========================
>>>>>>> Prof. Victor Lazzarini
>>>>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>>>>> Maynooth University,
>>>>>>> Maynooth, Co Kildare, Ireland
>>>>>>> Tel: 00 353 7086936
>>>>>>> Fax: 00 353 1 7086952
>>>>>>> 
>>>>>>>> On 26 Apr 2017, at 17:08, Michael Gogins 
>>>>>>>> wrote:
>>>>>>>> 
>>>>>>>> I've created issues in GitHub for removing the Boost dependency from
>>>>>>>> the Csound interfaces, from CsoundVST, and from CsoundAC.
>>>>>>>> 
>>>>>>>> I have pushed the fixes for the Csound interfaces and for CsoundVST,
>>>>>>>> but not tested them, which I will do later today.
>>>>>>>> 
>>>>>>>> It turns out I had already replaced boost::numeric::ublas with eigen3
>>>>>>>> in CsoundAC. It remains to replace Boost formatting and random
>>>>>>>> distributions. I am working on that now and expect it to be
>>>>>>>> straightforward, but the C++ requirement will move to -std=c++11. It
>>>>>>>> looks like the Boost random distributions more or less went into
>>>>>>>> C++11.
>>>>>>>> 
>>>>>>>> Best,
>>>>>>>> Mike
>>>>>>>> 
>>>>>>>> -----------------------------------------------------
>>>>>>>> Michael Gogins
>>>>>>>> Irreducible Productions
>>>>>>>> http://michaelgogins.tumblr.com
>>>>>>>> Michael dot Gogins at gmail dot com
>>>>>>>> 
>>>>>>>> 
>>>>>>>> On Tue, Apr 25, 2017 at 7:52 PM, Steven Yi  wrote:
>>>>>>>>> Thanks Michael!
>>>>>>>>> 
>>>>>>>>> I do think eigen3 is fine as a dependency.  It works at least with
>>>>>>>>> vcpkg as far as I can tell.  We'll have to test again once boost is
>>>>>>>>> off the list of dependencies as it sort of just killed the whole build
>>>>>>>>> and I'm not sure we have any data yet on how long it takes to build
>>>>>>>>> without boost.  (Stephen may have a better idea of it all though.)
>>>>>>>>> 
>>>>>>>>> On Tue, Apr 25, 2017 at 6:35 PM, Michael Gogins
>>>>>>>>>  wrote:
>>>>>>>>>> I will take care of this. I would also like to remove the boost
>>>>>>>>>> dependency from CsoundAC etc, not sure how hard that is. The
>>>>>>>>>> dependencies regarding matrix arithmetic would be replaced by Eigen
>>>>>>>>>> (eigen3), can Appveyor deal with that?
>>>>>>>>>> 
>>>>>>>>>> Best,
>>>>>>>>>> Mike
>>>>>>>>>> 
>>>>>>>>>> -----------------------------------------------------
>>>>>>>>>> Michael Gogins
>>>>>>>>>> Irreducible Productions
>>>>>>>>>> http://michaelgogins.tumblr.com
>>>>>>>>>> Michael dot Gogins at gmail dot com
>>>>>>>>>> 
>>>>>>>>>> 
>>>>>>>>>> On Tue, Apr 25, 2017 at 5:09 PM, Steven Yi  wrote:
>>>>>>>>>>> Hi All,
>>>>>>>>>>> 
>>>>>>>>>>> I've been working with Stephen Kyne on getting the Visual Studio
>>>>>>>>>>> build
>>>>>>>>>>> to run on Appveyor (a continuous integration system like Travis).
>>>>>>>>>>> We're getting stuck because downloading and building Boost takes a
>>>>>>>>>>> long time and thus the build times out.  As far as I can see, we seem
>>>>>>>>>>> to have a dependency on Boost in just a few places:
>>>>>>>>>>> 
>>>>>>>>>>> 1. interfaces/CsoundFile.cpp
>>>>>>>>>>> 2. CsoundAC
>>>>>>>>>>> 3. CsoundVST
>>>>>>>>>>> 
>>>>>>>>>>> I think for now our best bet to get a continuous Windows build with
>>>>>>>>>>> Appveyor going is to remove the dependency on Boost from
>>>>>>>>>>> interfaces/CsoundFile.cpp and just not build CsoundAC/CsoundVST.
>>>>>>>>>>> 
>>>>>>>>>>> For CsoundFile.cpp, there's just one use of boost:
>>>>>>>>>>> 
>>>>>>>>>>> // TODO: Replace boost throughout interfaces.
>>>>>>>>>>> #if !defined(__ANDROID__)
>>>>>>>>>>> boost::split(args, buffer, boost::is_any_of(" \t\n\r"),
>>>>>>>>>>> boost::token_compress_on);
>>>>>>>>>>> #endif
>>>>>>>>>>> 
>>>>>>>>>>> Could someone (Michael perhaps? I'm guessing this might be code of
>>>>>>>>>>> yours?) take a look and perhaps find a way to get rid of using Boost
>>>>>>>>>>> here?
>>>>>>>>>>> 
>>>>>>>>>>> Thanks!
>>>>>>>>>>> steven
>>>>>>> 
>>>>>> 
>>>> 
>> 
>> 
> 

There may be a pragma to turn off -Wc++11-narrowing.

Best,
Mike

-----------------------------------------------------
Michael Gogins
Irreducible Productions
http://michaelgogins.tumblr.com
Michael dot Gogins at gmail dot com


