chore(benchmarks): note the format_table/format_markdown column sync

[lesnik512]

What

format_table and format_markdown mirror the same 9 columns, per-column precision, and gated-vs-informational footer facts in two skins (fixed-width vs Markdown). Add a one-line pointer in each so an edit to one is a reminder to update the other.

Why not factor them out

Reviewed and deliberately not abstracted: with only two renderers and a stable column set, a shared column-descriptor table would cost more indirection than the duplication costs risk (premature DRY). The comment pins the drift hazard for ~zero cost; the break-even for factoring is a genuine third output format or churning columns.

Comment-only, no behavior change. ruff + ty clean, 20/20 tests/test_benchmarks.py pass.

🤖 Generated with Claude Code

[github-actions]

Benchmark gate

✅ gate passed

scenario	msg/s	delete/msg	WALrec/msg	WALB/msg	fpi	upd	del	dead_tup
consumer/w1/b10	918	1.000	6.78	916	242	5000	5000	10000
consumer/w1/b100	936	1.000	6.82	921	243	5000	5000	10000
consumer/w2/b10	1064	1.000	6.81	966	242	5000	5000	10000
consumer/w2/b100	1274	1.000	6.73	960	243	5000	5000	10000
consumer/w4/b10	954	1.000	6.74	945	242	5000	5000	10000
consumer/w4/b100	1437	1.000	6.87	1037	244	5000	5000	10000
consumer/w1/b100/tfbs100	5198	0.010	6.07	1116	243	5000	5000	10000
producer/w1/b100	2079	0.000	3.04	584	0	0	0	0

Gated (fails the build): delete_calls + tuple counters (upd/del/ins) + the producer's insert_calls/select_calls, exact; wal_records within a 10% band. msg/s, WAL bytes and total calls are informational (timing/FPI noise).