Honour q on the last chunk of a kitty transmission, bump to 2.0.5

[tomlm]

Summary

• A kitty continuation chunk may carry "only the m and optionally q keys", and notcurses relies on that: its first chunk has no q and its last chunk says q=2. XTerm.NET read the quiet level from the first chunk only, so it sent OK to a client that had asked for no reply. Once notcurses exited, that reply reached the shell as typed input: Gi=13871133;OK at the prompt after notcurses-info.
• New KittyTransmission.ApplyContinuation takes q from a later chunk, and the reply sent when the image completes follows it.
• Bumps version to 2.0.5.

Tests

• Quiet_on_the_last_chunk_silences_the_reply replays notcurses' chunk pattern. Without the fix it fails with exactly ESC _Gi=31;OK ESC \.
• A_chunked_transmission_without_quiet_still_replies checks that OK still comes when no chunk sets q.
• Full suite: 2237 passed, 1 skipped.

🤖 Generated with Claude Code

[github-actions]

Perf comparison — this change, against its base

3 run(s) of each side, alternating on one machine. Allocation is a count and is gated exactly. Time is a measurement, so its gate is derived from the spread this job just observed in itself rather than fixed in advance.

corpus	bytes/char	gen0/Mchar	ns/char	Δ time	noise	gate
scroll-ascii	0.00 → 0.00	0.00 → 0.00	3.38 → 3.38	-0.0%	±3%	8%
sgr-churn	0.00 → 0.00	0.00 → 0.00	8.97 → 9.08	+1.2%	±1%	4%
truecolor	0.00 → 0.00	0.00 → 0.00	9.99 → 10.15	+1.6%	±4%	13%
alt-redraw	0.00 → 0.00	0.00 → 0.00	14.35 → 14.42	+0.5%	±1%	4%
unicode	7.66 → 7.66	0.45 → 0.45	35.00 → 34.93	-0.2%	±2%	5%
flood	0.00 → 0.00	0.00 → 0.00	95.26 → 95.14	-0.1%	±0%	4%

Each corpus is gated at max(4%, 3 × its own noise). A wide noise column means this runner was busy and the timing half of the table should be read as advisory; the allocation half is exact either way.

assemblies measured

• base: XTerm.NET 2.0.4.0 mvid:edd04189-f427-47d9-a741-050e80dc445b
• head: XTerm.NET 2.0.5.0 mvid:8698ccc0-dc90-4700-8dd3-726e4b5fc26d

---

Perf comparison — cumulative, everything since 2.0.3

3 run(s) of each side, alternating on one machine. Allocation is a count and is gated exactly. Time is a measurement, so its gate is derived from the spread this job just observed in itself rather than fixed in advance.

corpus	bytes/char	gen0/Mchar	ns/char	Δ time	noise	gate
scroll-ascii	0.00 → 0.00	0.00 → 0.00	3.38 → 3.38	+0.2%	±1%	5%
sgr-churn	0.00 → 0.00	0.00 → 0.00	9.01 → 9.08	+0.7%	±1%	5%
truecolor	0.00 → 0.00	0.00 → 0.00	10.11 → 10.15	+0.4%	±4%	11%
alt-redraw	0.00 → 0.00	0.00 → 0.00	14.39 → 14.42	+0.2%	±1%	5%
unicode	7.66 → 7.66	0.45 → 0.45	35.82 → 34.93	-2.5%	±6%	18%
flood	0.00 → 0.00	0.00 → 0.00	95.46 → 95.14	-0.3%	±2%	6%

Each corpus is gated at max(5%, 3 × its own noise). A wide noise column means this runner was busy and the timing half of the table should be read as advisory; the allocation half is exact either way.

assemblies measured

• base: XTerm.NET 2.0.3.0 mvid:9ba0ce98-0d1e-4424-8aa1-9ab012cdc07a
• head: XTerm.NET 2.0.5.0 mvid:8698ccc0-dc90-4700-8dd3-726e4b5fc26d