On Wed, Apr 26, 2017 at 2:33 PM, Rory Walsh  wrote:
> I'm not sure how JUCE does it but I am sure It doesn't modify the VST SDK.
>
> On 26 April 2017 at 19:14, Victor Lazzarini 
> wrote:
>>
>> yep, I tried it again:
>>
>> /usr/local/lib/vstsdk2.4/public.sdk/source/vst2.x/audioeffect.cpp:512:20:
>> error: non-constant-expression cannot
>>       be narrowed from type 'int' to 'char' in initializer list
>> [-Wc++11-narrowing]
>>                         char temp[2] = {'0' + (char)digit, '\0'};
>>                                         ^~~~~~~~~~~~~~~~~
>> /usr/local/lib/vstsdk2.4/public.sdk/source/vst2.x/audioeffect.cpp:512:20:
>> note: insert an explicit cast to
>>       silence this issue
>>                         char temp[2] = {'0' + (char)digit, '\0'};
>>                                         ^~~~~~~~~~~~~~~~~
>>                                         static_cast( )
>> 1 error generated.
>>
>>
>> it needs some SDK surgery. Or maybe a compiler switch.
>> ========================
>> Prof. Victor Lazzarini
>> Dean of Arts, Celtic Studies, and Philosophy,
>> Maynooth University,
>> Maynooth, Co Kildare, Ireland
>> Tel: 00 353 7086936
>> Fax: 00 353 1 7086952
>>
>> > On 26 Apr 2017, at 19:02, Victor Lazzarini 
>> > wrote:
>> >
>> > Well, I might be wrong, but I remember adding -std=c++11 to cmake and
>> > making VST4cs fail to build with
>> > a complaint about an SDK header.
>> >
>> > ========================
>> > Prof. Victor Lazzarini
>> > Dean of Arts, Celtic Studies, and Philosophy,
>> > Maynooth University,
>> > Maynooth, Co Kildare, Ireland
>> > Tel: 00 353 7086936
>> > Fax: 00 353 1 7086952
>> >
>> >> On 26 Apr 2017, at 18:52, Rory Walsh  wrote:
>> >>
>> >> That's odd, I've been building Cabbage with c++11 for quite some time
>> >> without any problems, and it use the VST 2.4 SDK?
>> >>
>> >> On 26 April 2017 at 18:40, Victor Lazzarini 
>> >> wrote:
>> >> The problem is not with our code, but with the SDK.
>> >> ========================
>> >> Prof. Victor Lazzarini
>> >> Dean of Arts, Celtic Studies, and Philosophy,
>> >> Maynooth University,
>> >> Maynooth, Co Kildare, Ireland
>> >> Tel: 00 353 7086936
>> >> Fax: 00 353 1 7086952
>> >>
>> >>> On 26 Apr 2017, at 18:34, Michael Gogins 
>> >>> wrote:
>> >>>
>> >>> I agree that we should move to uniformly requiring at least C++11. If
>> >>> the VST host opcodes are a problem, I will fix them.
>> >>>
>> >>> Best,
>> >>> Mike
>> >>>
>> >>> -----------------------------------------------------
>> >>> Michael Gogins
>> >>> Irreducible Productions
>> >>> http://michaelgogins.tumblr.com
>> >>> Michael dot Gogins at gmail dot com
>> >>>
>> >>>
>> >>> On Wed, Apr 26, 2017 at 1:25 PM, Victor Lazzarini
>> >>>  wrote:
>> >>>> Some older GCCs (e.g. the one that comes in Debian Jessie) do not
>> >>>> have full c++11 support, but I think most
>> >>>> linuxes now come with the newer gccs (> 6.0 I guess). In raspbian
>> >>>> jessie, I had to do get a newer compiler
>> >>>> package (can’t remember which exactly). It can be a pain sometimes,
>> >>>> but I think we should roll out the code for
>> >>>> c++11.
>> >>>>
>> >>>> ========================
>> >>>> Prof. Victor Lazzarini
>> >>>> Dean of Arts, Celtic Studies, and Philosophy,
>> >>>> Maynooth University,
>> >>>> Maynooth, Co Kildare, Ireland
>> >>>> Tel: 00 353 7086936
>> >>>> Fax: 00 353 1 7086952
>> >>>>
>> >>>>> On 26 Apr 2017, at 18:11, Steven Yi  wrote:
>> >>>>>
>> >>>>> Thanks Michael for the update!  I think C++11 makes sense. We'll
>> >>>>> just
>> >>>>> have to be sure to test on all of the compilers we use. I think
>> >>>>> we'll
>> >>>>> be fine for Visual Studio (I found this link for VS2015, but not one
>> >>>>> yet for VS2017, but C++11 looks well covered even in VS2015:
>> >>>>> https://msdn.microsoft.com/en-us/library/hh567368.aspx).  I think
>> >>>>> the
>> >>>>> only other compilers that are used are GCC and Clang and they should
>> >>>>> both be fine on all of the platforms.  (Assuming Emscripten/Clang is
>> >>>>> fine.)  We'll just have to be sure to explicitly set C++11 for all
>> >>>>> of
>> >>>>> our platforms.
>> >>>>>
>> >>>>> On Wed, Apr 26, 2017 at 12:23 PM, Rory Walsh 
>> >>>>> wrote:
>> >>>>>> I agree.
>> >>>>>>
>> >>>>>> On 26 April 2017 at 17:12, Victor Lazzarini
>> >>>>>> 
>> >>>>>> wrote:
>> >>>>>>>
>> >>>>>>> I think it makes sense to move to c++11 for the whole codebase,
>> >>>>>>> after all
>> >>>>>>> it’s 2017. The only issue might be VST,
>> >>>>>>> which appears to have problems with it (I noticed when trying to
>> >>>>>>> compile
>> >>>>>>> the VST host opcodes).
>> >>>>>>> ========================
>> >>>>>>> Prof. Victor Lazzarini
>> >>>>>>> Dean of Arts, Celtic Studies, and Philosophy,
>> >>>>>>> Maynooth University,
>> >>>>>>> Maynooth, Co Kildare, Ireland
>> >>>>>>> Tel: 00 353 7086936
>> >>>>>>> Fax: 00 353 1 7086952
>> >>>>>>>
>> >>>>>>>> On 26 Apr 2017, at 17:08, Michael Gogins
>> >>>>>>>> 
>> >>>>>>>> wrote:
>> >>>>>>>>
>> >>>>>>>> I've created issues in GitHub for removing the Boost dependency
>> >>>>>>>> from
>> >>>>>>>> the Csound interfaces, from CsoundVST, and from CsoundAC.
>> >>>>>>>>
>> >>>>>>>> I have pushed the fixes for the Csound interfaces and for
>> >>>>>>>> CsoundVST,
>> >>>>>>>> but not tested them, which I will do later today.
>> >>>>>>>>
>> >>>>>>>> It turns out I had already replaced boost::numeric::ublas with
>> >>>>>>>> eigen3
>> >>>>>>>> in CsoundAC. It remains to replace Boost formatting and random
>> >>>>>>>> distributions. I am working on that now and expect it to be
>> >>>>>>>> straightforward, but the C++ requirement will move to -std=c++11.
>> >>>>>>>> It
>> >>>>>>>> looks like the Boost random distributions more or less went into
>> >>>>>>>> C++11.
>> >>>>>>>>
>> >>>>>>>> Best,
>> >>>>>>>> Mike
>> >>>>>>>>
>> >>>>>>>> -----------------------------------------------------
>> >>>>>>>> Michael Gogins
>> >>>>>>>> Irreducible Productions
>> >>>>>>>> http://michaelgogins.tumblr.com
>> >>>>>>>> Michael dot Gogins at gmail dot com
>> >>>>>>>>
>> >>>>>>>>
>> >>>>>>>> On Tue, Apr 25, 2017 at 7:52 PM, Steven Yi 
>> >>>>>>>> wrote:
>> >>>>>>>>> Thanks Michael!
>> >>>>>>>>>
>> >>>>>>>>> I do think eigen3 is fine as a dependency.  It works at least
>> >>>>>>>>> with
>> >>>>>>>>> vcpkg as far as I can tell.  We'll have to test again once boost
>> >>>>>>>>> is
>> >>>>>>>>> off the list of dependencies as it sort of just killed the whole
>> >>>>>>>>> build
>> >>>>>>>>> and I'm not sure we have any data yet on how long it takes to
>> >>>>>>>>> build
>> >>>>>>>>> without boost.  (Stephen may have a better idea of it all
>> >>>>>>>>> though.)
>> >>>>>>>>>
>> >>>>>>>>> On Tue, Apr 25, 2017 at 6:35 PM, Michael Gogins
>> >>>>>>>>>  wrote:
>> >>>>>>>>>> I will take care of this. I would also like to remove the boost
>> >>>>>>>>>> dependency from CsoundAC etc, not sure how hard that is. The
>> >>>>>>>>>> dependencies regarding matrix arithmetic would be replaced by
>> >>>>>>>>>> Eigen
>> >>>>>>>>>> (eigen3), can Appveyor deal with that?
>> >>>>>>>>>>
>> >>>>>>>>>> Best,
>> >>>>>>>>>> Mike
>> >>>>>>>>>>
>> >>>>>>>>>> -----------------------------------------------------
>> >>>>>>>>>> Michael Gogins
>> >>>>>>>>>> Irreducible Productions
>> >>>>>>>>>> http://michaelgogins.tumblr.com
>> >>>>>>>>>> Michael dot Gogins at gmail dot com
>> >>>>>>>>>>
>> >>>>>>>>>>
>> >>>>>>>>>> On Tue, Apr 25, 2017 at 5:09 PM, Steven Yi 
>> >>>>>>>>>> wrote:
>> >>>>>>>>>>> Hi All,
>> >>>>>>>>>>>
>> >>>>>>>>>>> I've been working with Stephen Kyne on getting the Visual
>> >>>>>>>>>>> Studio
>> >>>>>>>>>>> build
>> >>>>>>>>>>> to run on Appveyor (a continuous integration system like
>> >>>>>>>>>>> Travis).
>> >>>>>>>>>>> We're getting stuck because downloading and building Boost
>> >>>>>>>>>>> takes a
>> >>>>>>>>>>> long time and thus the build times out.  As far as I can see,
>> >>>>>>>>>>> we seem
>> >>>>>>>>>>> to have a dependency on Boost in just a few places:
>> >>>>>>>>>>>
>> >>>>>>>>>>> 1. interfaces/CsoundFile.cpp
>> >>>>>>>>>>> 2. CsoundAC
>> >>>>>>>>>>> 3. CsoundVST
>> >>>>>>>>>>>
>> >>>>>>>>>>> I think for now our best bet to get a continuous Windows build
>> >>>>>>>>>>> with
>> >>>>>>>>>>> Appveyor going is to remove the dependency on Boost from
>> >>>>>>>>>>> interfaces/CsoundFile.cpp and just not build
>> >>>>>>>>>>> CsoundAC/CsoundVST.
>> >>>>>>>>>>>
>> >>>>>>>>>>> For CsoundFile.cpp, there's just one use of boost:
>> >>>>>>>>>>>
>> >>>>>>>>>>> // TODO: Replace boost throughout interfaces.
>> >>>>>>>>>>> #if !defined(__ANDROID__)
>> >>>>>>>>>>> boost::split(args, buffer, boost::is_any_of(" \t\n\r"),
>> >>>>>>>>>>> boost::token_compress_on);
>> >>>>>>>>>>> #endif
>> >>>>>>>>>>>
>> >>>>>>>>>>> Could someone (Michael perhaps? I'm guessing this might be
>> >>>>>>>>>>> code of
>> >>>>>>>>>>> yours?) take a look and perhaps find a way to get rid of using
>> >>>>>>>>>>> Boost
>> >>>>>>>>>>> here?
>> >>>>>>>>>>>
>> >>>>>>>>>>> Thanks!
>> >>>>>>>>>>> steven
>> >>>>>>>
>> >>>>>>
>> >>>>
>> >>
>> >>
>> >
>>

