On Monday, 23 March 2015 at 14:19:34 UTC, Ben Boeckel wrote:
> On Mon, Mar 23, 2015 at 13:49:53 +0000, Szymon Gatner via Digitalmars-d-announce wrote:
>> Been waiting for this for almost 2 years, might as well wait for another release. Seriously tho, it is surprising how much little attention is put into C++/D integration considering all the recent fuss about this. And yes, I've tried mixed C++/D app on x64 Win, it crashes with simple writeln() call.
>
> At least you get that far; Android doesn't even get there (though those
> patches to even let people get toolchains up and running are "only"
> around a year old).
>
> --Ben

i really try not to be whiny about it but it is sooo frustrating. d advertises itself as easy to integrate with c/c++ and maybe in theory it is but in practice it is not true at all. simplest example from Adam's book I followed crashed miserable so I can only assume that -nobody- is mixing c++ with d on Win. I need it working on win so we can even start to think about using D for our projects but indeed iOS and Android is very next

On Monday, 23 March 2015 at 15:36:42 UTC, Martin Nowak wrote:
> On 03/23/2015 02:49 PM, Szymon Gatner wrote:
>> Been waiting for this for almost 2 years, might as well wait for another
>> release. Seriously tho, it is surprising how much little attention is
>> put into C++/D integration considering all the recent fuss about this.
>> And yes, I've tried mixed C++/D app on x64 Win, it crashes with simple
>> writeln() call.
>
> As a matter of fact we only have few Windows developers, so unless
> someone with interest in this platform steps up it will hardly improve.
> Daniel Murphy did a lot of D/C++ integration to get ddmd to work, and
> apparently, that also works on Windows, so you might ask him for details.
> https://auto-tester.puremagic.com/?projectid=10

sure I could bug individuals to make things work for me, then discover another problem rinse and repeat. thing is i dont want things to somehow work (possibly only until next release that will breake it) I want to know that this is something that is cared about and can be considered mature (and i am looking for the weakest definition of "mature" possible)

On Monday, 23 March 2015 at 13:49:55 UTC, Szymon Gatner wrote:
> On Friday, 20 March 2015 at 15:19:03 UTC, Martin Nowak wrote:
>> On 03/19/2015 08:02 AM, Rainer Schuetze wrote:
>>> The COFF32 lib is built through win64.mak. This is an excerpt from my
>>> build script to create lib32\phobos32mscoff.lib:
>>> 
>>> set dm_make=c:\l\dmc\bin\make
>>> set vs=vs12
>>> set vcdir=c:\l\%vs%\vc
>>> set cl32=%vcdir%/bin/cl.exe
>>> set ar32=%vcdir%/bin/lib.exe
>>> set MSLINK=%vcdir%\bin\link.exe
>>> set lib32coff=m:\s\d\rainers\lib32;%vcdir%\vc\lib;%sdkdir%\lib
>>> set sdkdir=c:\Program Files (x86)\Microsoft SDKs\Windows\v7.1A
>>> 
>>> set LINKCMD=%MSLINK%
>>> set LIB=%LIB32COFF%
>>> set ARGS=DMD=%DMD% MAKE=%dm_make% "CC=\"%cl32%\"" "AR=\"%ar32%\""
>>> VCDIR=%vcdir% "SDKDIR=%sdkdir%"
>>> 
>>> cd druntime
>>> %dm_make% -f win64.mak MODEL=32mscoff %ARGS% target
>>> if errorlevel 1 goto xit
>>> cd ..
>>> 
>>> cd phobos
>>> %dm_make% -f win64.mak MODEL=32mscoff %ARGS%
>>> LIB=..\lib32\phobos32mscoff.lib
>>> if errorlevel 1 goto xit
>>> cd ..
>>
>> It's seriously too late, unless someone else steps up and does it. I'd
>> need to change the build script to clean and build another Windows
>> target, and my time is bound to regression fixing.
>>
>
> Been waiting for this for almost 2 years, might as well wait for another release. Seriously tho, it is surprising how much little attention is put into C++/D integration considering all the recent fuss about this. And yes, I've tried mixed C++/D app on x64 Win, it crashes with simple writeln() call.

bug report?

