[Clam-devel] Re: Windows native compilation

David García Garzón dgarcia at iua.upf.edu
Tue Apr 29 00:09:41 PDT 2008


I should have "used the code, luke". Id3 just uses this macro in visual, in 
any other compiler it just ignores it. So I definitely removed the 
LINKOPTIONS lines.

The bad news is that this is not the problem you have. So, my second guess has 
adquired relevance: we are compiling as it were C++ (with g++) and linking as 
it were C (with gcc). The puzzling fact is that crosscompilation works 
without problems. So we need to know which is the test is failing. The pasted 
config.log, is missing the lines that say that. That information is given by 
the console output (checking id3 header/compiling/linking/runnig, id3 
header/compiling/linking/run failed) or maybe some lines above in the 
config.log. Could you paste the full config.log or attaching it to the list?

As you can see adding support to a 'new' platform is somewhat tedious and 
sloow. This is why i wanted it solved as soon as we can and then enjoy 
coding.


On Dimarts 29 Abril 2008, David García Garzón wrote:
> (I move the IRC conversation to the list)
>
> Jun, the error you talk about on the irc
> (http://clam.pastebin.com/d41b5600c) is weird as it complaints about
> symbols of the standard c++ library just as it were linking using gcc
> instead of g++ (indeed that could be the problem).
>
> But, checking the code i saw an irregularity on the libid3 checks that
> could be also the reason:
>     if sys.platform == 'win32' :
>         if crosscompiling :
>             env.Append( CPPFLAGS=['-DID3LIB_LINKOPTION=3'] )
>         else :
>             env.Append( CPPFLAGS=['-DID3LIB_LINKOPTION=1'] )
>
> Notice that the condition on crosscompiling is never reached when
> crosscompiling as when crosscompiling, sys.platform == "linux". Thus, the
> croscompiled version is not defining the macro, neither with 3 nor 1, and
> it works. Being that a LINKOPTION it could be the cause.
>
> So I commented out that code and commited. The crosscompiled version still
> worked as testfarm shows (http://clam.iua.upf.edu/testfarm/). Please, Jun,
> check the native compilation. If you still have problems send me an email.
> It might be that the check code do link with gcc instead of g++ but this
> should not be a problem as the crosscompiled is working.
>
> If the next error is about not finding winnm library or some that could be
> related when testing portmidi, then consider commenting out the two related
> lines in CLAM/scons/libs/clam_dependent_libs_checks.py (arround 170) and
> later there is a native only check on directx (arround 221). Proceed in a
> similar way, that is, commenting out the win32 specific code.
>
> In mid-week I check the irc log at home at night and maybe at office but
> just when i am working at place which lately is not that true. Tomorrow i
> will be busy with classes until Spain 15:00 (21:00 at your place). But send
> your emails to the list, maybe someone else can answer you.



-- 
David García Garzón
(Work) dgarcia at iua dot upf anotherdot es
http://www.iua.upf.edu/~dgarcia
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.clam-project.org/pipermail/clam-devel-clam-project.org/attachments/20080429/ed355ba5/attachment.sig>


More information about the clam-devel mailing list