yes, I assume there is something like that. Or -Wno-c++11-narrowing.
========================
Prof. Victor Lazzarini
Dean of Arts, Celtic Studies, and Philosophy,
Maynooth University,
Maynooth, Co Kildare, Ireland
Tel: 00 353 7086936
Fax: 00 353 1 7086952 

> On 26 Apr 2017, at 20:05, Michael Gogins  wrote:
> 
> There may be a pragma to turn off -Wc++11-narrowing.
> 
> Best,
> Mike
> 
> -----------------------------------------------------
> Michael Gogins
> Irreducible Productions
> http://michaelgogins.tumblr.com
> Michael dot Gogins at gmail dot com
> 
> 
> On Wed, Apr 26, 2017 at 2:33 PM, Rory Walsh  wrote:
>> I'm not sure how JUCE does it but I am sure It doesn't modify the VST SDK.
>> 
>> On 26 April 2017 at 19:14, Victor Lazzarini 
>> wrote:
>>> 
>>> yep, I tried it again:
>>> 
>>> /usr/local/lib/vstsdk2.4/public.sdk/source/vst2.x/audioeffect.cpp:512:20:
>>> error: non-constant-expression cannot
>>>      be narrowed from type 'int' to 'char' in initializer list
>>> [-Wc++11-narrowing]
>>>                        char temp[2] = {'0' + (char)digit, '\0'};
>>>                                        ^~~~~~~~~~~~~~~~~
>>> /usr/local/lib/vstsdk2.4/public.sdk/source/vst2.x/audioeffect.cpp:512:20:
>>> note: insert an explicit cast to
>>>      silence this issue
>>>                        char temp[2] = {'0' + (char)digit, '\0'};
>>>                                        ^~~~~~~~~~~~~~~~~
>>>                                        static_cast( )
>>> 1 error generated.
>>> 
>>> 
>>> it needs some SDK surgery. Or maybe a compiler switch.
>>> ========================
>>> Prof. Victor Lazzarini
>>> Dean of Arts, Celtic Studies, and Philosophy,
>>> Maynooth University,
>>> Maynooth, Co Kildare, Ireland
>>> Tel: 00 353 7086936
>>> Fax: 00 353 1 7086952
>>> 
>>>> On 26 Apr 2017, at 19:02, Victor Lazzarini 
>>>> wrote:
>>>> 
>>>> Well, I might be wrong, but I remember adding -std=c++11 to cmake and
>>>> making VST4cs fail to build with
>>>> a complaint about an SDK header.
>>>> 
>>>> ========================
>>>> Prof. Victor Lazzarini
>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>> Maynooth University,
>>>> Maynooth, Co Kildare, Ireland
>>>> Tel: 00 353 7086936
>>>> Fax: 00 353 1 7086952
>>>> 
>>>>> On 26 Apr 2017, at 18:52, Rory Walsh  wrote:
>>>>> 
>>>>> That's odd, I've been building Cabbage with c++11 for quite some time
>>>>> without any problems, and it use the VST 2.4 SDK?
>>>>> 
>>>>> On 26 April 2017 at 18:40, Victor Lazzarini 
>>>>> wrote:
>>>>> The problem is not with our code, but with the SDK.
>>>>> ========================
>>>>> Prof. Victor Lazzarini
>>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>>> Maynooth University,
>>>>> Maynooth, Co Kildare, Ireland
>>>>> Tel: 00 353 7086936
>>>>> Fax: 00 353 1 7086952
>>>>> 
>>>>>> On 26 Apr 2017, at 18:34, Michael Gogins 
>>>>>> wrote:
>>>>>> 
>>>>>> I agree that we should move to uniformly requiring at least C++11. If
>>>>>> the VST host opcodes are a problem, I will fix them.
>>>>>> 
>>>>>> Best,
>>>>>> Mike
>>>>>> 
>>>>>> -----------------------------------------------------
>>>>>> Michael Gogins
>>>>>> Irreducible Productions
>>>>>> http://michaelgogins.tumblr.com
>>>>>> Michael dot Gogins at gmail dot com
>>>>>> 
>>>>>> 
>>>>>> On Wed, Apr 26, 2017 at 1:25 PM, Victor Lazzarini
>>>>>>  wrote:
>>>>>>> Some older GCCs (e.g. the one that comes in Debian Jessie) do not
>>>>>>> have full c++11 support, but I think most
>>>>>>> linuxes now come with the newer gccs (> 6.0 I guess). In raspbian
>>>>>>> jessie, I had to do get a newer compiler
>>>>>>> package (can’t remember which exactly). It can be a pain sometimes,
>>>>>>> but I think we should roll out the code for
>>>>>>> c++11.
>>>>>>> 
>>>>>>> ========================
>>>>>>> Prof. Victor Lazzarini
>>>>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>>>>> Maynooth University,
>>>>>>> Maynooth, Co Kildare, Ireland
>>>>>>> Tel: 00 353 7086936
>>>>>>> Fax: 00 353 1 7086952
>>>>>>> 
>>>>>>>> On 26 Apr 2017, at 18:11, Steven Yi  wrote:
>>>>>>>> 
>>>>>>>> Thanks Michael for the update!  I think C++11 makes sense. We'll
>>>>>>>> just
>>>>>>>> have to be sure to test on all of the compilers we use. I think
>>>>>>>> we'll
>>>>>>>> be fine for Visual Studio (I found this link for VS2015, but not one
>>>>>>>> yet for VS2017, but C++11 looks well covered even in VS2015:
>>>>>>>> https://msdn.microsoft.com/en-us/library/hh567368.aspx).  I think
>>>>>>>> the
>>>>>>>> only other compilers that are used are GCC and Clang and they should
>>>>>>>> both be fine on all of the platforms.  (Assuming Emscripten/Clang is
>>>>>>>> fine.)  We'll just have to be sure to explicitly set C++11 for all
>>>>>>>> of
>>>>>>>> our platforms.
>>>>>>>> 
>>>>>>>> On Wed, Apr 26, 2017 at 12:23 PM, Rory Walsh 
>>>>>>>> wrote:
>>>>>>>>> I agree.
