public inbox for gcc@gcc.gnu.org
 help / color / mirror / Atom feed
* small patch for sparc-linux-gnulibc1
@ 1997-10-17 19:44 David S. Miller
  1997-10-17 23:35 ` Jeffrey A Law
  1997-10-20  2:07 ` Andreas Schwab
  0 siblings, 2 replies; 5+ messages in thread
From: David S. Miller @ 1997-10-17 19:44 UTC (permalink / raw)
  To: egcs

I can't get a clean build without the following change.

I'm not certain this is the correct fix though.  gbl-ctors.h has this
bit in it:

#if defined (WINNT) || defined (NEED_ATEXIT)
extern int atexit (void (*) (void));
#else
extern void atexit (void (*) (void));
#endif

It seems to me if NEED_ATEXIT is not defined, this means a prototype
extern is not necessary.  Perhaps someone can help figure out what
really should be happening here.

Fri Oct 17 17:09:05 1997  David S. Miller  <davem@tanya.rutgers.edu>

	* config/sparc/linux.h (NEED_ATEXIT): Define.

--- config/sparc/linux.h.~1~	Wed Sep 24 20:33:46 1997
+++ config/sparc/linux.h	Fri Oct 17 17:07:43 1997
@@ -27,6 +27,9 @@
 #undef HAVE_ATEXIT
 #define HAVE_ATEXIT
 
+#undef NEED_ATEXIT
+#define NEED_ATEXIT
+
 /* Linux uses ctype from glibc.a. I am not sure how complete it is.
    For now, we play safe. It may change later. */
 

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: small patch for sparc-linux-gnulibc1
  1997-10-17 19:44 small patch for sparc-linux-gnulibc1 David S. Miller
@ 1997-10-17 23:35 ` Jeffrey A Law
  1997-10-18 12:39   ` David S. Miller
  1997-10-20  2:07 ` Andreas Schwab
  1 sibling, 1 reply; 5+ messages in thread
From: Jeffrey A Law @ 1997-10-17 23:35 UTC (permalink / raw)
  To: David S. Miller; +Cc: egcs

  In message < 199710172212.SAA04011@jenolan.rutgers.edu >you write:
  > 
  > I can't get a clean build without the following change.
  > 
  > I'm not certain this is the correct fix though.  gbl-ctors.h has this
  > bit in it:
  > 
  > #if defined (WINNT) || defined (NEED_ATEXIT)
  > extern int atexit (void (*) (void));
  > #else
  > extern void atexit (void (*) (void));
  > #endif
  > 
  > It seems to me if NEED_ATEXIT is not defined, this means a prototype
  > extern is not necessary.  Perhaps someone can help figure out what
  > really should be happening here.
  > 
  > Fri Oct 17 17:09:05 1997  David S. Miller  <davem@tanya.rutgers.edu>
  > 
  > 	* config/sparc/linux.h (NEED_ATEXIT): Define.
Hmmm.  At first I was going to say use autoconf for this, but it
seems to me that this is really a target issue, so autoconf doesn't
help.

Similarly for my comment about the gbl-ctors.h change in my last message.

So I'm not sure what to do.

Jeff

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: small patch for sparc-linux-gnulibc1
  1997-10-17 23:35 ` Jeffrey A Law
@ 1997-10-18 12:39   ` David S. Miller
  0 siblings, 0 replies; 5+ messages in thread
From: David S. Miller @ 1997-10-18 12:39 UTC (permalink / raw)
  To: law; +Cc: egcs

   Date: Sat, 18 Oct 1997 00:21:34 -0600
   From: Jeffrey A Law <law@cygnus.com>

   Hmmm.  At first I was going to say use autoconf for this, but it
   seems to me that this is really a target issue, so autoconf doesn't
   help.

   Similarly for my comment about the gbl-ctors.h change in my last
   message.

   So I'm not sure what to do.

I think the second extern should be blown away.  If the defines say "I
HAVE_ATEXIT" and "I do not NEED_ATEXIT" the extern should not be
there.  Else the meaning of these macros should be clarified.

Later,
David "Sparc" Miller
davem@caip.rutgers.edu

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: small patch for sparc-linux-gnulibc1
  1997-10-17 19:44 small patch for sparc-linux-gnulibc1 David S. Miller
  1997-10-17 23:35 ` Jeffrey A Law
@ 1997-10-20  2:07 ` Andreas Schwab
  1997-10-20 14:00   ` David S. Miller
  1 sibling, 1 reply; 5+ messages in thread
From: Andreas Schwab @ 1997-10-20  2:07 UTC (permalink / raw)
  To: David S. Miller; +Cc: egcs

David S Miller <davem@jenolan.rutgers.edu> writes:

|> I can't get a clean build without the following change.

|> I'm not certain this is the correct fix though.  gbl-ctors.h has this
|> bit in it:

|> #if defined (WINNT) || defined (NEED_ATEXIT)
|> extern int atexit (void (*) (void));
|> #else
|> extern void atexit (void (*) (void));
|> #endif

|> It seems to me if NEED_ATEXIT is not defined, this means a prototype
|> extern is not necessary.  Perhaps someone can help figure out what
|> really should be happening here.

Note that the second prototype is broken anyway.  atexit has always
returned int, not void.

-- 
Andreas Schwab                                      "And now for something
schwab@issan.informatik.uni-dortmund.de              completely different"
schwab@gnu.org

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: small patch for sparc-linux-gnulibc1
  1997-10-20  2:07 ` Andreas Schwab
@ 1997-10-20 14:00   ` David S. Miller
  0 siblings, 0 replies; 5+ messages in thread
From: David S. Miller @ 1997-10-20 14:00 UTC (permalink / raw)
  To: schwab; +Cc: egcs

   From: Andreas Schwab <schwab@issan.informatik.uni-dortmund.de>
   Date: 20 Oct 1997 11:07:10 +0200

   |> #if defined (WINNT) || defined (NEED_ATEXIT)
   |> extern int atexit (void (*) (void));
   |> #else
   |> extern void atexit (void (*) (void));
   |> #endif

   Note that the second prototype is broken anyway.  atexit has always
   returned int, not void.

Broken or not, all the libc5 based systems I do work on have that
prototype in both the headers and the man pages.  So we have to handle
it.

Later,
David S. Miller
davem@caip.rutgers.edu

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~1997-10-20 14:00 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
1997-10-17 19:44 small patch for sparc-linux-gnulibc1 David S. Miller
1997-10-17 23:35 ` Jeffrey A Law
1997-10-18 12:39   ` David S. Miller
1997-10-20  2:07 ` Andreas Schwab
1997-10-20 14:00   ` David S. Miller

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).