[obv] varobj.c: Fix GCC false warning

[obv] varobj.c: Fix GCC false warning

[Jan Kratochvil]

Hi,

CFLAGS=-O0 ./configure --without-python; make
varobj.c: In function ‘value_get_print_value’:
varobj.c:2578:22: error: ‘str_addr’ may be used uninitialized in this function

Reproducible with FSF GCC 4.4, 4.5, HEAD.

Checked in.


Thanks,
Jan


http://sourceware.org/ml/gdb-cvs/2010-10/msg00100.html

--- src/gdb/ChangeLog	2010/10/15 23:08:04	1.12264
+++ src/gdb/ChangeLog	2010/10/17 08:43:45	1.12265
@@ -1,3 +1,8 @@
+2010-10-17  Jan Kratochvil  <jan.kratochvil@redhat.com>
+
+	Fix GCC false warning.
+	* varobj.c (value_get_print_value) <str_addr>: Initialize it.
+
 2010-10-16  Pierre Muller  <muller@ics.u-strasbg.fr>
 
 	* p-typeprint.c (pascal_type_print_method_args): Fix problem in
--- src/gdb/varobj.c	2010/10/15 18:54:13	1.162
+++ src/gdb/varobj.c	2010/10/17 08:43:46	1.163
@@ -2486,7 +2486,8 @@
   long len = 0;
   char *encoding = NULL;
   struct gdbarch *gdbarch = NULL;
-  CORE_ADDR str_addr;
+  /* Initialize it just to avoid a GCC false warning.  */
+  CORE_ADDR str_addr = 0;
   int string_print = 0;
 
   if (value == NULL)

Re: [obv] varobj.c: Fix GCC false warning

[Andreas Schwab]

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

> --- src/gdb/varobj.c	2010/10/15 18:54:13	1.162
> +++ src/gdb/varobj.c	2010/10/17 08:43:46	1.163
> @@ -2486,7 +2486,8 @@
>    long len = 0;
>    char *encoding = NULL;
>    struct gdbarch *gdbarch = NULL;
> -  CORE_ADDR str_addr;
> +  /* Initialize it just to avoid a GCC false warning.  */
> +  CORE_ADDR str_addr = 0;
>    int string_print = 0;

What happens if string_print is set but gdbpy_extract_lazy_string was
not called?

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

[patch] varobj.c: Reset string_print [Re: [obv] varobj.c: Fix GCC false warning]

[Jan Kratochvil]

On Sun, 17 Oct 2010 12:02:08 +0200, Andreas Schwab wrote:
> What happens if string_print is set but gdbpy_extract_lazy_string was
> not called?

I agree.

BTW this is unrelated to the [obv] fix, in fact GCC warned on a falsely
uninitialized value <! HAVE_PYTHON> and GCC missed a real uninitialized value
<HAVE_PYTHON>.

I guess this patch should fix it.
No regressions on {x86_64,x86_64-m32,i686}-fedora14snapshot-linux-gnu.


Thanks,
Jan


gdb/
2010-10-17  Jan Kratochvil  <jan.kratochvil@redhat.com>

	* varobj.c (value_get_print_value): Reset STRING_PRINT if STR_ADDR is
	not set.

--- a/gdb/varobj.c
+++ b/gdb/varobj.c
@@ -2557,9 +2557,14 @@ value_get_print_value (struct value *value, enum varobj_display_formats format,
 			make_cleanup (xfree, thevalue);
 		      }
 		    else
-		      gdbpy_print_stack ();
+		      {
+			gdbpy_print_stack ();
+			string_print = 0;
+		      }
 		  }
 	      }
+	    else
+	      string_print = 0;
 	    if (replacement)
 	      value = replacement;
 	  }