bug#15365: 24.3.50; Segfault when calling vc-git-log-edit-toggle-amend

[Eli Zaretskii]

> From: Dmitry Gutov <dgutov@yandex.ru>
> Date: Fri, 13 Sep 2013 14:55:27 +0300
> 
> Edit a file in a Git repository, then press
> 
> C-x v v, then C-c C-e

Too bad this requires to commit to a git repo.  If you cannot find a
recipe that doesn't need that, please show values of variables
involved in this abort.  The relevant code fragment is this:

      /* If window start is unchanged, we can reuse the whole matrix
         as is, after adjusting glyph positions.  No need to compute
         the window end again, since its offset from Z hasn't changed.  */
      r0 = MATRIX_FIRST_TEXT_ROW (current_matrix);
      if (CHARPOS (start) == MATRIX_ROW_START_CHARPOS (r0) + Z_delta
          && BYTEPOS (start) == MATRIX_ROW_START_BYTEPOS (r0) + Z_delta_bytes
          /* PT must not be in a partially visible line.  */
          && !(PT >= MATRIX_ROW_START_CHARPOS (row) + Z_delta
               && MATRIX_ROW_BOTTOM_Y (row) > window_text_bottom_y (w)))
        {
          /* Adjust positions in the glyph matrix.  */
          if (Z_delta || Z_delta_bytes)
            {
              struct glyph_row *r1
                = MATRIX_BOTTOM_TEXT_ROW (current_matrix, w);
              increment_matrix_positions (w->current_matrix,
                                          MATRIX_ROW_VPOS (r0, current_matrix),
                                          MATRIX_ROW_VPOS (r1, current_matrix),
                                          Z_delta, Z_delta_bytes);
            }

          /* Set the cursor.  */
          row = row_containing_pos (w, PT, r0, NULL, 0);
          if (row)
            set_cursor_from_row (w, row, current_matrix, 0, 0, 0, 0);
          else
            emacs_abort ();
          return 1;
        }

So the values of PT, Z, and Z_BYTE are important.  Also, the following
GDB commands can tell important details:

  (gdb) pmtxrows current_matrix

Note that you need to be in the call-stack frame of this code, frame
#3 according to the backtrace you show, to be able to reference these
variables.

Thanks.

> (gdb) xbacktrace
> "redisplay_internal (C function)" (0xbf6108)
> (gdb)
> 
> (For all I know, the above may be not accurate, because I loaded the
> src/.gdbinit file only after Emacs started up, worked and crashed.)

It is accurate.  However, if you start GDB from the Emacs's src
directory, you will be free of that doubt.