Jonathan Kelly wrote:
> skaller wrote:
>   
>> On Thu, 2006-09-07 at 23:39 +1000, Jonathan Kelly wrote:
>>
>>   
>>     
>>> I was trying to use the mingw within the cygwin environment ... and put 
>>> the mingw directories first in the path. The idea comes from the mingw 
>>> site somewhere ...
>>>     
>>>       
>> Ah, then you need:
>>
>> ./configure --target=nocygwin
>>
>> This is for developing under Cygwin, but making binaries which
>> don't require Cygwin1.dll. This, you can actually run the results
>> directly from CMD.EXE prompt like:
>>
>> bin\flx_arun somefile.dll
>>
>> after somefile.dll is built by Felix. However you still
>> compile and develop code under Cygwin.
>>
>>
>>   
>>     
> For posterity, and others maybe thinking like me ... (must remember 
> "reply-all"!!
>
> Doesn't seem to be the case. Just to be clear, I have cygwin AND the 
> mingw toolset (but not MSYS) installed, and have put the mingw directory 
> first in my bash path. When I do this, configure reports the os as WIN32 
> and not cygwin.
>
>   

This was John's reply that got off-list because I didn't reply all ...

On Fri, 2006-09-08 at 22:49 +1000, Jonathan Kelly wrote:


>> > > ./configure --target=nocygwin
>>     

> > Doesn't seem to be the case. Just to be clear, I have cygwin AND the 
> > mingw toolset (but not MSYS) installed, and have put the mingw directory 
> > first in my bash path. 
>   

You shouldn't do that. Unkludge your PATH  :) 

To call mingw gcc from Cygwin you do this:

        gcc -mnocygwin .. ...

You do not call the mingw gcc directly. That is what the

        --target=nocygwin

means. You will see that if you trace the configuration
diagnostics .. the target platform is changed to win32.
However the GNU toolchain is selected, not the MSVC++
toolchain. The GNU compiler selected is

        g++ -mnocywgin

which is the Cygwin gcc NOT the mingw gcc. However the
Cygwin gcc recognizes the -mnocygwin option and switches
to the mingw compiler and libraries 'under the covers'.

A key point of this is that filenames are still given
in Unix format, not Win32 format .. the compiler still
generates pure win32 cygwin1.dll independent binaries,
but it names them with unix conventions. Thus you write:

        g++ -mnocygwin fred/program.cxx

rather than

        g++ -mnocygwin fred\program.cxx

which of course bash wouldn't parse properly ..  :) 

In theory, what you're doing SHOULD be workable by using
switches like 

        --buildcxx=g++ 

you'd be able to use your mingw g++ directly, even with

        --build=cygwin

OR with

        --build=win32

but I'd try to avoid that because I don't know if it will
actually work  :) 


-- John Skaller <skaller at users dot sf dot net> Felix, successor to 
C++: http://felix.sf.net



-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
Felix-language mailing list
Felix-language@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/felix-language

Reply via email to