[PATCH] Fix PR gdb/290

[PATCH] Fix PR gdb/290

[kettenis]

Checked in as obvious.  Wonder why GCC didn't warn about this.  Should
we add some more -Wxxx flags?

Mark


Index: ChangeLog
from  Mark Kettenis  <kettenis@gnu.org>

	* doublest.c (store_typed_floating): Don't try to return a value.
	Fixes PR gdb/290.

Index: doublest.c
===================================================================
RCS file: /cvs/src/src/gdb/doublest.c,v
retrieving revision 1.8
diff -u -p -r1.8 doublest.c
--- doublest.c 2002/01/22 19:57:40 1.8
+++ doublest.c 2002/02/09 17:24:44
@@ -732,9 +732,9 @@ store_typed_floating (void *addr, const 
   memset (addr, 0, TYPE_LENGTH (type));
 
   if (TYPE_FLOATFORMAT (type) == NULL)
-    return store_floating (addr, TYPE_LENGTH (type), val);
-
-  floatformat_from_doublest (TYPE_FLOATFORMAT (type), &val, addr);
+    store_floating (addr, TYPE_LENGTH (type), val);
+  else
+    floatformat_from_doublest (TYPE_FLOATFORMAT (type), &val, addr);
 }
 
 /* Convert a floating-point number of type FROM_TYPE from a

Re: [PATCH] Fix PR gdb/290

[Andrew Cagney]

[added gcc]


> Checked in as obvious.  Wonder why GCC didn't warn about this.  Should
> we add some more -Wxxx flags?


I was looking at it and wondering the same thing - I know I've fixed 
botched return type warnings.  Trying:

static int f1 (void) { return 1; }
static void f2 (void) { return 1; }
static void f3 (void) { return; }
static int f4 (void) { return; }

I get:

-Wimplicit -Wreturn-type -Wcomment -Wtrigraphs -Wformat -Wparentheses 
-Wpointer-arith -Wuninitialized -Werror
doublest.c: In function `f2':
doublest.c:741: warning: `return' with a value, in function returning void
doublest.c: In function `f4':
doublest.c:743: warning: `return' with no value, in function returning 
non-void

so ok so far.  Hmm, I think this tells the story:

static void f5 (int a) { return f3(); }

Returning the result from a function that returns void (f3()) doesn't 
attract a warning.

Bug or feature?

	Andrew


> Mark
> 
> 
> Index: ChangeLog
> from  Mark Kettenis  <kettenis@gnu.org>
> 
> * doublest.c (store_typed_floating): Don't try to return a value.
> 	Fixes PR gdb/290.
> 
> Index: doublest.c
> ===================================================================
> RCS file: /cvs/src/src/gdb/doublest.c,v
> retrieving revision 1.8
> diff -u -p -r1.8 doublest.c
> --- doublest.c 2002/01/22 19:57:40 1.8
> +++ doublest.c 2002/02/09 17:24:44
> @@ -732,9 +732,9 @@ store_typed_floating (void *addr, const 
>    memset (addr, 0, TYPE_LENGTH (type));
>  
>    if (TYPE_FLOATFORMAT (type) == NULL)
> -    return store_floating (addr, TYPE_LENGTH (type), val);
> -
> -  floatformat_from_doublest (TYPE_FLOATFORMAT (type), &val, addr);
> +    store_floating (addr, TYPE_LENGTH (type), val);
> +  else
> +    floatformat_from_doublest (TYPE_FLOATFORMAT (type), &val, addr);
>  }
>  
>  /* Convert a floating-point number of type FROM_TYPE from a
> 
> 

Re: [PATCH] Fix PR gdb/290

[Joe Buck]

Andrew writes:
> static void f3 (void) { return; }
> ...
> so ok so far.  Hmm, I think this tells the story:
> 
> static void f5 (int a) { return f3(); }
> 
> Returning the result from a function that returns void (f3()) doesn't 
> attract a warning.

> Bug or feature?

For C++, it is a non-optional feature; the standard requires this to work
correctly (a void function may say "return foop();" if foop's return type
is void).  The reason for this will become clear after you write templates
for a while; without it "void" becomes more of an oddball case.

For C89, I don't think it's legal though it may be reasonable to accept
as an extension; I don't know about C99.

Re: [PATCH] Fix PR gdb/290

[Andrew Cagney]

> For C89, I don't think it's legal though it may be reasonable to accept
> as an extension; I don't know about C99.


More info is at:


http://sources.redhat.com/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gdb&pr=290


It upset Sun's ISO C compiler.

enjoy,
Andrew

Re: [PATCH] Fix PR gdb/290

[Andreas Schwab]

Joe Buck <jbuck@synopsys.COM> writes:

|> For C89, I don't think it's legal though it may be reasonable to accept
|> as an extension; I don't know about C99.

The C99 standard is quite clear about this:

    6.8.6.4[#1] A return statement with an expression shall not appear in
    a function whose return type is void.  A return statement without an
    expression shall only appear in a function whose return type is void.

Andreas.

-- 
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE GmbH, Deutschherrnstr. 15-19, D-90429 Nürnberg
Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: [PATCH] Fix PR gdb/290

[Joe Buck]

> Joe Buck <jbuck@synopsys.COM> writes:
> 
> |> For C89, I don't think it's legal though it may be reasonable to accept
> |> as an extension; I don't know about C99.
> 
> The C99 standard is quite clear about this:
> 
>     6.8.6.4[#1] A return statement with an expression shall not appear in
>     a function whose return type is void.  A return statement without an
>     expression shall only appear in a function whose return type is void.

OK, then this can only be done in C++.  Any C code should avoid it.

for 
void foo();
void bar() { return foo();}

(which is legal C++)

I've just checked, and gcc 2.95.2 and 3.0.3 reject this type of C code
iff -ansi -pedantic is specified.  By default they accept it.  Anyone
developing code that must compile with many compilers probably should
be using -pedantic.