Re: emacs-28 c05864dd25: Avoid another segfault in 'face_at_buffer_position'

[Robert Pluim <rpluim@gmail.com>]

>>>>> On Fri, 14 Jan 2022 09:12:24 -0500 (EST), Eli Zaretskii <eliz@gnu.org> said:

    Eli> branch: emacs-28
    Eli> commit c05864dd25c820fe250672d32d6ec77ba946670d
    Eli> Author: Eli Zaretskii <eliz@gnu.org>
    Eli> Commit: Eli Zaretskii <eliz@gnu.org>

    Eli>     Avoid another segfault in 'face_at_buffer_position'
    
    Eli>     * src/xfaces.c (face_at_buffer_position): Make really sure the
    Eli>     default face is usable.  (Bug#53254)
    Eli> ---
    Eli>  src/xfaces.c | 8 ++++++--
    Eli>  1 file changed, 6 insertions(+), 2 deletions(-)

    Eli> diff --git a/src/xfaces.c b/src/xfaces.c
    Eli> index b9fb7c0ac1..dff5ae346d 100644
    Eli> --- a/src/xfaces.c
    Eli> +++ b/src/xfaces.c
    Eli> @@ -6425,8 +6425,12 @@ face_at_buffer_position (struct window *w, ptrdiff_t pos,
 
    Eli>      default_face = FACE_FROM_ID_OR_NULL (f, face_id);
    Eli>      if (!default_face)
    Eli> -      default_face = FACE_FROM_ID (f,
    Eli> -				   lookup_basic_face (w, f, DEFAULT_FACE_ID));
    Eli> +      {
    Eli> +	if (FRAME_FACE_CACHE (f)->used == 0)
    Eli> +	  recompute_basic_faces (f);
    Eli> +	default_face = FACE_FROM_ID (f,
    Eli> +				     lookup_basic_face (w, f, DEFAULT_FACE_ID));
    Eli> +      }
    Eli>    }

If youʼre going to need to check the face cache, how about doing that
unconditionally? ie removing FACE_FROM_ID_OR_NULL and the
!default_face bit?

Robert
--