RFA: fix locale configure checks

RFA: fix locale configure checks

[Tom Tromey]

Today while exploring the charset code I happened to notice that gdb
never calls setlocale.

The code is there, but configure never does the checks that might
define HAVE_LOCALE_H, HAVE_SETLOCALE, or HAVE_LC_MESSAGES.

This patch fixes the problem.  I just copied what GCC does here.

Built & regtested on x86-64.  I also examined the config.h on my x86
box by hand.

Ok?

Tom

:ADDPATCH build:

2008-08-21  Tom Tromey  <tromey@redhat.com>

	* config.in, configure: Rebuild.
	* configure.ac: Check for locale.h, setlocale.  Call
	AM_LC_MESSAGES.
	* acinclude.m4: Include lcmessage.m4.

diff --git a/gdb/acinclude.m4 b/gdb/acinclude.m4
index 4e8f590..2a951e2 100644
--- a/gdb/acinclude.m4
+++ b/gdb/acinclude.m4
@@ -26,6 +26,9 @@ sinclude(../config/tcl.m4)
 dnl For dependency tracking macros.
 sinclude([../config/depstand.m4])
 
+dnl For AM_LC_MESSAGES
+sinclude([../config/lcmessage.m4])
+
 #
 # Sometimes the native compiler is a bogus stub for gcc or /usr/ucb/cc. This
 # makes configure think it's cross compiling. If --target wasn't used, then
diff --git a/gdb/configure.ac b/gdb/configure.ac
index 680fba0..2d2c955 100644
--- a/gdb/configure.ac
+++ b/gdb/configure.ac
@@ -647,6 +647,7 @@ AC_CHECK_HEADERS(link.h, [], [],
 # include <nlist.h>
 #endif
 ])
+AC_CHECK_HEADERS(locale.h)
 AC_CHECK_HEADERS(machine/reg.h)
 AC_CHECK_HEADERS(poll.h sys/poll.h)
 AC_CHECK_HEADERS(proc_service.h thread_db.h gnu/libc-version.h)
@@ -713,6 +714,7 @@ AC_CHECK_HEADERS(ctype.h time.h)
 AC_CHECK_DECLS([free, malloc, realloc])
 AC_CHECK_DECLS([strerror, strstr])
 AC_CHECK_DECLS([getopt, snprintf, vsnprintf])
+AM_LC_MESSAGES
 
 # ----------------------- #
 # Checks for structures.  #
@@ -753,6 +755,7 @@ AC_CHECK_FUNCS(poll)
 AC_CHECK_FUNCS(pread64)
 AC_CHECK_FUNCS(sbrk)
 AC_CHECK_FUNCS(setpgid setpgrp setsid)
+AC_CHECK_FUNCS(setlocale)
 AC_CHECK_FUNCS(sigaction sigprocmask sigsetmask)
 AC_CHECK_FUNCS(socketpair)
 AC_CHECK_FUNCS(syscall)

Re: RFA: fix locale configure checks

[Tom Tromey]

>>>>> "Tom" == Tom Tromey <tromey@redhat.com> writes:

Tom> Today while exploring the charset code I happened to notice that gdb
Tom> never calls setlocale.

Tom> The code is there, but configure never does the checks that might
Tom> define HAVE_LOCALE_H, HAVE_SETLOCALE, or HAVE_LC_MESSAGES.

Tom> This patch fixes the problem.  I just copied what GCC does here.

I've updated this patch to account for recent changes to configure.ac.

I am re-regtesting it, but I don't anticipate any problems.
Ok if it passes?

