test: fix stale row/function reuse across filter + xref-select blocks

Two latent ordering bugs, exposed once the suite actually opened the big function:
- the 'sub_1*' filter round-trip left the list filtered, so reusing biggest_i (an
  unfiltered row index) clamped move_cursor to a tiny stub function (4-line
  decompile, no calls), cascading into ~7 skipped/failed checks. Clear the filter
  before reusing biggest_i.
- the xref-select sub-test navigates 'dis' to another function; the following
  column-cursor block assumed the original function was still loaded at call_idx,
  so h/l ran on an uncached (empty) line and clamped x to 0. Reload the original
  function and materialize the call line first.

pilot suite 59/59.

0 files changed, 0 insertions, 0 deletions