Maybe when the configure script is run, it could #define something and then in the source file that uses the load_symbols() function it could call the function using #ifdef. The configure script could check for existance of the needed header. I guess it is still a plug for now, but makes it R5 compatible. -----Original Message----- From: Axel Dörfler" [mailto:axeld@xxxxxxxxxxxxxxxx] Sent: Thursday, May 02, 2002 12:52 PM To: openbeos@xxxxxxxxxxxxx Subject: [openbeos] Re: Building OpenBFS... > What??? > > That header should be replaced with a correct one and put into our > sources. And put into the next installation preferably. This is much > better than having people edit header files themselves (and > potentially > screw them up). They can just comment out the line in question - there is no need to screw up any headers files. Since it's a problem of *some* of the Be headers, I don't want to add the whole header file - other kernel drivers will have the same problem, and it would be a bit dumb to have this header in every driver folder. The simple solution would be to just add the prototype to the kernel_interface.cpp file, that way it should always compile without any problems. Adios... Axel.