[patch/rfa] xm-aix4.h cleanup

[patch/rfa] xm-aix4.h cleanup

[Joel Brobecker]

[-- Attachment #1: Type: text/plain, Size: 1133 bytes --]

Peter, Kevin,

Based on a previous discussion, and Peter's testing, do you think I can
commit the attached patch? I am currently testing it on our 2 AIX
machines (4.3.2 and 5.1). I would like to commit it once I have the
results, provided they are positive (I expect them to be, given Peter's
reports of his own testing).

2004-08-09  Joel Brobecker  <brobecker@gnat.com>

        General cleanup of the AIX xm include file:
        * config/xm-aix4.h (fd_set): Remove undefine.
        Remove include of <sys/select.h>.
        (HAVE_TERMIO): Remove define.
        (USG): Likewise.
        (NULL): Remove redefine.
        (vfork): Remove define.
        (termdef): Remove.
        (SIGWINCH_HANDLER): Remove.
        (SIGWINCH_HANDLER_BODY): Remove.

This leaves:

  . FIVE_ARG_PTRACE: Will become OBE thanks to a patch from MarkK.

  . USE_O_NOCTTY: I think the idea for that one is to always use O_NOCTTY,
    regardless of the system. So the macro becomes moot. Will send a patch.

  . #pragma alloca: Need to continue the discussion.

All in all, I think we can get rid of that file within the next 2 weeks.

Thanks,
-- 
Joel

[-- Attachment #2: xm-aix.diff --]
[-- Type: text/plain, Size: 2330 bytes --]

Index: xm-aix4.h
===================================================================
RCS file: /cvs/src/src/gdb/config/xm-aix4.h,v
retrieving revision 1.6
diff -u -p -r1.6 xm-aix4.h
--- xm-aix4.h	2 Dec 2001 02:57:13 -0000	1.6
+++ xm-aix4.h	9 Aug 2004 06:26:54 -0000
@@ -20,27 +20,6 @@
    Foundation, Inc., 59 Temple Place - Suite 330,
    Boston, MA 02111-1307, USA.  */
 
-/* The following text is taken from config/rs6000.mh:
- * # The IBM version of /usr/include/rpc/rpc.h has a bug -- it says
- * # `extern fd_set svc_fdset;' without ever defining the type fd_set.
- * # Unfortunately this occurs in the vx-share code, which is not configured
- * # like the rest of GDB (e.g. it doesn't include "defs.h").
- * # We circumvent this bug by #define-ing fd_set here, but undefining it in
- * # the xm-rs6000.h file before ordinary modules try to use it.  FIXME, IBM!
- * MH_CFLAGS='-Dfd_set=int'
- * So, here we do the undefine...which has to occur before we include
- * <sys/select.h> below.
- */
-#undef fd_set
-
-#include <sys/select.h>
-
-/* At least as of AIX 3.2, we have termios.  */
-#define	HAVE_TERMIOS 1
-/* #define HAVE_TERMIO 1 */
-
-#define	USG 1
-
 #define FIVE_ARG_PTRACE
 
 /* This system requires that we open a terminal with O_NOCTTY for it to
@@ -48,44 +27,6 @@
 
 #define	USE_O_NOCTTY
 
-/* Brain death inherited from PC's pervades.  */
-#undef NULL
-#define NULL 0
-
 /* The IBM compiler requires this in order to properly compile alloca().  */
 #pragma alloca
 
-/* There is no vfork.  */
-
-#define	vfork	fork
-
-char *termdef ();
-
-/* Signal handler for SIGWINCH `window size changed'. */
-
-#define	SIGWINCH_HANDLER  aix_resizewindow
-extern void aix_resizewindow (int);
-
-/* `lines_per_page' and `chars_per_line' are local to utils.c. Rectify this. */
-
-#define	SIGWINCH_HANDLER_BODY	\
-									\
-/* Respond to SIGWINCH `window size changed' signal, and reset GDB's	\
-   window settings appropriately. */					\
-									\
-void 						\
-aix_resizewindow (signo)			\
-     int signo;					\
-{						\
-  int fd = fileno (stdout);			\
-  if (isatty (fd)) {				\
-    int val;					\
-						\
-    val = atoi (termdef (fd, 'l'));		\
-    if (val > 0)				\
-      lines_per_page = val;			\
-    val = atoi (termdef (fd, 'c'));		\
-    if (val > 0)				\
-      chars_per_line = val;			\
-  }						\
-}

Re: [patch/rfa] xm-aix4.h cleanup

[Joel Brobecker]

[-- Attachment #1: Type: text/plain, Size: 697 bytes --]

> This leaves:
> 
>   . FIVE_ARG_PTRACE: Will become OBE thanks to a patch from MarkK.
> 
>   . USE_O_NOCTTY: I think the idea for that one is to always use O_NOCTTY,
>     regardless of the system. So the macro becomes moot. Will send a patch.
> 
>   . #pragma alloca: Need to continue the discussion.

