Re: SIGSEGV -- redisplay bug

[Richard Stallman]

  /* Increment positions in glyphs.  */
  for (area = 0; area < LAST_AREA; ++area)
    for (i = 0; i < row->used[area]; ++i)
      if (BUFFERP (row->glyphs[area][i].object)
          && row->glyphs[area][i].charpos > 0)
        row->glyphs[area][i].charpos += delta;

    (gdb) p area
    $1 = 4159280

I don't see how anything here could enable area to get bigger than 2,
unless area is stored in the stack and some absurd pointers or indices
make row->glyphs[area][i].charpos alias with that stack slot.

But such aliasing should be impossible as long as AREA and I are
reasonable.  The glyphs are in the heap, nowhere near the stack.

Could you look at the values of row->used[0], row->used[1]
and row->used[2]?