Fix Windows IPC and profiling build

[pcarrier]

No description provided.

[indent]

PR Summary

Fixes the Windows build of blit (including the profiling cargo profile) and two Windows runtime bugs, with no changes to protocol, APIs, or unix/linux behavior.

• Windows IPC connect race: connect_ipc now retries ERROR_PIPE_BUSY (up to ~1s) so a client that probes right after a server accepts another client no longer fails while the server is creating its next pipe instance.
• Windows PTY working directory: respawn_child now passes the requested directory as lpCurrentDirectory to CreateProcessW instead of null, so restarted Windows PTYs start in the right directory (matching spawn_pty and unix).
• Build hygiene (cfg gating): platform-only symbols are correctly gated — kicked_reason to #[cfg(unix)], OpenProcess/WaitForSingleObject/WAIT_OBJECT_0 to #[cfg(all(windows, test))], FEATURE_CREATE_EXEC to unix, plus let _ = binds for mode (fssync) and app_id/instance_id (non-Linux) — all matching actual usage.
• build.ps1: new local PowerShell script that reproduces the Windows build (browser wasm → js packages → cargo build -p blit-cli --profile profiling). Dev tooling only; CI is unchanged.

Issues

Review closed.

_{View session}

[github-actions]

Coverage

Crate	Lines	Functions	Regions
alacritty-driver	75.7% (934/1234)	78.7% (74/94)	79.0% (1521/1926)
browser	0.0% (0/825)	0.0% (0/69)	0.0% (0/1404)
cli	40.0% (6227/15563)	46.0% (619/1346)	40.9% (9702/23715)
compositor	35.1% (6045/17209)	49.8% (502/1009)	34.9% (8309/23786)
desktop	78.7% (4409/5605)	71.9% (388/540)	75.4% (6119/8120)
fonts	85.3% (756/886)	89.9% (71/79)	86.4% (1485/1719)
fssync	92.6% (5520/5962)	94.4% (501/531)	92.8% (10189/10983)
gateway	35.1% (698/1990)	40.6% (69/170)	32.1% (1059/3303)
git	87.7% (4642/5295)	90.2% (378/419)	87.5% (7408/8463)
guest	82.8% (2347/2836)	80.5% (265/329)	81.4% (3911/4805)
lsp	77.2% (2688/3483)	79.4% (262/330)	74.9% (4210/5619)
proxy	19.2% (172/898)	20.5% (26/127)	21.0% (293/1392)
remote	91.7% (18064/19694)	94.7% (1270/1341)	89.4% (29291/32748)
sd-notify	73.9% (68/92)	100.0% (6/6)	83.2% (109/131)
server	58.7% (35532/60536)	65.7% (3053/4647)	60.2% (54280/90237)
ssh	32.2% (165/512)	48.2% (27/56)	31.4% (261/830)
upsidedown	31.4% (391/1247)	27.8% (55/198)	34.8% (797/2287)
webrtc-forwarder	8.5% (238/2805)	10.7% (22/205)	6.3% (289/4595)
webserver	65.1% (1288/1978)	67.6% (177/262)	67.2% (2159/3214)
Total	60.7% (90184/148650)	66.0% (7765/11758)	61.7% (141392/229277)

[blacksmith-sh]

Found 16 test failures on Blacksmith runners:

Failures

Test	View Logs
[chromium] › tests/ mobile-keyboard.spec.ts:200:5 › a tap while the keyboard failed to rise retries instead of hiding	View Logs
[chromium] › tests/ mobile-keyboard.spec.ts:256:5 › the switcher stays in view when the keyboard pans the v isual viewport	View Logs
[chromium] › tests/ mobile-keyboard.spec.ts:293:5 › key-line taps cancel the touch so the keyboard stays up	View Logs
[chromium] › tests/ mobile-keyboard.spec.ts:362:5 › the icon's keyboard survives focus landing on a surface canvas	View Logs
[chromium] › tests/ mobile-keyboard.spec.ts:455:5 › iOS hops focus through a host when the target already h olds it	View Logs
[chromium] › tests/ mobile-keyboard.spec.ts:61:5 › keyboard rises only from the toggle and the key line tra cks it	View Logs
[chromium] › tests/ parked-drag.spec.ts:50:7 › Parked pane drag › a parked terminal is inert and drags into the main view	View Logs
[chromium] › tests/ parked-drag.spec.ts:94:7 › Parked pane drag › a parked card drags into a specific BSP p ane	View Logs
[chromium] › tests/ text-prediction.spec.ts:136:7 › host text prediction › a proposal shows in a chip on th e cursor's line and is not forwarded	View Logs
[chromium] › tests/ text-prediction.spec.ts:184:7 › host text prediction › with prediction off the field st ays empty, as it always was	View Logs

...and 6 more test failures. View all on Blacksmith

^{Need help on this PR? Tag @codesmith-bot with what you need.}