Path: utzoo!utgpu!watserv1!watmath!att!att!linac!pacific.mps.ohio-state.edu!zaphod.mps.ohio-state.edu!swrinde!ucsd!ucbvax!bloom-beacon!LIGHTNING.MCRCIM.MCGILL.EDU!mouse
From: mouse@LIGHTNING.MCRCIM.MCGILL.EDU
Newsgroups: comp.windows.x
Subject: Re:  X11R3 and Gcc 1.37.1
Message-ID: <9011042259.AA00244@lightning.McRCIM.McGill.EDU>
Date: 4 Nov 90 22:59:36 GMT
Sender: daemon@athena.mit.edu (Mr Background)
Organization: The Internet
Lines: 48

> I have just [overnight] recompiled X11R3 with the Gnu C compiler,
> gcc-1.37.1.  However I now have a bad X11.  Can anyone throw any
> light on the problem?

> "xinit" results in the messages:
>         Getting interface configuration: Operation not supported on socket
>         sunOpenFrameBuffer: Inappropriate ioctl for device
>         sunOpenFrameBuffer: Inappropriate ioctl for device

>         Fatal server bug!
>         Can't find any displays

>         giving up.
>         xinit:  Connection refused (errno 61):  unable to connect to X server

I think this is a FAQ (but I suppose that doesn't help someone without
news access...).  The problem lies with gcc's include files.  You need
to run (or rerun, or perhaps fix and rerun :-) gcc's fixincludes
script.  Try a simple test.  Put the following two lines in a file, say
foo.c:

	#include <sys/ioctl.h>
	SIOCGIFCONF

Run `cc -E foo.c' and `gcc -E foo.c', and compare the last line of
output from them.  You will[%] see that the gcc output has 'x' where
the cc output has 'i'.  They should be (functionally) identical.  Play
with fixincludes etc until this works (if this doesn't work, there's
little hope for X).

[%] Almost certainly.  Every time I've heard of this, the problem has
    been as I've described here.

> Running my old [Sun cc compiled] X11R3, I still can't run any of my
> new clients. For example, I get:

>         gav% xlogo
>         XIO:  fatal IO error 45 (Operation not supported on socket) on X server "gav:0.0"
>               after 30 requests (17 known processed) with 0 events remaining.

This is more of the same thing.  The problem is that when gcc uses your
Sun-provided include files, all the ioctl defines (like SIOCGIFCONF)
come out with the wrong values, so all ioctls fail.

					der Mouse

			old: mcgill-vision!mouse
			new: mouse@larry.mcrcim.mcgill.edu