>>>>>>>>> 
>>>>>>>>> On 26 April 2017 at 17:12, Victor Lazzarini
>>>>>>>>> 
>>>>>>>>> wrote:
>>>>>>>>>> 
>>>>>>>>>> I think it makes sense to move to c++11 for the whole codebase,
>>>>>>>>>> after all
>>>>>>>>>> it’s 2017. The only issue might be VST,
>>>>>>>>>> which appears to have problems with it (I noticed when trying to
>>>>>>>>>> compile
>>>>>>>>>> the VST host opcodes).
>>>>>>>>>> ========================
>>>>>>>>>> Prof. Victor Lazzarini
>>>>>>>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>>>>>>>> Maynooth University,
>>>>>>>>>> Maynooth, Co Kildare, Ireland
>>>>>>>>>> Tel: 00 353 7086936
>>>>>>>>>> Fax: 00 353 1 7086952
>>>>>>>>>> 
>>>>>>>>>>> On 26 Apr 2017, at 17:08, Michael Gogins
>>>>>>>>>>> 
>>>>>>>>>>> wrote:
>>>>>>>>>>> 
>>>>>>>>>>> I've created issues in GitHub for removing the Boost dependency
>>>>>>>>>>> from
>>>>>>>>>>> the Csound interfaces, from CsoundVST, and from CsoundAC.
>>>>>>>>>>> 
>>>>>>>>>>> I have pushed the fixes for the Csound interfaces and for
>>>>>>>>>>> CsoundVST,
>>>>>>>>>>> but not tested them, which I will do later today.
>>>>>>>>>>> 
>>>>>>>>>>> It turns out I had already replaced boost::numeric::ublas with
>>>>>>>>>>> eigen3
>>>>>>>>>>> in CsoundAC. It remains to replace Boost formatting and random
>>>>>>>>>>> distributions. I am working on that now and expect it to be
>>>>>>>>>>> straightforward, but the C++ requirement will move to -std=c++11.
>>>>>>>>>>> It
>>>>>>>>>>> looks like the Boost random distributions more or less went into
>>>>>>>>>>> C++11.
>>>>>>>>>>> 
>>>>>>>>>>> Best,
>>>>>>>>>>> Mike
>>>>>>>>>>> 
>>>>>>>>>>> -----------------------------------------------------
>>>>>>>>>>> Michael Gogins
>>>>>>>>>>> Irreducible Productions
>>>>>>>>>>> http://michaelgogins.tumblr.com
>>>>>>>>>>> Michael dot Gogins at gmail dot com
>>>>>>>>>>> 
>>>>>>>>>>> 
>>>>>>>>>>> On Tue, Apr 25, 2017 at 7:52 PM, Steven Yi 
>>>>>>>>>>> wrote:
>>>>>>>>>>>> Thanks Michael!
>>>>>>>>>>>> 
>>>>>>>>>>>> I do think eigen3 is fine as a dependency.  It works at least
>>>>>>>>>>>> with
>>>>>>>>>>>> vcpkg as far as I can tell.  We'll have to test again once boost
>>>>>>>>>>>> is
>>>>>>>>>>>> off the list of dependencies as it sort of just killed the whole
>>>>>>>>>>>> build
>>>>>>>>>>>> and I'm not sure we have any data yet on how long it takes to
>>>>>>>>>>>> build
>>>>>>>>>>>> without boost.  (Stephen may have a better idea of it all
>>>>>>>>>>>> though.)
>>>>>>>>>>>> 
>>>>>>>>>>>> On Tue, Apr 25, 2017 at 6:35 PM, Michael Gogins
>>>>>>>>>>>>  wrote:
>>>>>>>>>>>>> I will take care of this. I would also like to remove the boost
>>>>>>>>>>>>> dependency from CsoundAC etc, not sure how hard that is. The
>>>>>>>>>>>>> dependencies regarding matrix arithmetic would be replaced by
>>>>>>>>>>>>> Eigen
>>>>>>>>>>>>> (eigen3), can Appveyor deal with that?
>>>>>>>>>>>>> 
>>>>>>>>>>>>> Best,
>>>>>>>>>>>>> Mike
>>>>>>>>>>>>> 
>>>>>>>>>>>>> -----------------------------------------------------
>>>>>>>>>>>>> Michael Gogins
>>>>>>>>>>>>> Irreducible Productions
>>>>>>>>>>>>> http://michaelgogins.tumblr.com
>>>>>>>>>>>>> Michael dot Gogins at gmail dot com
>>>>>>>>>>>>> 
>>>>>>>>>>>>> 
>>>>>>>>>>>>> On Tue, Apr 25, 2017 at 5:09 PM, Steven Yi 
>>>>>>>>>>>>> wrote:
>>>>>>>>>>>>>> Hi All,
>>>>>>>>>>>>>> 
>>>>>>>>>>>>>> I've been working with Stephen Kyne on getting the Visual
>>>>>>>>>>>>>> Studio
>>>>>>>>>>>>>> build
>>>>>>>>>>>>>> to run on Appveyor (a continuous integration system like
>>>>>>>>>>>>>> Travis).
>>>>>>>>>>>>>> We're getting stuck because downloading and building Boost
>>>>>>>>>>>>>> takes a
>>>>>>>>>>>>>> long time and thus the build times out.  As far as I can see,
>>>>>>>>>>>>>> we seem
>>>>>>>>>>>>>> to have a dependency on Boost in just a few places:
>>>>>>>>>>>>>> 
>>>>>>>>>>>>>> 1. interfaces/CsoundFile.cpp
>>>>>>>>>>>>>> 2. CsoundAC
>>>>>>>>>>>>>> 3. CsoundVST
>>>>>>>>>>>>>> 
>>>>>>>>>>>>>> I think for now our best bet to get a continuous Windows build
>>>>>>>>>>>>>> with
>>>>>>>>>>>>>> Appveyor going is to remove the dependency on Boost from
>>>>>>>>>>>>>> interfaces/CsoundFile.cpp and just not build
>>>>>>>>>>>>>> CsoundAC/CsoundVST.
>>>>>>>>>>>>>> 
>>>>>>>>>>>>>> For CsoundFile.cpp, there's just one use of boost:
>>>>>>>>>>>>>> 
>>>>>>>>>>>>>> // TODO: Replace boost throughout interfaces.
>>>>>>>>>>>>>> #if !defined(__ANDROID__)
>>>>>>>>>>>>>> boost::split(args, buffer, boost::is_any_of(" \t\n\r"),
>>>>>>>>>>>>>> boost::token_compress_on);
>>>>>>>>>>>>>> #endif
>>>>>>>>>>>>>> 
>>>>>>>>>>>>>> Could someone (Michael perhaps? I'm guessing this might be
>>>>>>>>>>>>>> code of
>>>>>>>>>>>>>> yours?) take a look and perhaps find a way to get rid of using
>>>>>>>>>>>>>> Boost
>>>>>>>>>>>>>> here?
>>>>>>>>>>>>>> 
>>>>>>>>>>>>>> Thanks!
>>>>>>>>>>>>>> steven
>>>>>>>>>> 
>>>>>>>>> 
>>>>>>> 
>>>>> 
>>>>> 
>

