bug#65060: 29.1.50; display_count_lines segv

[Eli Zaretskii]

> From: Kai Ma <justksqsf@gmail.com>
> Date: Sat, 12 Aug 2023 15:28:16 +0800
> Cc: 65060@debbugs.gnu.org
> 
> > On Aug 12, 2023, at 14:50, Eli Zaretskii <eliz@gnu.org> wrote:
> > 
> > Ping!  Can you please help me finish investigating this bug by
> > providing the information I asked for?  Armed with that information, I
> > think I will be able to find a solution.
> 
> Sorry for the late reply!
> 
> cursor (before do) = 0x0
> * GPT_BYTE = 3072
> * GAP_SIZE = 2000
> * BEG_ADDR = 0x0
> * current_buffer->text->beg = 0x0
> 
> cursor (before memchr) = 0x0
> * GPT_BYTE = 3072
> * GAP_SIZE = 2000
> * BEG_ADDR = 0x0
> * current_buffer->text->beg = 0x0

Thanks.  My guess was correct: the buffer in question was killed.

Could you please try the patch below, and see if it avoids the
crashes?  The patch causes Emacs to signal an error when
format-mode-line is called for a dead buffer, so if I did this
correctly, you should see that Emacs no longer crashes, but there are
error messages about a dead buffer in *Messages*.  If this is what
happens, you should then take this up with the dirvish developers, and
ask them to avoid calling format-mode-line for a killed buffer.

diff --git a/src/fns.c b/src/fns.c
index 2ed62d6..af5f947 100644
--- a/src/fns.c
+++ b/src/fns.c
@@ -6123,6 +6123,9 @@ DEFUN ("line-number-at-pos", Fline_number_at_pos,
 {
   ptrdiff_t pos_byte, start_byte = BEGV_BYTE;
 
+  if (!BUFFER_LIVE_P (current_buffer))
+    error ("Attempt to count lines in a dead buffer");
+
   if (MARKERP (position))
     {
       /* We don't trust the byte position if the marker's buffer is
diff --git a/src/xdisp.c b/src/xdisp.c
index 2250897..1141707 100644
--- a/src/xdisp.c
+++ b/src/xdisp.c
@@ -27653,6 +27653,8 @@ DEFUN ("format-mode-line", Fformat_mode_line, 
Sformat_mode_line,
   if (NILP (buffer))
     buffer = w->contents;
   CHECK_BUFFER (buffer);
+  if (!BUFFER_LIVE_P (XBUFFER (buffer)))
+    error ("Attempt to format a mode line for a dead buffer");
 
   /* Make formatting the modeline a non-op when noninteractive, otherwise
      there will be problems later caused by a partially initialized frame.  */