Actually, I didn't look closely enough, even after Mark's answer.
This is already dealt with, so we can remove this one as well.
So I'm suggesting this patch instead (which such scraps the alloca
#pragma in addition to what the previous patch was already removing).

Is this OK?

(the only little detail is that I am not sure I can verify this works
or not. I'll see what I can do).

-- 
Joel

[-- Attachment #2: xm-aix.diff --]
[-- Type: text/plain, Size: 2330 bytes --]

Index: xm-aix4.h
===================================================================
RCS file: /cvs/src/src/gdb/config/xm-aix4.h,v
retrieving revision 1.6
diff -u -p -r1.6 xm-aix4.h
--- xm-aix4.h	2 Dec 2001 02:57:13 -0000	1.6
+++ xm-aix4.h	9 Aug 2004 06:46:44 -0000
@@ -20,27 +20,6 @@
    Foundation, Inc., 59 Temple Place - Suite 330,
    Boston, MA 02111-1307, USA.  */
 
-/* The following text is taken from config/rs6000.mh:
- * # The IBM version of /usr/include/rpc/rpc.h has a bug -- it says
- * # `extern fd_set svc_fdset;' without ever defining the type fd_set.
- * # Unfortunately this occurs in the vx-share code, which is not configured
- * # like the rest of GDB (e.g. it doesn't include "defs.h").
- * # We circumvent this bug by #define-ing fd_set here, but undefining it in
- * # the xm-rs6000.h file before ordinary modules try to use it.  FIXME, IBM!
- * MH_CFLAGS='-Dfd_set=int'
- * So, here we do the undefine...which has to occur before we include
- * <sys/select.h> below.
- */
-#undef fd_set
-
-#include <sys/select.h>
-
-/* At least as of AIX 3.2, we have termios.  */
-#define	HAVE_TERMIOS 1
-/* #define HAVE_TERMIO 1 */
-
-#define	USG 1
-
 #define FIVE_ARG_PTRACE
 
 /* This system requires that we open a terminal with O_NOCTTY for it to
@@ -48,44 +27,3 @@
 
 #define	USE_O_NOCTTY
 
-/* Brain death inherited from PC's pervades.  */
-#undef NULL
-#define NULL 0
-
-/* The IBM compiler requires this in order to properly compile alloca().  */
-#pragma alloca
-
-/* There is no vfork.  */
-
-#define	vfork	fork
-
-char *termdef ();
-
-/* Signal handler for SIGWINCH `window size changed'. */
-
-#define	SIGWINCH_HANDLER  aix_resizewindow
-extern void aix_resizewindow (int);
-
-/* `lines_per_page' and `chars_per_line' are local to utils.c. Rectify this. */
-
-#define	SIGWINCH_HANDLER_BODY	\
-									\
-/* Respond to SIGWINCH `window size changed' signal, and reset GDB's	\
-   window settings appropriately. */					\
-									\
-void 						\
-aix_resizewindow (signo)			\
-     int signo;					\
-{						\
-  int fd = fileno (stdout);			\
-  if (isatty (fd)) {				\
-    int val;					\
-						\
-    val = atoi (termdef (fd, 'l'));		\
-    if (val > 0)				\
-      lines_per_page = val;			\
-    val = atoi (termdef (fd, 'c'));		\
-    if (val > 0)				\
-      chars_per_line = val;			\
-  }						\
-}

Re: [patch/rfa] xm-aix4.h cleanup

[Kevin Buettner]

On Sun, 8 Aug 2004 23:48:26 -0700
Joel Brobecker <brobecker@gnat.com> wrote:

> > This leaves:
> > 
> >   . FIVE_ARG_PTRACE: Will become OBE thanks to a patch from MarkK.
> > 
> >   . USE_O_NOCTTY: I think the idea for that one is to always use O_NOCTTY,
> >     regardless of the system. So the macro becomes moot. Will send a patch.
> > 
> >   . #pragma alloca: Need to continue the discussion.
> 
> Actually, I didn't look closely enough, even after Mark's answer.
> This is already dealt with, so we can remove this one as well.
> So I'm suggesting this patch instead (which such scraps the alloca
> #pragma in addition to what the previous patch was already removing).
> 
> Is this OK?
> 
> (the only little detail is that I am not sure I can verify this works
> or not. I'll see what I can do).

I'm in favor of this patch so long as AIX still builds okay and the test
results look good.

Kevin

Re: [patch/rfa] xm-aix4.h cleanup

[Joel Brobecker]

> I'm in favor of this patch so long as AIX still builds okay and the test
> results look good.

Thanks Kevin. This patch was successfullly tested on AIX 4.3.2 and 5.1.
So I checked it in.

Now, on to the last 2 entries....

-- 
Joel