[PATCH] c-exp.y, memory leak

[PATCH] c-exp.y, memory leak

[msnyder]

[-- Attachment #1: Type: text/plain, Size: 0 bytes --]



[-- Attachment #2: 171.txt --]
[-- Type: text/plain, Size: 852 bytes --]

2007-08-04  Michael Snyder  <msnyder@access-company.com>

	* c-exp.y (parse_number): Memory leak.

Index: c-exp.y
===================================================================
RCS file: /cvs/src/src/gdb/c-exp.y,v
retrieving revision 1.38
diff -p -r1.38 c-exp.y
*** c-exp.y	12 Jun 2007 15:33:03 -0000	1.38
--- c-exp.y	5 Aug 2007 01:49:51 -0000
*************** parse_number (p, len, parsed_float, puti
*** 1096,1104 ****
  	    putithere->typed_val_float.type = 
  	      builtin_type (current_gdbarch)->builtin_long_double;
  	  else
! 	    return ERROR;
  	}
  
        return FLOAT;
      }
  
--- 1096,1108 ----
  	    putithere->typed_val_float.type = 
  	      builtin_type (current_gdbarch)->builtin_long_double;
  	  else
! 	    {
! 	      free (s);
! 	      return ERROR;
! 	    }
  	}
  
+       free (s);
        return FLOAT;
      }
  

Re: [PATCH] c-exp.y, memory leak

[Mark Kettenis]

> Date: Sat, 4 Aug 2007 18:51:09 -0700 (PDT)
> From: msnyder@sonic.net
> 
> 2007-08-04  Michael Snyder  <msnyder@access-company.com>
> 
> 	* c-exp.y (parse_number): Memory leak.

Shouldn't this use xfree()?

> Index: c-exp.y
> ===================================================================
> RCS file: /cvs/src/src/gdb/c-exp.y,v
> retrieving revision 1.38
> diff -p -r1.38 c-exp.y
> *** c-exp.y	12 Jun 2007 15:33:03 -0000	1.38
> --- c-exp.y	5 Aug 2007 01:49:51 -0000
> *************** parse_number (p, len, parsed_float, puti
> *** 1096,1104 ****
>   	    putithere->typed_val_float.type = 
>   	      builtin_type (current_gdbarch)->builtin_long_double;
>   	  else
> ! 	    return ERROR;
>   	}
>   
>         return FLOAT;
>       }
>   
> --- 1096,1108 ----
>   	    putithere->typed_val_float.type = 
>   	      builtin_type (current_gdbarch)->builtin_long_double;
>   	  else
> ! 	    {
> ! 	      free (s);
> ! 	      return ERROR;
> ! 	    }
>   	}
>   
> +       free (s);
>         return FLOAT;
>       }

Re: [PATCH] c-exp.y, memory leak

[Michael Snyder]

> > Date: Sat, 4 Aug 2007 18:51:09 -0700 (PDT)
> > From: msnyder@sonic.net
> >
> > 2007-08-04  Michael Snyder  <msnyder@access-company.com>
> >
> > * c-exp.y (parse_number): Memory leak.
>
> Shouldn't this use xfree()?

Ah, right, that's what I meant to say.   ;-)

>
> > Index: c-exp.y
> > ===================================================================
> > RCS file: /cvs/src/src/gdb/c-exp.y,v
> > retrieving revision 1.38
> > diff -p -r1.38 c-exp.y
> > *** c-exp.y 12 Jun 2007 15:33:03 -0000 1.38
> > --- c-exp.y 5 Aug 2007 01:49:51 -0000
> > *************** parse_number (p, len, parsed_float, puti
> > *** 1096,1104 ****
> >       putithere->typed_val_float.type =
> >         builtin_type (current_gdbarch)->builtin_long_double;
> >     else
> > !     return ERROR;
> >   }
> >
> >         return FLOAT;
> >       }
> >
> > --- 1096,1108 ----
> >       putithere->typed_val_float.type =
> >         builtin_type (current_gdbarch)->builtin_long_double;
> >     else
> > !     {
> > !       free (s);
> > !       return ERROR;
> > !     }
> >   }
> >
> > +       free (s);
> >         return FLOAT;
> >       }
>

Re: [PATCH] c-exp.y, memory leak

[msnyder]

>> Date: Sat, 4 Aug 2007 18:51:09 -0700 (PDT)
>> From: msnyder@sonic.net
>>
>> 2007-08-04  Michael Snyder  <msnyder@access-company.com>
>>
>> 	* c-exp.y (parse_number): Memory leak.
>
> Shouldn't this use xfree()?

Actually, although arguably they should -- I'm not so sure.

It's kind of a mess -- c-exp.y uses malloc, and then Makefile.in
runs sed on c-exp.c to transform malloc to xmalloc.  No such sed
transform is done to convert free to xfree, and there is at least
one other call to free in there that does not get converted.

I'm reluctant to mess with the sed scripts.  "free" is just too
common a string.

Re: [PATCH] c-exp.y, memory leak

[msnyder]

>>> Date: Sat, 4 Aug 2007 18:51:09 -0700 (PDT)
>>> From: msnyder@sonic.net
>>>
>>> 2007-08-04  Michael Snyder  <msnyder@access-company.com>
>>>
>>> 	* c-exp.y (parse_number): Memory leak.
>>
>> Shouldn't this use xfree()?
>
> Actually, although arguably they should -- I'm not so sure.
>
> It's kind of a mess -- c-exp.y uses malloc, and then Makefile.in
> runs sed on c-exp.c to transform malloc to xmalloc.  No such sed
> transform is done to convert free to xfree, and there is at least
> one other call to free in there that does not get converted.
>
> I'm reluctant to mess with the sed scripts.  "free" is just too
> common a string.

Alright -- I'm going to commit this using "free" rather than
"xfree", on the premise that the file already calls free once,
and doesn't call xfree at all.