JinjingandOrca 81dfa1db55 feat(gpu): add the GPU fallback ladder, cache purge, and driver-identity modules ...

`--disable-gpu` does not stop Chromium from spawning a GPU child for the Viz
display compositor, so a driver that CHECK-crashes at GPU init keeps killing
every launch even with the fallback applied. These are the pieces a real
remediation needs, added standalone with no callers yet:

- a bounded, additive tier ladder (disable-gpu -> +compositing/ANGLE swiftshader
  -> +in-process-gpu, which removes the crashing child entirely)
- a GPU/shader cache purge, so a cache written by the crashing driver is not
  replayed into the tier we just escalated to
- a summarizer for app.getGPUInfo('complete'); nothing in src/main called it, so
  every GPU crash bundle arrived with no vendor, device, or driver version

Co-authored-by: Orca <help@stably.ai>

2026-07-25 12:15:00 -07:00

..

cli

fix(runtime): report the effective ask timeout so a clamped wait isn't misreported (#10550)

2026-07-25 05:42:05 -07:00

main

feat(gpu): add the GPU fallback ladder, cache purge, and driver-identity modules

2026-07-25 12:15:00 -07:00

preload

feat(codex): surface a stalled config sync instead of failing silently (#10449)

2026-07-24 18:30:54 -07:00

relay

fix(agent-hooks): install OpenCode status plugin into the WSL guest so status works over WSL (#10328)

2026-07-25 02:30:59 -07:00

renderer

fix(tasks): keep repos with a pending remote-identity probe in the picker (#10527)

2026-07-25 04:04:19 -07:00

shared

fix(tasks): keep repos with a pending remote-identity probe in the picker (#10527)

2026-07-25 04:04:19 -07:00

types

…