Hi, Samuel Thibault, le Sun 16 Dec 2007 15:13:22 +0000, a écrit : > Theodore Tso, le Sun 16 Dec 2007 10:03:12 -0500, a écrit : > > On Sun, Dec 16, 2007 at 02:47:35AM +0000, Samuel Thibault wrote: > > > However, we still have a problem on the resize part. On GNU/Hurd (and > > > also on some BSD variants IIRC), the _IOW macro encodes the type in the > > > actual ioctl number, so that it can't be completely arbitrary. The > > > attached patch defines what is needed for struct ext2_new_group_input to > > > work. > > > > What platforms has this been tested on. Just the Hurd, or have you > > also tested on FreeBSD? > > Just the Hurd, I don't have a FreeBSD box. > > > Or I could just #ifdef it for the Hurd for now, and change it to a #if > > defined(__hurd__) || defined(__fooBSD__) later.... > > GNU Hurd doesn't define __hurd__, it defines __GNU__ (which is not > defined on other GNU systems like GNU/Linux or GNU/*BSD).
Some BSD people told me that their macro was #define _IOW(g,n,t) _IOC(IOC_IN, (g), (n), sizeof(t)) so that the type can be arbitrary with a need for the hurdish _IOT definition. Samuel -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
