Problem report #24

Problem report #24

[Dan Nicolaescu]

CID: 24
Checker: FORWARD_NULL (help)
File: emacs/src/xfaces.c
Function: Fx_list_fonts
Description: Variable "f" tracked as NULL was passed to a function that dereferences it.

Event assign_zero: Variable "f" assigned value 0.
Also see events: [var_deref_model]

3022 	      f = NULL;
3023 	      face = Qnil;
3024 	    }
3025 	
3026 	  /* Determine the width standard for comparison with the fonts we find.  */
3027 	

At conditional (1): "face == Qnil" taking false path

3028 	  if (NILP (face))
3029 	    size = 0;
3030 	  else
3031 	    {
3032 	      /* This is of limited utility since it works with character
3033 		 widths.  Keep it for compatibility.  --gerd.  */

Event var_deref_model: Variable "f" tracked as NULL was passed to a function that dereferences it. [model]
Also see events: [assign_zero]

3034 	      int face_id = lookup_named_face (f, face, 0, 0);
3035 	      struct face *face = (face_id < 0
3036 				   ? NULL
3037 				   : FACE_FROM_ID (f, face_id));
3038 	

Re: Problem report #24

[Stefan Monnier]

> Event assign_zero: Variable "f" assigned value 0.
> Also see events: [var_deref_model]

> 3022 	      f = NULL;
> 3023 	      face = Qnil;
> 3024 	    }
> 3025 	
> 3026 	  /* Determine the width standard for comparison with the fonts we find.  */
> 3027 	

> At conditional (1): "face == Qnil" taking false path

> 3028 	  if (NILP (face))
> 3029 	    size = 0;
> 3030 	  else
> 3031 	    {
> 3032 	      /* This is of limited utility since it works with character
> 3033 		 widths.  Keep it for compatibility.  --gerd.  */

> Event var_deref_model: Variable "f" tracked as NULL was passed to a function that dereferences it. [model]
> Also see events: [assign_zero]

If f is NULL, then face=Qnil, so the false path of face==Qnil won't
be taken.


        Stefan

Re: Problem report #24

[Jan D.]

Stefan Monnier wrote:
>> Event assign_zero: Variable "f" assigned value 0.
>> Also see events: [var_deref_model]
> 
>> 3022 	      f = NULL;
>> 3023 	      face = Qnil;
>> 3024 	    }
>> 3025 	
>> 3026 	  /* Determine the width standard for comparison with the fonts we find.  */
>> 3027 	
> 
>> At conditional (1): "face == Qnil" taking false path
> 
>> 3028 	  if (NILP (face))
>> 3029 	    size = 0;
>> 3030 	  else
>> 3031 	    {
>> 3032 	      /* This is of limited utility since it works with character
>> 3033 		 widths.  Keep it for compatibility.  --gerd.  */
> 
>> Event var_deref_model: Variable "f" tracked as NULL was passed to a function that dereferences it. [model]
>> Also see events: [assign_zero]
> 
> If f is NULL, then face=Qnil, so the false path of face==Qnil won't
> be taken.
> 

face could be Qnil upon entry to the function?

	Jan D.

Re: Problem report #24

[Stefan Monnier]

>>> Event assign_zero: Variable "f" assigned value 0.
>>> Also see events: [var_deref_model]
>> 
>>> 3022 	      f = NULL;
>>> 3023 	      face = Qnil;
>>> 3024 	    }
>>> 3025 	
>>> 3026 	  /* Determine the width standard for comparison with the fonts we find.  */
>>> 3027 	
>> 
>>> At conditional (1): "face == Qnil" taking false path
>> 
>>> 3028 	  if (NILP (face))
>>> 3029 	    size = 0;
>>> 3030 	  else
>>> 3031 	    {
>>> 3032 	      /* This is of limited utility since it works with character
>>> 3033 		 widths.  Keep it for compatibility.  --gerd.  */
>> 
>>> Event var_deref_model: Variable "f" tracked as NULL was passed to a function that dereferences it. [model]
>>> Also see events: [assign_zero]
>> If f is NULL, then face=Qnil, so the false path of face==Qnil won't
>> be taken.
>> 

> face could be Qnil upon entry to the function?

Sure, no problem: the problem report is when f==NULL and face!=Qnil.


        Stefan

Re: Problem report #24

[Richard Stallman]

    > face could be Qnil upon entry to the function?

    Sure, no problem: the problem report is when f==NULL and face!=Qnil.

I am not sure what "no problem" means here.
Are you saying there is no bug?

Re: Problem report #24

[Stefan Monnier]

>> face could be Qnil upon entry to the function?
>     Sure, no problem: the problem report is when f==NULL and face!=Qnil.

> I am not sure what "no problem" means here.
> Are you saying there is no bug?

Exactly: I see no bug.


        Stefan