[PD-dev] 0.48.0 release status
Dan Wilcox
danomatika at gmail.com
Sat Jul 22 01:28:10 CEST 2017
Right, don't need the _WIN32 for MINGW which has GNU libc.
> On Jul 22, 2017, at 1:26 AM, Miller Puckette <msp at ucsd.edu> wrote:
>
> Wouldn't it be...
>
> #if !defined(_MSC_VER) || (_MSC_VER >= 17000))
>
> i.e., always true unlese we're MSVC and < 17000 ?
>
> (I think the line below would stop it from being compiled in with mingw).
>
> cheers
> Miller
>
> On Sat, Jul 22, 2017 at 12:35:23AM +0200, Dan Wilcox wrote:
>> Simplest fix is to replace the
>>
>> #ifndef _WIN32
>>
>> with
>>
>> #if !defined(_WIN32) || (defined(_MSC_VER) && (_MSC_VER >= 17000))
>>
>> I think that is pretty safe and can be added now without any real issues. If Pd is compiled with a newer MSVC than 2008, boom those are available.
>>
>> Another, longer & probably more difficult option, is to provide implementations for MSVC < 2012 such as:
>>
>> // versions of MSVC prior to 2012 do not have these
>> #if defined(_WIN32) && defined(_MSC_VER) && (_MSC_VER < 17000)
>> #define asinh(x) log(x + sqrt(x2 + 1))
>> #define acosh(x) log(x + sqrt(x2 - 1))
>> #define atanh(x) (log(1+x) - log(1-x))/2
>> #define isnan(x) _isnan(x)
>> #define isinf(x) (!_finite(x))
>> ...
>> #endif
>>
>> I have no idea exactly how express calls the root C functions, so this approach may not work anyway.
>>
>>> On Jul 21, 2017, at 10:39 PM, Miller Puckette <msp at ucsd.edu> wrote:
>>>
>>> OK, will try that for next release :)
>>>
>>> Anyhow, I have to profile MSVC against mingw and (hopefully) move the
>>> whole compile chain over. But certainly not now.
>>>
>>> Miller
>>>
>>> On Fri, Jul 21, 2017 at 10:30:26PM +0200, Dan Wilcox wrote:
>>>> Looks like the preprocessor define is _MSC_VER: https://stackoverflow.com/questions/70013/how-to-detect-if-im-compiling-code-with-visual-studio-2008 <https://stackoverflow.com/questions/70013/how-to-detect-if-im-compiling-code-with-visual-studio-2008>
>>>>
>>>> So something like the following could probably work:
>>>>
>>>> #ifdef _WIN32
>>>> #if (_MSC_VER < 17000)
>>>> #define ATANH_NOT_AVAILABLE
>>>> #endif
>>>> #endif
>>>> #include <math.h>
>>>>
>>>>> On Jul 21, 2017, at 10:21 PM, Dan Wilcox <danomatika at gmail.com> wrote:
>>>>>
>>>>> It's most probably working because you are compiling on a newer system with a newer Visual Studio which has the newer math.h which has the function.
>>>>>
>>>>> Older versions pre-2012 do not, hence the current non-usage.
>>>>>
>>>>>> On Jul 21, 2017, at 9:55 PM, Alexandre Torres Porres <porres at gmail.com <mailto:porres at gmail.com>> wrote:
>>>>>>
>>>>>>
>>>>>>
>>>>>> 2017-07-21 16:44 GMT-03:00 Dan Wilcox <danomatika at gmail.com <mailto:danomatika at gmail.com>>:
>>>>>> Those functions are in the windows math.h if you're using Visual Studio 2012 or greater.
>>>>>>
>>>>>> Hey, I just realized we have an [atanh] object for cyclone, I actually did that one myself, and I just tested, it's working on windows!
>>>>>>
>>>>>> here's the source https://github.com/porres/pd-cyclone/blob/master/cyclone_src/binaries/control/atanh.c <https://github.com/porres/pd-cyclone/blob/master/cyclone_src/binaries/control/atanh.c>
>>>>>>
>>>>>> so, definitely, those functions may exist for windows, not sure what building system you're using, but it may need an update
>>>>>
>>>>> --------
>>>>> Dan Wilcox
>>>>> @danomatika <http://twitter.com/danomatika>
>>>>> danomatika.com <http://danomatika.com/>
>>>>> robotcowboy.com <http://robotcowboy.com/>
>>>>>
>>>>>
>>>>>
>>>>
>>>> --------
>>>> Dan Wilcox
>>>> @danomatika <http://twitter.com/danomatika>
>>>> danomatika.com <http://danomatika.com/>
>>>> robotcowboy.com <http://robotcowboy.com/>
>>>>
>>>>
>>>>
>>
>> --------
>> Dan Wilcox
>> @danomatika <http://twitter.com/danomatika>
>> danomatika.com <http://danomatika.com/>
>> robotcowboy.com <http://robotcowboy.com/>
>>
>>
>>
--------
Dan Wilcox
@danomatika <http://twitter.com/danomatika>
danomatika.com <http://danomatika.com/>
robotcowboy.com <http://robotcowboy.com/>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.puredata.info/pipermail/pd-dev/attachments/20170722/00a03a34/attachment-0001.html>
More information about the Pd-dev
mailing list