q1 — polar maximality and the Q5 integer slice

Replay:

sh compute/q1/run_all.sh

Script	What it certifies
`polar_vertices.py` / `polar.c`	polars of $D_5,L_5,Q_5,R_5$ have $\max\\|x\\|^2=5/4<2$
`replay_max_vertex.py`	Fraction GE rebuild of each recorded vertex
`integer_q5_44.c`	integer $T_{Q_5}$ distributions at $N=44$ are empty
`dump_q5_tables.py`	C tables match Gegenbauer; $Q_5$ histogram passes
`check_q5_44_empty.py`	header / witness / empty-scan replay
`a4_containing.py`	height $\\|s\\|\ge 2$ ; discrete extras cap at 20
`integer_restricted.py`	$T_{L_5}$ still has integer hits at 41, 42, 43
`restricted_duals.py`	no new interpolating dual below 44 on $T_{Q_5}$

The unrestricted range remains $40\le\tau_5\le 44$ .