-Wno-c++11-narrowing does it.
========================
Prof. Victor Lazzarini
Dean of Arts, Celtic Studies, and Philosophy,
Maynooth University,
Maynooth, Co Kildare, Ireland
Tel: 00 353 7086936
Fax: 00 353 1 7086952 

> On 26 Apr 2017, at 20:08, Victor Lazzarini  wrote:
> 
> yes, I assume there is something like that. Or -Wno-c++11-narrowing.
> ========================
> Prof. Victor Lazzarini
> Dean of Arts, Celtic Studies, and Philosophy,
> Maynooth University,
> Maynooth, Co Kildare, Ireland
> Tel: 00 353 7086936
> Fax: 00 353 1 7086952 
> 
>> On 26 Apr 2017, at 20:05, Michael Gogins  wrote:
>> 
>> There may be a pragma to turn off -Wc++11-narrowing.
>> 
>> Best,
>> Mike
>> 
>> -----------------------------------------------------
>> Michael Gogins
>> Irreducible Productions
>> http://michaelgogins.tumblr.com
>> Michael dot Gogins at gmail dot com
>> 
>> 
>> On Wed, Apr 26, 2017 at 2:33 PM, Rory Walsh  wrote:
>>> I'm not sure how JUCE does it but I am sure It doesn't modify the VST SDK.
>>> 
>>> On 26 April 2017 at 19:14, Victor Lazzarini 
>>> wrote:
>>>> 
>>>> yep, I tried it again:
>>>> 
>>>> /usr/local/lib/vstsdk2.4/public.sdk/source/vst2.x/audioeffect.cpp:512:20:
>>>> error: non-constant-expression cannot
>>>>     be narrowed from type 'int' to 'char' in initializer list
>>>> [-Wc++11-narrowing]
>>>>                       char temp[2] = {'0' + (char)digit, '\0'};
>>>>                                       ^~~~~~~~~~~~~~~~~
>>>> /usr/local/lib/vstsdk2.4/public.sdk/source/vst2.x/audioeffect.cpp:512:20:
>>>> note: insert an explicit cast to
>>>>     silence this issue
>>>>                       char temp[2] = {'0' + (char)digit, '\0'};
>>>>                                       ^~~~~~~~~~~~~~~~~
>>>>                                       static_cast( )
>>>> 1 error generated.
>>>> 
>>>> 
>>>> it needs some SDK surgery. Or maybe a compiler switch.
>>>> ========================
>>>> Prof. Victor Lazzarini
>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>> Maynooth University,
>>>> Maynooth, Co Kildare, Ireland
>>>> Tel: 00 353 7086936
>>>> Fax: 00 353 1 7086952
>>>> 
>>>>> On 26 Apr 2017, at 19:02, Victor Lazzarini 
>>>>> wrote:
>>>>> 
>>>>> Well, I might be wrong, but I remember adding -std=c++11 to cmake and
>>>>> making VST4cs fail to build with
>>>>> a complaint about an SDK header.
>>>>> 
>>>>> ========================
>>>>> Prof. Victor Lazzarini
>>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>>> Maynooth University,
>>>>> Maynooth, Co Kildare, Ireland
>>>>> Tel: 00 353 7086936
>>>>> Fax: 00 353 1 7086952
>>>>> 
>>>>>> On 26 Apr 2017, at 18:52, Rory Walsh  wrote:
>>>>>> 
>>>>>> That's odd, I've been building Cabbage with c++11 for quite some time
>>>>>> without any problems, and it use the VST 2.4 SDK?
>>>>>> 
>>>>>> On 26 April 2017 at 18:40, Victor Lazzarini 
>>>>>> wrote:
>>>>>> The problem is not with our code, but with the SDK.
>>>>>> ========================
>>>>>> Prof. Victor Lazzarini
>>>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>>>> Maynooth University,
>>>>>> Maynooth, Co Kildare, Ireland
>>>>>> Tel: 00 353 7086936
>>>>>> Fax: 00 353 1 7086952
>>>>>> 
>>>>>>> On 26 Apr 2017, at 18:34, Michael Gogins 
>>>>>>> wrote:
>>>>>>> 
>>>>>>> I agree that we should move to uniformly requiring at least C++11. If
>>>>>>> the VST host opcodes are a problem, I will fix them.
>>>>>>> 
>>>>>>> Best,
>>>>>>> Mike
>>>>>>> 
>>>>>>> -----------------------------------------------------
>>>>>>> Michael Gogins
>>>>>>> Irreducible Productions
>>>>>>> http://michaelgogins.tumblr.com
>>>>>>> Michael dot Gogins at gmail dot com
>>>>>>> 
>>>>>>> 
>>>>>>> On Wed, Apr 26, 2017 at 1:25 PM, Victor Lazzarini
>>>>>>>  wrote:
>>>>>>>> Some older GCCs (e.g. the one that comes in Debian Jessie) do not
>>>>>>>> have full c++11 support, but I think most
>>>>>>>> linuxes now come with the newer gccs (> 6.0 I guess). In raspbian
>>>>>>>> jessie, I had to do get a newer compiler
>>>>>>>> package (can’t remember which exactly). It can be a pain sometimes,
>>>>>>>> but I think we should roll out the code for
>>>>>>>> c++11.
>>>>>>>> 
>>>>>>>> ========================
>>>>>>>> Prof. Victor Lazzarini
>>>>>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>>>>>> Maynooth University,
>>>>>>>> Maynooth, Co Kildare, Ireland
>>>>>>>> Tel: 00 353 7086936
>>>>>>>> Fax: 00 353 1 7086952
>>>>>>>> 
>>>>>>>>> On 26 Apr 2017, at 18:11, Steven Yi  wrote:
>>>>>>>>> 
>>>>>>>>> Thanks Michael for the update!  I think C++11 makes sense. We'll
>>>>>>>>> just
>>>>>>>>> have to be sure to test on all of the compilers we use. I think
>>>>>>>>> we'll
>>>>>>>>> be fine for Visual Studio (I found this link for VS2015, but not one
>>>>>>>>> yet for VS2017, but C++11 looks well covered even in VS2015:
>>>>>>>>> https://msdn.microsoft.com/en-us/library/hh567368.aspx).  I think
>>>>>>>>> the
>>>>>>>>> only other compilers that are used are GCC and Clang and they should
>>>>>>>>> both be fine on all of the platforms.  (Assuming Emscripten/Clang is
>>>>>>>>> fine.)  We'll just have to be sure to explicitly set C++11 for all
>>>>>>>>> of
>>>>>>>>> our platforms.
>>>>>>>>> 
>>>>>>>>> On Wed, Apr 26, 2017 at 12:23 PM, Rory Walsh 
>>>>>>>>> wrote:
>>>>>>>>>> I agree.
>>>>>>>>>> 
>>>>>>>>>> On 26 April 2017 at 17:12, Victor Lazzarini
>>>>>>>>>> 
>>>>>>>>>> wrote:
>>>>>>>>>>> 
>>>>>>>>>>> I think it makes sense to move to c++11 for the whole codebase,
>>>>>>>>>>> after all
>>>>>>>>>>> it’s 2017. The only issue might be VST,
>>>>>>>>>>> which appears to have problems with it (I noticed when trying to
