fix(daemon): a full disk under cmux-tui offers Restart with the reason, keeps starting it, and a refused New Workspace says why (restore fuzz h2, h3) by lawrencecchen · Pull Request #30496 · manaflow-ai/cmux

test(daemon): red tests for restore fuzz h2 (full disk): Restart cmux-tui offered with the reason after a failed ensure, and the reconnect keeps starting the daemon with backoff

lawrencecchen

committed

4d96f95

test(app): red test for restore fuzz h3 (full disk): a person's refused New Workspace shows one 'did not apply: the disk is full' refusal; a scripted one gets the typed error and no HUD

lawrencecchen

committed

ddb642d

fix(daemon): a local cmux-tui whose start fails (a full disk) offers Restart with the reason, and the app keeps starting it until it runs

lawrencecchen

committed

249eede

fix(app): a person's New Workspace refused on a full disk shows one "did not apply: the disk is full"; a scripted run keeps the typed error

lawrencecchen

committed

7488cea

refactor(daemon): recovery wiring as DaemonService.installRecovery(); the notice is internal and holds its daemon

lawrencecchen

committed

56ce1c4

test(app): h3 red test checks the action ran and the create reached the daemon, and names the commands the stand-in saw

lawrencecchen

committed

a5edcf4

test(daemon): DaemonRecoveryNoticeTests makes its stand-in's bridge on the main actor (fleet compile, job b7b1b48e)

lawrencecchen

committed

d553f01

test(app): h3 red test binds the app's actions (AppActions.bind), as the app does at launch

lawrencecchen

committed

fff3cc7

test(daemon): DaemonRestartTests waits for Retry Now's own ensure; Restart already shows before it (h2: every failed start offers it)

lawrencecchen

committed

4c8a6c2

test: the h2/h3 tests assert the localized reason and the typed start_failure, not English text (REVIEW.md rule 1)

lawrencecchen

committed

da8798b

Navigation Menu

Repository navigation

fix(daemon): a full disk under cmux-tui offers Restart with the reason, keeps starting it, and a refused New Workspace says why (restore fuzz h2, h3) - #30496

fix(daemon): a full disk under cmux-tui offers Restart with the reason, keeps starting it, and a refused New Workspace says why (restore fuzz h2, h3)#30496
lawrencecchen merged 10 commits into
feat-cmux-nextfrom
fuzz-red/restore-h2h3

Commits on Nov 1, 2026

test(daemon): red tests for restore fuzz h2 (full disk): Restart cmux-tui offered with the reason after a failed ensure, and the reconnect keeps starting the daemon with backoff

test(app): red test for restore fuzz h3 (full disk): a person's refused New Workspace shows one 'did not apply: the disk is full' refusal; a scripted one gets the typed error and no HUD

fix(daemon): a local cmux-tui whose start fails (a full disk) offers Restart with the reason, and the app keeps starting it until it runs

fix(app): a person's New Workspace refused on a full disk shows one "did not apply: the disk is full"; a scripted run keeps the typed error

refactor(daemon): recovery wiring as DaemonService.installRecovery(); the notice is internal and holds its daemon

test(app): h3 red test checks the action ran and the create reached the daemon, and names the commands the stand-in saw

test(daemon): DaemonRecoveryNoticeTests makes its stand-in's bridge on the main actor (fleet compile, job b7b1b48e)

test(app): h3 red test binds the app's actions (AppActions.bind), as the app does at launch

test(daemon): DaemonRestartTests waits for Retry Now's own ensure; Restart already shows before it (h2: every failed start offers it)

test: the h2/h3 tests assert the localized reason and the typed start_failure, not English text (REVIEW.md rule 1)