[patch/5.2/commit] Zap __func__

[patch/5.2/commit] Zap __func__

[Andrew Cagney]

[-- Attachment #1: Type: text/plain, Size: 117 bytes --]

Just FYI,

I've committed this to the 5.2 branch - zap more __func__s.  It's brutal 
but it works :-)

enjoy,
Andrew

[-- Attachment #2: diffs --]
[-- Type: text/plain, Size: 1154 bytes --]

2002-07-15  Andrew Cagney  <ac131313@redhat.com>

	* dwarf2cfi.c: Replace __func__ with "?func?".

Index: dwarf2cfi.c
===================================================================
RCS file: /cvs/src/src/gdb/dwarf2cfi.c,v
retrieving revision 1.1.2.5
diff -u -r1.1.2.5 dwarf2cfi.c
--- dwarf2cfi.c	3 Jul 2002 06:00:26 -0000	1.1.2.5
+++ dwarf2cfi.c	15 Jul 2002 16:40:58 -0000
@@ -860,7 +860,7 @@
     }
   else
     internal_error (__FILE__, __LINE__,
-		    "%s(): Internal error: fde->cie_ptr==NULL !", __func__);
+		    "%s(): Internal error: fde->cie_ptr==NULL !", "?func?");
 }
 
 static void
@@ -1361,7 +1361,7 @@
 	      orig_context->reg[fs->regs.reg[i].loc.reg].loc.addr;
 	  default:
 	    internal_error (__FILE__, __LINE__,
-			    "%s: unknown register rule", __func__);
+			    "%s: unknown register rule", "?func?");
 	  }
 	break;
       case REG_SAVED_EXP:
@@ -1379,7 +1379,7 @@
 	break;
       default:
 	internal_error (__FILE__, __LINE__,
-			"%s: unknown register rule", __func__);
+			"%s: unknown register rule", "?func?");
       }
   get_reg ((char *) &context->ra, context, fs->retaddr_column);
   unwind_tmp_obstack_free ();

Re: [patch/5.2/commit] Zap __func__

[Daniel Jacobowitz]

On Mon, Jul 15, 2002 at 12:42:49PM -0400, Andrew Cagney wrote:
> Just FYI,
> 
> I've committed this to the 5.2 branch - zap more __func__s.  It's brutal 
> but it works :-)
> 
> enjoy,
> Andrew

> 2002-07-15  Andrew Cagney  <ac131313@redhat.com>
> 
> 	* dwarf2cfi.c: Replace __func__ with "?func?".

Er, hunh?

First of all, is there any reason that __FUNCTION__ is not adequately
portable?  I think it is.  Second of all, if you're going to remove
__func__ you could at least replace it with the name of the function.

-- 
Daniel Jacobowitz                           Carnegie Mellon University
MontaVista Software                         Debian GNU/Linux Developer

Re: [patch/5.2/commit] Zap __func__

[Andrew Cagney]

> On Mon, Jul 15, 2002 at 12:42:49PM -0400, Andrew Cagney wrote:
> 
>> Just FYI,
>> 
>> I've committed this to the 5.2 branch - zap more __func__s.  It's brutal 
>> but it works :-)
>> 
>> enjoy,
>> Andrew
> 
> 
>> 2002-07-15  Andrew Cagney  <ac131313@redhat.com>
>> 
>> 	* dwarf2cfi.c: Replace __func__ with "?func?".
> 
> 
> Er, hunh?
> 
> First of all, is there any reason that __FUNCTION__ is not adequately
> portable?  I think it is.  Second of all, if you're going to remove
> __func__ you could at least replace it with the name of the function.

Remember this is a branch and those ``__func__''s were only printed when 
there was an internal_error() - I don't think anyone is going to notice 
:-).  The correct clean fix was committed to the mainline a few hours 
earlier (I looked at back patching it but noticed too many differences).

As for __FUNCTION__, that isn''t part of ISO C 90.

enjoy,
Andrew

Re: [patch/5.2/commit] Zap __func__

[Daniel Jacobowitz]

On Mon, Jul 15, 2002 at 04:09:35PM -0400, Andrew Cagney wrote:
> >On Mon, Jul 15, 2002 at 12:42:49PM -0400, Andrew Cagney wrote:
> >
> >>Just FYI,
> >>
> >>I've committed this to the 5.2 branch - zap more __func__s.  It's brutal 
> >>but it works :-)
> >>
> >>enjoy,
> >>Andrew
> >
> >
> >>2002-07-15  Andrew Cagney  <ac131313@redhat.com>
> >>
> >>	* dwarf2cfi.c: Replace __func__ with "?func?".
> >
> >
> >Er, hunh?
> >
> >First of all, is there any reason that __FUNCTION__ is not adequately
> >portable?  I think it is.  Second of all, if you're going to remove
> >__func__ you could at least replace it with the name of the function.
> 
> Remember this is a branch and those ``__func__''s were only printed when 
> there was an internal_error() - I don't think anyone is going to notice 
> :-).  The correct clean fix was committed to the mainline a few hours 
> earlier (I looked at back patching it but noticed too many differences).

-           internal_error (__FILE__, __LINE__,
-                           "%s: unknown register rule", __func__);
+           internal_error (__FILE__, __LINE__, "bad switch");

Is there any reason not to leave the error message as it was?  It's
just a matter of "update_context: unknown register rule".  Replacing a
clear internal error with "bad switch" doesn't seem like a good move.

(having an internal_error here is a little shady anyway, it's like
abort()ing on user input)

> As for __FUNCTION__, that isn''t part of ISO C 90.

Yep, you're right.

-- 
Daniel Jacobowitz                           Carnegie Mellon University
MontaVista Software                         Debian GNU/Linux Developer

Re: [patch/5.2/commit] Zap __func__

[Andrew Cagney]

> Remember this is a branch and those ``__func__''s were only printed when 
>> there was an internal_error() - I don't think anyone is going to notice 
>> :-).  The correct clean fix was committed to the mainline a few hours 
>> earlier (I looked at back patching it but noticed too many differences).
> 
> 
> -           internal_error (__FILE__, __LINE__,
> -                           "%s: unknown register rule", __func__);
> +           internal_error (__FILE__, __LINE__, "bad switch");
> 
> Is there any reason not to leave the error message as it was?  It's
> just a matter of "update_context: unknown register rule".  Replacing a
> clear internal error with "bad switch" doesn't seem like a good move.

If it really is an internal error due to a messed up switch then I think 
the message is sufficient.  However ....

> (having an internal_error here is a little shady anyway, it's like
> abort()ing on user input)

.... it sounds like these should be replaced by error() (or complain()). 
  If it is triggered by a bad input file then, yes, you are correct.

Andrew