>>>>>>>>>>> compile
>>>>>>>>>>> the VST host opcodes).
>>>>>>>>>>> ========================
>>>>>>>>>>> Prof. Victor Lazzarini
>>>>>>>>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>>>>>>>>> Maynooth University,
>>>>>>>>>>> Maynooth, Co Kildare, Ireland
>>>>>>>>>>> Tel: 00 353 7086936
>>>>>>>>>>> Fax: 00 353 1 7086952
>>>>>>>>>>> 
>>>>>>>>>>>> On 26 Apr 2017, at 17:08, Michael Gogins
>>>>>>>>>>>> 
>>>>>>>>>>>> wrote:
>>>>>>>>>>>> 
>>>>>>>>>>>> I've created issues in GitHub for removing the Boost dependency
>>>>>>>>>>>> from
>>>>>>>>>>>> the Csound interfaces, from CsoundVST, and from CsoundAC.
>>>>>>>>>>>> 
>>>>>>>>>>>> I have pushed the fixes for the Csound interfaces and for
>>>>>>>>>>>> CsoundVST,
>>>>>>>>>>>> but not tested them, which I will do later today.
>>>>>>>>>>>> 
>>>>>>>>>>>> It turns out I had already replaced boost::numeric::ublas with
>>>>>>>>>>>> eigen3
>>>>>>>>>>>> in CsoundAC. It remains to replace Boost formatting and random
>>>>>>>>>>>> distributions. I am working on that now and expect it to be
>>>>>>>>>>>> straightforward, but the C++ requirement will move to -std=c++11.
>>>>>>>>>>>> It
>>>>>>>>>>>> looks like the Boost random distributions more or less went into
>>>>>>>>>>>> C++11.
>>>>>>>>>>>> 
>>>>>>>>>>>> Best,
>>>>>>>>>>>> Mike
>>>>>>>>>>>> 
>>>>>>>>>>>> -----------------------------------------------------
>>>>>>>>>>>> Michael Gogins
>>>>>>>>>>>> Irreducible Productions
>>>>>>>>>>>> http://michaelgogins.tumblr.com
>>>>>>>>>>>> Michael dot Gogins at gmail dot com
>>>>>>>>>>>> 
>>>>>>>>>>>> 
>>>>>>>>>>>> On Tue, Apr 25, 2017 at 7:52 PM, Steven Yi 
>>>>>>>>>>>> wrote:
>>>>>>>>>>>>> Thanks Michael!
>>>>>>>>>>>>> 
>>>>>>>>>>>>> I do think eigen3 is fine as a dependency.  It works at least
>>>>>>>>>>>>> with
>>>>>>>>>>>>> vcpkg as far as I can tell.  We'll have to test again once boost
>>>>>>>>>>>>> is
>>>>>>>>>>>>> off the list of dependencies as it sort of just killed the whole
>>>>>>>>>>>>> build
>>>>>>>>>>>>> and I'm not sure we have any data yet on how long it takes to
>>>>>>>>>>>>> build
>>>>>>>>>>>>> without boost.  (Stephen may have a better idea of it all
>>>>>>>>>>>>> though.)
>>>>>>>>>>>>> 
>>>>>>>>>>>>> On Tue, Apr 25, 2017 at 6:35 PM, Michael Gogins
>>>>>>>>>>>>>  wrote:
>>>>>>>>>>>>>> I will take care of this. I would also like to remove the boost
>>>>>>>>>>>>>> dependency from CsoundAC etc, not sure how hard that is. The
>>>>>>>>>>>>>> dependencies regarding matrix arithmetic would be replaced by
>>>>>>>>>>>>>> Eigen
>>>>>>>>>>>>>> (eigen3), can Appveyor deal with that?
>>>>>>>>>>>>>> 
>>>>>>>>>>>>>> Best,
>>>>>>>>>>>>>> Mike
>>>>>>>>>>>>>> 
>>>>>>>>>>>>>> -----------------------------------------------------
>>>>>>>>>>>>>> Michael Gogins
>>>>>>>>>>>>>> Irreducible Productions
>>>>>>>>>>>>>> http://michaelgogins.tumblr.com
>>>>>>>>>>>>>> Michael dot Gogins at gmail dot com
>>>>>>>>>>>>>> 
>>>>>>>>>>>>>> 
>>>>>>>>>>>>>> On Tue, Apr 25, 2017 at 5:09 PM, Steven Yi 
>>>>>>>>>>>>>> wrote:
>>>>>>>>>>>>>>> Hi All,
>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>> I've been working with Stephen Kyne on getting the Visual
>>>>>>>>>>>>>>> Studio
>>>>>>>>>>>>>>> build
>>>>>>>>>>>>>>> to run on Appveyor (a continuous integration system like
>>>>>>>>>>>>>>> Travis).
>>>>>>>>>>>>>>> We're getting stuck because downloading and building Boost
>>>>>>>>>>>>>>> takes a
>>>>>>>>>>>>>>> long time and thus the build times out.  As far as I can see,
>>>>>>>>>>>>>>> we seem
>>>>>>>>>>>>>>> to have a dependency on Boost in just a few places:
>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>> 1. interfaces/CsoundFile.cpp
>>>>>>>>>>>>>>> 2. CsoundAC
>>>>>>>>>>>>>>> 3. CsoundVST
>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>> I think for now our best bet to get a continuous Windows build
>>>>>>>>>>>>>>> with
>>>>>>>>>>>>>>> Appveyor going is to remove the dependency on Boost from
>>>>>>>>>>>>>>> interfaces/CsoundFile.cpp and just not build
>>>>>>>>>>>>>>> CsoundAC/CsoundVST.
>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>> For CsoundFile.cpp, there's just one use of boost:
>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>> // TODO: Replace boost throughout interfaces.
>>>>>>>>>>>>>>> #if !defined(__ANDROID__)
>>>>>>>>>>>>>>> boost::split(args, buffer, boost::is_any_of(" \t\n\r"),
>>>>>>>>>>>>>>> boost::token_compress_on);
>>>>>>>>>>>>>>> #endif
>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>> Could someone (Michael perhaps? I'm guessing this might be
>>>>>>>>>>>>>>> code of
>>>>>>>>>>>>>>> yours?) take a look and perhaps find a way to get rid of using
>>>>>>>>>>>>>>> Boost
>>>>>>>>>>>>>>> here?
>>>>>>>>>>>>>>> 
>>>>>>>>>>>>>>> Thanks!
>>>>>>>>>>>>>>> steven
>>>>>>>>>>> 
>>>>>>>>>> 
>>>>>>>> 
>>>>>> 

Hi all,

Thanks to the boost removal, the build time was greatly reduced. It's now building on appveyor in about 30 minutes. Once I get the caching working it should be quicker from then on. The appveyor.yml config file should be almost complete at this stage, it's even packaging all binaries to a zip file which is available as an artifact to download.

I'd like to get the csound installer for windows working next and have that produced as the artifact. Not sure the work involved there but we would have an end to end automation then. From there we can just add new features/opcodes to the build and maintain some stability.

Also for the VST SDK, the latest version seems backwards compatible with VST2. It might be worth moving to support both if it was deemed beneficial. I'd assume it at least supports C++ 11/14. 

