[patch] Fix --with-system-readline vs. readline-6.0+

[patch] Fix --with-system-readline vs. readline-6.0+

[Jan Kratochvil]

Hi,

configure --with-system-readline with system readline-6.0 do cause:
.../gdb/tui/tui-io.c:567: undefined reference to `readline_echoing_p'
.../gdb/tui/tui-io.c:525: undefined reference to `readline_echoing_p'
.../gdb/tui/tui-io.c:530: undefined reference to `readline_echoing_p'
	http://koji.fedoraproject.org/koji/getfile?taskID=1515524&name=build.log

There are now various vendor patches out there
	http://www.mail-archive.com/pld-cvs-commit@lists.pld-linux.org/msg174978.html
	http://lists.opensuse.org/opensuse-commit/2009-03/msg00362.html
	http://osdir.com/ml/fedora-extras-commits/2009-07/msg05039.html

After bundled readline gets upgraded to 6.0 this AC_DEFINE should be reversed
and tui/tui-io.c to be updated to use _rl_echoing_p of the bundled version.


Thanks,
Jan


gdb/
2009-07-31  Jan Kratochvil  <jan.kratochvil@redhat.com>

	Fix compatibility of --with-system-readline and readline-6.0+.
	* configure.ac <--with-system-readline> (for readline_echoing_p): New
	test.
	* config.in: Regenerate.
	* configure: Regenerate.

--- a/gdb/configure.ac
+++ b/gdb/configure.ac
@@ -536,6 +536,21 @@ if test "$with_system_readline" = yes; then
   READLINE=-lreadline
   READLINE_DEPS=
   READLINE_CFLAGS=
+
+  # readline-6.0 started to use the name `_rl_echoing_p'.
+  # `$(READLINE_DIR)/' of bundled readline would not resolve in configure.
+
+  AC_MSG_CHECKING([for readline_echoing_p])
+  save_LIBS=$LIBS
+  LIBS="$LIBS $READLINE"
+  AC_LINK_IFELSE(AC_LANG_PROGRAM(,[[extern int readline_echoing_p;
+				    return readline_echoing_p;]]),
+		 [READLINE_ECHOING_P=yes],
+		 [READLINE_ECHOING_P=no
+		  AC_DEFINE([readline_echoing_p], [_rl_echoing_p],
+			    [readline-6.0 started to use different name.])])
+  LIBS="$save_LIBS"
+  AC_MSG_RESULT([$READLINE_ECHOING_P])
 else
   READLINE='$(READLINE_DIR)/libreadline.a'
   READLINE_DEPS='$(READLINE)'

Re: [patch] Fix --with-system-readline vs. readline-6.0+

[Andreas Schwab]

Jan Kratochvil <jan.kratochvil@redhat.com> writes:

> After bundled readline gets upgraded to 6.0 this AC_DEFINE should be reversed
> and tui/tui-io.c to be updated to use _rl_echoing_p of the bundled version.

Actually it should not use this internal variable at all.

Andreas.

-- 
Andreas Schwab, schwab@redhat.com
GPG Key fingerprint = D4E8 DBE3 3813 BB5D FA84  5EC7 45C6 250E 6F00 984E
"And now for something completely different."

Re: [patch] Fix --with-system-readline vs. readline-6.0+

[Tom Tromey]

>>>>> "Andreas" == Andreas Schwab <schwab@redhat.com> writes:

Jan> After bundled readline gets upgraded to 6.0 this AC_DEFINE should
Jan> be reversed and tui/tui-io.c to be updated to use _rl_echoing_p of
Jan> the bundled version.

Andreas> Actually it should not use this internal variable at all.

Is there some nice way to do that?  Like, some exported readline API?

I'm inclined to approve this patch, on the theory that it will keep the
TUI working a while longer.  However, if there is a better possibility
we can use now, I won't.

Tom

Re: [patch] Fix --with-system-readline vs. readline-6.0+

[Andreas Schwab]

Tom Tromey <tromey@redhat.com> writes:

>>>>>> "Andreas" == Andreas Schwab <schwab@redhat.com> writes:
>
> Jan> After bundled readline gets upgraded to 6.0 this AC_DEFINE should
> Jan> be reversed and tui/tui-io.c to be updated to use _rl_echoing_p of
> Jan> the bundled version.
>
> Andreas> Actually it should not use this internal variable at all.
>
> Is there some nice way to do that?  Like, some exported readline API?

Not that I know of, but Chet will probably add one if asked.

Andreas.

-- 
Andreas Schwab, schwab@redhat.com
GPG Key fingerprint = D4E8 DBE3 3813 BB5D FA84  5EC7 45C6 250E 6F00 984E
"And now for something completely different."