[elephant-devel] Testing

[Ian Eslick]

Ok, I got myself a windows installation and set it up only to get to  
the same point you're at.  I have no idea what's going on with db.h  
either.  I'll try looking at this again tomorrow.

You know, I switched to Mac OS because I hated working with cygwin  
under MS Windows... :)

On Feb 19, 2007, at 6:20 PM, Frank Schorr wrote:

> I haven't really understood what happens when a dll is built with  
> cygwin.
> dlltool generates the .def file and the exports.o used in the
> next gcc step. See
> http://www.gnu.org/software/binutils/manual/html_chapter/ 
> binutils_13.html
> http://www.cygwin.com/cygwin-ug-net/dll.html
>
>
> I was able to get a libmemutil.dll without errors by:
>
> gcc -mno-cygwin -mwindows -std=c99 -c libmemutil.c
> dlltool -z libmeutil.def --export-all-symbols -e exports.o -l  
> libmemutil.lib libmemutil.o
> gcc -shared -mno-cygwin -mwindows libmemutil.o exports.o -o  
> libmemutil.dll
>
>
> A while ago, when I made the dlls for elephant 0.6.0, my first try  
> was MSVC, guided by
> http://bc.tech.coop/blog/041007.html
> However, the available free MSVC appeared to have no preset dll- 
> project any more.
> I had to try compiler switches ... and found cygwin easier.
> This should mean that not every lisp-on-windows-user has
> MSVC installed (and is able to use it. cygwin is easy to install).
>
> Best regards,
> Frank
>
>> -----Ursprüngliche Nachricht-----
>> Von: Elephant bugs and development <elephant-devel at common-lisp.net>
>> Gesendet: 19.02.07 21:26:09
>> An: Elephant bugs and development <elephant-devel at common-lisp.net>
>> Betreff: Re: [elephant-devel] Testing
>
>
>> Thanks for trying that.  Do you know if there's a way with the cygwin
>> gcc compiler to use a .def file instead of generating a .o with the
>> exports?  I'm pretty sure MSVC has one.  We'll also need a set of
>> build commands for MSVC...
>>
>> I suspect we may have broken the build of libmemutil when we included
>> sys/types to get access to u_int64 and u_int32, etc.   Without a
>> local system to play with I'm not sure I can debug this.  I guess the
>> thing to do is look at db.h and sys/types to try to figure out what
>> was intended and whether they're compatible.
>>
>> I'll see about getting access to a Windows box to try to reproduce
>> and test these things on my side.
>>
>> Thanks,
>> Ian
>>
>> On Feb 19, 2007, at 2:09 PM, Frank Schorr wrote:
>>
>>>
>>> gcc -shared -mno-cygwin -mwindows -std=c99 --export-all-symbols
>>> libberkeley-db.def libberkeley-db.c -o libberkeley-db.dll
>>>
>>> produces
>>>
>>> cc1: error: unrecognized command line option "-fexport-all-symbols"
>>>
>>> Further:
>>>
>>> gcc -mno-cygwin -mwindows -c -Wall -std=c99 -L/c/Programme/Oracle/
>>> Berkeley\ DB\ 4.5.20/lib/ -I/c/Programme/Oracle/Berkeley\ DB\
>>> 4.5.20/include/ libberkeley-db.c
>>>
>>> produces
>>>
>>> In file included from libberkeley-db.c:165:
>>> /c/Programme/Oracle/Berkeley DB 4.5.20/include/db.h:99: error:
>>> conflicting types
>>>  for 'ssize_t'
>>> /usr/lib/gcc/i686-pc-mingw32/3.4.4/../../../../i686-pc-mingw32/
>>> include/sys/types
>>> .h:104: error: previous declaration of 'ssize_t' was here
>>> libberkeley-db.c: In function `case_cmp':
>>> libberkeley-db.c:1233: warning: implicit declaration of function
>>> `_strnicmp'
>>>
>>> Hope this can help
>>> Please let me know what to do next.
>>> Frank
>>>
> ______________________________________________________________________ 
> ____
> Erweitern Sie FreeMail zu einem noch leistungsstärkeren E-Mail- 
> Postfach!		
> Mehr Infos unter http://freemail.web.de/home/landingpad/?mc=021131
>
> _______________________________________________
> elephant-devel site list
> elephant-devel at common-lisp.net
> http://common-lisp.net/mailman/listinfo/elephant-devel