This is the mail archive of the cygwin@sources.redhat.com mailing list for the Cygwin project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: libodbccp32.a


Chris Faylor wrote:
> >
> >Exactly right! Sometimes it takes someone else to state the obvious
> >before I see the problem.  Thanks Danny, Thanks Chris!
> 
> I think you mean Chuck, not Chris.

Whatever. 'sokay.

> 
> >The only thing is that this build was looking picked up a reference to
> >win32.h from perl.h. win32.h is not defined on a stock cygwin configuration,
> >I actually edited perl.h but I guess a symlink to windows.h would be just as
> >good.

No. perl.h has the following stanza:

#ifdef WIN32
#  include "win32.h"
#endif

That's "win32.h", not <win32.h>.  There's a file in
<perl-src>/win32/win32.h, which is probably copied up to the toplevel
during a win32-native configure. (I'm just guessing here)

The correct fix is probably to replace the #ifdef statement with:

#if defined(WIN32) && !defined(__CYGWIN__)

But the presence of that #include "win32.h" has never cause me any build
trouble; it's odd that it would only cause difficulties when building
DBD::ODBC but not other modules.

> >
> >Should that be standard in cygwin? Should this be an issue for the perl
> >porters? Or is this just another README!
> 
> I don't know why perl is expecting win32.h but this is not a standard windows
> header, AFAIK.  I don't think that cygwin has ever had a win32.h, either.

--Chuck

--
Want to unsubscribe from this list?
Send a message to cygwin-unsubscribe@sourceware.cygnus.com


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]