Thanks,
Stephen

---

From: Csound-developers <CSOUND-DEV@LISTSERV.HEANET.IE> on behalf of Victor Lazzarini <Victor.Lazzarini@NUIM.IE>
Sent: 26 April 2017 20:21
To: CSOUND-DEV@LISTSERV.HEANET.IE
Subject: Re: [Csnd-dev] Removing boost from CsoundFile.cpp

 

-Wno-c++11-narrowing does it.
========================
Prof. Victor Lazzarini
Dean of Arts, Celtic Studies, and Philosophy,
Maynooth University,
Maynooth, Co Kildare, Ireland
Tel: 00 353 7086936
Fax: 00 353 1 7086952

> On 26 Apr 2017, at 20:08, Victor Lazzarini <Victor.Lazzarini@nuim.ie> wrote:
>
> yes, I assume there is something like that. Or -Wno-c++11-narrowing.
> ========================
> Prof. Victor Lazzarini
> Dean of Arts, Celtic Studies, and Philosophy,
> Maynooth University,
> Maynooth, Co Kildare, Ireland
> Tel: 00 353 7086936
> Fax: 00 353 1 7086952
>
>> On 26 Apr 2017, at 20:05, Michael Gogins <michael.gogins@GMAIL.COM> wrote:
>>
>> There may be a pragma to turn off -Wc++11-narrowing.
>>
>> Best,
>> Mike
>>
>> -----------------------------------------------------
>> Michael Gogins
>> Irreducible Productions
>> http://michaelgogins.tumblr.com

Michael Gogins michaelgogins.tumblr.com Computer music, photography, some poetry and fiction, and something more or less like theology or philosophy of religion

>> Michael dot Gogins at gmail dot com
>>
>>
>> On Wed, Apr 26, 2017 at 2:33 PM, Rory Walsh <rorywalsh@ear.ie> wrote:
>>> I'm not sure how JUCE does it but I am sure It doesn't modify the VST SDK.
>>>
>>> On 26 April 2017 at 19:14, Victor Lazzarini <Victor.Lazzarini@nuim.ie>
>>> wrote:
>>>>
>>>> yep, I tried it again:
>>>>
>>>> /usr/local/lib/vstsdk2.4/public.sdk/source/vst2.x/audioeffect.cpp:512:20:
>>>> error: non-constant-expression cannot
>>>>     be narrowed from type 'int' to 'char' in initializer list
>>>> [-Wc++11-narrowing]
>>>>                       char temp[2] = {'0' + (char)digit, '\0'};
>>>>                                       ^~~~~~~~~~~~~~~~~
>>>> /usr/local/lib/vstsdk2.4/public.sdk/source/vst2.x/audioeffect.cpp:512:20:
>>>> note: insert an explicit cast to
>>>>     silence this issue
>>>>                       char temp[2] = {'0' + (char)digit, '\0'};
>>>>                                       ^~~~~~~~~~~~~~~~~
>>>>                                       static_cast<char>( )
>>>> 1 error generated.
>>>>
>>>>
>>>> it needs some SDK surgery. Or maybe a compiler switch.
>>>> ========================
>>>> Prof. Victor Lazzarini
>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>> Maynooth University,
>>>> Maynooth, Co Kildare, Ireland
>>>> Tel: 00 353 7086936
>>>> Fax: 00 353 1 7086952
>>>>
>>>>> On 26 Apr 2017, at 19:02, Victor Lazzarini <Victor.Lazzarini@NUIM.IE>
>>>>> wrote:
>>>>>
>>>>> Well, I might be wrong, but I remember adding -std=c++11 to cmake and
>>>>> making VST4cs fail to build with
>>>>> a complaint about an SDK header.
>>>>>
>>>>> ========================
>>>>> Prof. Victor Lazzarini
>>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>>> Maynooth University,
>>>>> Maynooth, Co Kildare, Ireland
>>>>> Tel: 00 353 7086936
>>>>> Fax: 00 353 1 7086952
>>>>>
>>>>>> On 26 Apr 2017, at 18:52, Rory Walsh <rorywalsh@ear.ie> wrote:
>>>>>>
>>>>>> That's odd, I've been building Cabbage with c++11 for quite some time
>>>>>> without any problems, and it use the VST 2.4 SDK?
>>>>>>
>>>>>> On 26 April 2017 at 18:40, Victor Lazzarini <Victor.Lazzarini@nuim.ie>
>>>>>> wrote:
>>>>>> The problem is not with our code, but with the SDK.
>>>>>> ========================
>>>>>> Prof. Victor Lazzarini
>>>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>>>> Maynooth University,
>>>>>> Maynooth, Co Kildare, Ireland
>>>>>> Tel: 00 353 7086936
>>>>>> Fax: 00 353 1 7086952
>>>>>>
>>>>>>> On 26 Apr 2017, at 18:34, Michael Gogins <michael.gogins@GMAIL.COM>
>>>>>>> wrote:
>>>>>>>
>>>>>>> I agree that we should move to uniformly requiring at least C++11. If
>>>>>>> the VST host opcodes are a problem, I will fix them.
>>>>>>>
>>>>>>> Best,
>>>>>>> Mike
>>>>>>>
>>>>>>> -----------------------------------------------------
>>>>>>> Michael Gogins
>>>>>>> Irreducible Productions
>>>>>>> http://michaelgogins.tumblr.com

Michael Gogins michaelgogins.tumblr.com Computer music, photography, some poetry and fiction, and something more or less like theology or philosophy of religion

>>>>>>> Michael dot Gogins at gmail dot com
>>>>>>>
>>>>>>>
>>>>>>> On Wed, Apr 26, 2017 at 1:25 PM, Victor Lazzarini
>>>>>>> <Victor.Lazzarini@nuim.ie> wrote:
>>>>>>>> Some older GCCs (e.g. the one that comes in Debian Jessie) do not
>>>>>>>> have full c++11 support, but I think most
>>>>>>>> linuxes now come with the newer gccs (> 6.0 I guess). In raspbian
>>>>>>>> jessie, I had to do get a newer compiler
>>>>>>>> package (can’t remember which exactly). It can be a pain sometimes,
>>>>>>>> but I think we should roll out the code for
>>>>>>>> c++11.
>>>>>>>>
>>>>>>>> ========================
>>>>>>>> Prof. Victor Lazzarini
>>>>>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>>>>>> Maynooth University,
>>>>>>>> Maynooth, Co Kildare, Ireland
>>>>>>>> Tel: 00 353 7086936
>>>>>>>> Fax: 00 353 1 7086952
>>>>>>>>
>>>>>>>>> On 26 Apr 2017, at 18:11, Steven Yi <stevenyi@gmail.com> wrote:
>>>>>>>>>
>>>>>>>>> Thanks Michael for the update!  I think C++11 makes sense. We'll
>>>>>>>>> just
>>>>>>>>> have to be sure to test on all of the compilers we use. I think
>>>>>>>>> we'll
>>>>>>>>> be fine for Visual Studio (I found this link for VS2015, but not one
>>>>>>>>> yet for VS2017, but C++11 looks well covered even in VS2015:
>>>>>>>>> https://msdn.microsoft.com/en-us/library/hh567368.aspx).  I think

Support For C++11-14-17 Features (Modern C++) msdn.microsoft.com Visual C++ implements the vast majority of features in the C++11 core language specification, as well as many C++14 Library features and some features proposed for C++17.