Tom

2008-11-25  Tom Tromey  <tromey@redhat.com>

	* config.in, configure: Rebuild.
	* configure.ac: Check for locale.h, setlocale.  Call
	AM_LC_MESSAGES.
	* acinclude.m4: Include lcmessage.m4.

diff --git a/gdb/acinclude.m4 b/gdb/acinclude.m4
index 4e8f590..2a951e2 100644
--- a/gdb/acinclude.m4
+++ b/gdb/acinclude.m4
@@ -26,6 +26,9 @@ sinclude(../config/tcl.m4)
 dnl For dependency tracking macros.
 sinclude([../config/depstand.m4])
 
+dnl For AM_LC_MESSAGES
+sinclude([../config/lcmessage.m4])
+
 #
 # Sometimes the native compiler is a bogus stub for gcc or /usr/ucb/cc. This
 # makes configure think it's cross compiling. If --target wasn't used, then
diff --git a/gdb/configure.ac b/gdb/configure.ac
index 8725aa6..4e0cf7d 100644
--- a/gdb/configure.ac
+++ b/gdb/configure.ac
@@ -696,7 +696,7 @@ AC_CHECK_HEADERS([nlist.h machine/reg.h poll.h sys/poll.h proc_service.h \
 		  sys/resource.h sys/procfs.h sys/ptrace.h ptrace.h \
 		  sys/reg.h sys/debugreg.h sys/select.h sys/syscall.h \
 		  sys/types.h sys/wait.h wait.h termios.h termio.h \
-		  sgtty.h unistd.h elf_hp.h ctype.h time.h])
+		  sgtty.h unistd.h elf_hp.h ctype.h time.h locale.h])
 AC_CHECK_HEADERS(link.h, [], [],
 [#if HAVE_SYS_TYPES_H
 # include <sys/types.h>
@@ -740,6 +740,7 @@ AC_CHECK_HEADERS(term.h, [], [],
 
 AC_CHECK_DECLS([free, malloc, realloc, strerror, strstr, getopt,
                 snprintf, vsnprintf])
+AM_LC_MESSAGES
 
 # ----------------------- #
 # Checks for structures.  #
@@ -775,7 +776,7 @@ AC_FUNC_VFORK
 AC_CHECK_FUNCS([canonicalize_file_name realpath getrusage getuid \
                 getgid poll pread64 sbrk setpgid setpgrp setsid \
 		sigaction sigprocmask sigsetmask socketpair syscall \
-		ttrace wborder])
+		ttrace wborder setlocale])
 
 # Check the return and argument types of ptrace.  No canned test for
 # this, so roll our own.

Re: RFA: fix locale configure checks

[Thiago Jung Bauermann]

El mar, 25-11-2008 a las 10:22 -0700, Tom Tromey escribió:
> >>>>> "Tom" == Tom Tromey <tromey@redhat.com> writes:
> Tom> Today while exploring the charset code I happened to notice that gdb
> Tom> never calls setlocale.
> 
> Tom> The code is there, but configure never does the checks that might
> Tom> define HAVE_LOCALE_H, HAVE_SETLOCALE, or HAVE_LC_MESSAGES.

I just noticed that GDB sets LC_MESSAGES and LC_CTYPE. Do you or anybody
know if there a reason that it doesn't set LC_ALL instead?
-- 
[]'s
Thiago Jung Bauermann
IBM Linux Technology Center

Re: RFA: fix locale configure checks

[Tom Tromey]

>>>>> "Thiago" == Thiago Jung Bauermann <bauerman@br.ibm.com> writes:

Thiago> I just noticed that GDB sets LC_MESSAGES and LC_CTYPE. Do you or anybody
Thiago> know if there a reason that it doesn't set LC_ALL instead?

cvs annotate claims I checked in this code in 2002.  Funny.

I can't find the original note.  I don't know why it works this way;
it might just be boilerplate from when I gettextized a number of
programs.

Tom

Re: RFA: fix locale configure checks

[Joel Brobecker]

> 2008-11-25  Tom Tromey  <tromey@redhat.com>
> 
> 	* config.in, configure: Rebuild.
> 	* configure.ac: Check for locale.h, setlocale.  Call
> 	AM_LC_MESSAGES.
> 	* acinclude.m4: Include lcmessage.m4.

Looks good to me.

Thank you,
-- 
Joel

Re: RFA: fix locale configure checks

[Tom Tromey]

>> 2008-11-25  Tom Tromey  <tromey@redhat.com>
>> * config.in, configure: Rebuild.
>> * configure.ac: Check for locale.h, setlocale.  Call
>> AM_LC_MESSAGES.
>> * acinclude.m4: Include lcmessage.m4.

Joel> Looks good to me.

Thanks, I checked this in.

Tom