------- Comment #5 from paolo dot carlini at oracle dot com 2010-04-15 03:09 ------- (In reply to comment #4) > MinGW doesn't have sys/ioctl.h; could we test HAVE_SYS_IOCTL_H? Are there > likely to be any platforms we support that define FIONREAD but don't have > sys/ioctl.h?
Dave, if you look at the beginning of the file, it looks like Solaris 2.5 is one such case. Now, leaving aside for the moment that likely clean-ups are possible (because Rainer is removing support for some old Solaris* versions, if I remember correctly), I wonder whether we should simply special case mingw32 and conditional to the macro being defined (don't remember: _MINGW32?) just include the required headers, and use ioctlsocket in place of ioctl. I think that in such kind of low level .cc file it's all the elegance we can hope for ;) -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43738