>>>>>>>>> the
>>>>>>>>> only other compilers that are used are GCC and Clang and they should
>>>>>>>>> both be fine on all of the platforms.  (Assuming Emscripten/Clang is
>>>>>>>>> fine.)  We'll just have to be sure to explicitly set C++11 for all
>>>>>>>>> of
>>>>>>>>> our platforms.
>>>>>>>>>
>>>>>>>>> On Wed, Apr 26, 2017 at 12:23 PM, Rory Walsh <rorywalsh@ear.ie>
>>>>>>>>> wrote:
>>>>>>>>>> I agree.
>>>>>>>>>>
>>>>>>>>>> On 26 April 2017 at 17:12, Victor Lazzarini
>>>>>>>>>> <Victor.Lazzarini@nuim.ie>
>>>>>>>>>> wrote:
>>>>>>>>>>>
>>>>>>>>>>> I think it makes sense to move to c++11 for the whole codebase,
>>>>>>>>>>> after all
>>>>>>>>>>> it’s 2017. The only issue might be VST,
>>>>>>>>>>> which appears to have problems with it (I noticed when trying to
>>>>>>>>>>> compile
>>>>>>>>>>> the VST host opcodes).
>>>>>>>>>>> ========================
>>>>>>>>>>> Prof. Victor Lazzarini
>>>>>>>>>>> Dean of Arts, Celtic Studies, and Philosophy,
>>>>>>>>>>> Maynooth University,
>>>>>>>>>>> Maynooth, Co Kildare, Ireland
>>>>>>>>>>> Tel: 00 353 7086936
>>>>>>>>>>> Fax: 00 353 1 7086952
>>>>>>>>>>>
>>>>>>>>>>>> On 26 Apr 2017, at 17:08, Michael Gogins
>>>>>>>>>>>> <michael.gogins@gmail.com>
>>>>>>>>>>>> wrote:
>>>>>>>>>>>>
>>>>>>>>>>>> I've created issues in GitHub for removing the Boost dependency
>>>>>>>>>>>> from
>>>>>>>>>>>> the Csound interfaces, from CsoundVST, and from CsoundAC.
>>>>>>>>>>>>
>>>>>>>>>>>> I have pushed the fixes for the Csound interfaces and for
>>>>>>>>>>>> CsoundVST,
>>>>>>>>>>>> but not tested them, which I will do later today.
>>>>>>>>>>>>
>>>>>>>>>>>> It turns out I had already replaced boost::numeric::ublas with
>>>>>>>>>>>> eigen3
>>>>>>>>>>>> in CsoundAC. It remains to replace Boost formatting and random
>>>>>>>>>>>> distributions. I am working on that now and expect it to be
>>>>>>>>>>>> straightforward, but the C++ requirement will move to -std=c++11.
>>>>>>>>>>>> It
>>>>>>>>>>>> looks like the Boost random distributions more or less went into
>>>>>>>>>>>> C++11.
>>>>>>>>>>>>
>>>>>>>>>>>> Best,
>>>>>>>>>>>> Mike
>>>>>>>>>>>>
>>>>>>>>>>>> -----------------------------------------------------
>>>>>>>>>>>> Michael Gogins
>>>>>>>>>>>> Irreducible Productions
>>>>>>>>>>>> http://michaelgogins.tumblr.com

Michael Gogins michaelgogins.tumblr.com Computer music, photography, some poetry and fiction, and something more or less like theology or philosophy of religion

>>>>>>>>>>>> Michael dot Gogins at gmail dot com
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>> On Tue, Apr 25, 2017 at 7:52 PM, Steven Yi <stevenyi@gmail.com>
>>>>>>>>>>>> wrote:
>>>>>>>>>>>>> Thanks Michael!
>>>>>>>>>>>>>
>>>>>>>>>>>>> I do think eigen3 is fine as a dependency.  It works at least
>>>>>>>>>>>>> with
>>>>>>>>>>>>> vcpkg as far as I can tell.  We'll have to test again once boost
>>>>>>>>>>>>> is
>>>>>>>>>>>>> off the list of dependencies as it sort of just killed the whole
>>>>>>>>>>>>> build
>>>>>>>>>>>>> and I'm not sure we have any data yet on how long it takes to
>>>>>>>>>>>>> build
>>>>>>>>>>>>> without boost.  (Stephen may have a better idea of it all
>>>>>>>>>>>>> though.)
>>>>>>>>>>>>>
>>>>>>>>>>>>> On Tue, Apr 25, 2017 at 6:35 PM, Michael Gogins
>>>>>>>>>>>>> <michael.gogins@gmail.com> wrote:
>>>>>>>>>>>>>> I will take care of this. I would also like to remove the boost
>>>>>>>>>>>>>> dependency from CsoundAC etc, not sure how hard that is. The
>>>>>>>>>>>>>> dependencies regarding matrix arithmetic would be replaced by
>>>>>>>>>>>>>> Eigen
>>>>>>>>>>>>>> (eigen3), can Appveyor deal with that?
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Best,
>>>>>>>>>>>>>> Mike
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> -----------------------------------------------------
>>>>>>>>>>>>>> Michael Gogins
>>>>>>>>>>>>>> Irreducible Productions
>>>>>>>>>>>>>> http://michaelgogins.tumblr.com

Michael Gogins michaelgogins.tumblr.com Computer music, photography, some poetry and fiction, and something more or less like theology or philosophy of religion

>>>>>>>>>>>>>> Michael dot Gogins at gmail dot com
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> On Tue, Apr 25, 2017 at 5:09 PM, Steven Yi <stevenyi@gmail.com>
>>>>>>>>>>>>>> wrote:
>>>>>>>>>>>>>>> Hi All,
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> I've been working with Stephen Kyne on getting the Visual
>>>>>>>>>>>>>>> Studio
>>>>>>>>>>>>>>> build
>>>>>>>>>>>>>>> to run on Appveyor (a continuous integration system like
>>>>>>>>>>>>>>> Travis).
>>>>>>>>>>>>>>> We're getting stuck because downloading and building Boost
>>>>>>>>>>>>>>> takes a
>>>>>>>>>>>>>>> long time and thus the build times out.  As far as I can see,
>>>>>>>>>>>>>>> we seem
>>>>>>>>>>>>>>> to have a dependency on Boost in just a few places:
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> 1. interfaces/CsoundFile.cpp
>>>>>>>>>>>>>>> 2. CsoundAC
>>>>>>>>>>>>>>> 3. CsoundVST
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> I think for now our best bet to get a continuous Windows build
>>>>>>>>>>>>>>> with
>>>>>>>>>>>>>>> Appveyor going is to remove the dependency on Boost from
>>>>>>>>>>>>>>> interfaces/CsoundFile.cpp and just not build
>>>>>>>>>>>>>>> CsoundAC/CsoundVST.
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> For CsoundFile.cpp, there's just one use of boost:
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> // TODO: Replace boost throughout interfaces.
>>>>>>>>>>>>>>> #if !defined(__ANDROID__)
>>>>>>>>>>>>>>> boost::split(args, buffer, boost::is_any_of(" \t\n\r"),
>>>>>>>>>>>>>>> boost::token_compress_on);
>>>>>>>>>>>>>>> #endif
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> Could someone (Michael perhaps? I'm guessing this might be
>>>>>>>>>>>>>>> code of
>>>>>>>>>>>>>>> yours?) take a look and perhaps find a way to get rid of using
>>>>>>>>>>>>>>> Boost
>>>>>>>>>>>>>>> here?
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> Thanks!
>>>>>>>>>>>>>>> steven
>>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>
>>>
>