On Monday, 23 March 2015 at 16:01:51 UTC, John Colvin wrote:
> On Monday, 23 March 2015 at 13:49:55 UTC, Szymon Gatner wrote:
>> On Friday, 20 March 2015 at 15:19:03 UTC, Martin Nowak wrote:
>>> On 03/19/2015 08:02 AM, Rainer Schuetze wrote:
>>>> The COFF32 lib is built through win64.mak. This is an excerpt from my
>>>> build script to create lib32\phobos32mscoff.lib:
>>>> 
>>>> set dm_make=c:\l\dmc\bin\make
>>>> set vs=vs12
>>>> set vcdir=c:\l\%vs%\vc
>>>> set cl32=%vcdir%/bin/cl.exe
>>>> set ar32=%vcdir%/bin/lib.exe
>>>> set MSLINK=%vcdir%\bin\link.exe
>>>> set lib32coff=m:\s\d\rainers\lib32;%vcdir%\vc\lib;%sdkdir%\lib
>>>> set sdkdir=c:\Program Files (x86)\Microsoft SDKs\Windows\v7.1A
>>>> 
>>>> set LINKCMD=%MSLINK%
>>>> set LIB=%LIB32COFF%
>>>> set ARGS=DMD=%DMD% MAKE=%dm_make% "CC=\"%cl32%\"" "AR=\"%ar32%\""
>>>> VCDIR=%vcdir% "SDKDIR=%sdkdir%"
>>>> 
>>>> cd druntime
>>>> %dm_make% -f win64.mak MODEL=32mscoff %ARGS% target
>>>> if errorlevel 1 goto xit
>>>> cd ..
>>>> 
>>>> cd phobos
>>>> %dm_make% -f win64.mak MODEL=32mscoff %ARGS%
>>>> LIB=..\lib32\phobos32mscoff.lib
>>>> if errorlevel 1 goto xit
>>>> cd ..
>>>
>>> It's seriously too late, unless someone else steps up and does it. I'd
>>> need to change the build script to clean and build another Windows
>>> target, and my time is bound to regression fixing.
>>>
>>
>> Been waiting for this for almost 2 years, might as well wait for another release. Seriously tho, it is surprising how much little attention is put into C++/D integration considering all the recent fuss about this. And yes, I've tried mixed C++/D app on x64 Win, it crashes with simple writeln() call.
>
> bug report?


i am typing from my phone so hard to find but i did post it on forums some time ago

On Monday, 23 March 2015 at 16:05:55 UTC, Szymon Gatner wrote:
>
>
> i am typing from my phone so hard to find but i did post it on forums some time ago

Posting about something on the forums won't get it fixed, make a bug report if you're encountering a bug.

On Monday, 23 March 2015 at 16:05:55 UTC, Szymon Gatner wrote:
> On Monday, 23 March 2015 at 16:01:51 UTC, John Colvin wrote:
>> On Monday, 23 March 2015 at 13:49:55 UTC, Szymon Gatner wrote:
>>> On Friday, 20 March 2015 at 15:19:03 UTC, Martin Nowak wrote:
>>>> On 03/19/2015 08:02 AM, Rainer Schuetze wrote:
>>>>> The COFF32 lib is built through win64.mak. This is an excerpt from my
>>>>> build script to create lib32\phobos32mscoff.lib:
>>>>> 
>>>>> set dm_make=c:\l\dmc\bin\make
>>>>> set vs=vs12
>>>>> set vcdir=c:\l\%vs%\vc
>>>>> set cl32=%vcdir%/bin/cl.exe
>>>>> set ar32=%vcdir%/bin/lib.exe
>>>>> set MSLINK=%vcdir%\bin\link.exe
>>>>> set lib32coff=m:\s\d\rainers\lib32;%vcdir%\vc\lib;%sdkdir%\lib
>>>>> set sdkdir=c:\Program Files (x86)\Microsoft SDKs\Windows\v7.1A
>>>>> 
>>>>> set LINKCMD=%MSLINK%
>>>>> set LIB=%LIB32COFF%
>>>>> set ARGS=DMD=%DMD% MAKE=%dm_make% "CC=\"%cl32%\"" "AR=\"%ar32%\""
>>>>> VCDIR=%vcdir% "SDKDIR=%sdkdir%"
>>>>> 
>>>>> cd druntime
>>>>> %dm_make% -f win64.mak MODEL=32mscoff %ARGS% target
>>>>> if errorlevel 1 goto xit
>>>>> cd ..
>>>>> 
>>>>> cd phobos
>>>>> %dm_make% -f win64.mak MODEL=32mscoff %ARGS%
>>>>> LIB=..\lib32\phobos32mscoff.lib
>>>>> if errorlevel 1 goto xit
>>>>> cd ..
>>>>
>>>> It's seriously too late, unless someone else steps up and does it. I'd
>>>> need to change the build script to clean and build another Windows
>>>> target, and my time is bound to regression fixing.
>>>>
>>>
>>> Been waiting for this for almost 2 years, might as well wait for another release. Seriously tho, it is surprising how much little attention is put into C++/D integration considering all the recent fuss about this. And yes, I've tried mixed C++/D app on x64 Win, it crashes with simple writeln() call.
>>
>> bug report?
>
>
> i am typing from my phone so hard to find but i did post it on forums some time ago

here:
http://forum.dlang.org/thread/hsglkscatlniiuacpkhe@forum.dlang.org#post-mailman.397.1409844359.5783.digitalmars-d-learn:40puremagic.com

On Monday, 23 March 2015 at 16:11:12 UTC, Kapps wrote:
> On Monday, 23 March 2015 at 16:05:55 UTC, Szymon Gatner wrote:
>>
>>
>> i am typing from my phone so hard to find but i did post it on forums some time ago
>
> Posting about something on the forums won't get it fixed, make a bug report if you're encountering a bug.

i posted it on forums because i wasnt sure if it was indeed a bug or just me. also it is hard to provide minimal app when mixing code and under msvc as potentially full project is needed. i do get your point tho

"Szymon Gatner"  wrote in message news:oofoormyfxkefokvkuzz@forum.dlang.org...

> i really try not to be whiny about it but it is sooo frustrating. d advertises itself as easy to integrate with c/c++ and maybe in theory it is but in practice it is not true at all. simplest example from Adam's book I followed crashed miserable so I can only assume that -nobody- is mixing c++ with d on Win.

DDMD is mixing D and C++ on all the autotester platforms.  It's not that simple, but it should be possible. 

"Szymon Gatner"  wrote in message news:tthkrzwwobmdzbufedye@forum.dlang.org...

> sure I could bug individuals to make things work for me, then discover another problem rinse and repeat. thing is i dont want things to somehow work (possibly only until next release that will breake it) I want to know that this is something that is cared about and can be considered mature (and i am looking for the weakest definition of "mature" possible)

The only way bugs will get fixed is if you file them.  D/C++ interop is definitely _not_ mature, but it is cared about.

eg This release will fix most of the issues with using variadic functions across the D/C++ boundary. 

On Tuesday, 24 March 2015 at 00:26:13 UTC, Daniel Murphy wrote:
> "Szymon Gatner"  wrote in message news:oofoormyfxkefokvkuzz@forum.dlang.org...
>
>> i really try not to be whiny about it but it is sooo frustrating. d advertises itself as easy to integrate with c/c++ and maybe in theory it is but in practice it is not true at all. simplest example from Adam's book I followed crashed miserable so I can only assume that -nobody- is mixing c++ with d on Win.
>
> DDMD is mixing D and C++ on all the autotester platforms.  It's not that simple, but it should be possible.

Honestly, I am not willing to try even less mature compiler (last month it was finished?) to try the feature that does not work in the reference one.

Don't get me wrong, we do want to be early adopters, but there is nothing to adopt yet.