Skip to content

Restore workflow on main - #311

Merged
JonJagger merged 1 commit into
mainfrom
restore-workflow-on-main
Sep 5, 2026
Merged

JonJagger merged 1 commit into
mainfrom
restore-workflow-on-main

Conversation

@JonJagger

Copy link
Copy Markdown
Member

No description provided.

@JonJagger
JonJagger merged commit 7061e4d into main Sep 5, 2026
10 checks passed
@JonJagger
JonJagger deleted the restore-workflow-on-main branch September 5, 2026 17:31
JonJagger added a commit that referenced this pull request Sep 5, 2026
  This branch is to become main, and the two have been redoing the same
  work since #299 split them. main reached the same place by a different
  route in #300: NodeImages, the lib/ move, the monotonic clock. Every
  conflict here came from that one commit, and the overlap only grows the
  longer the merge is put off, so it happens now rather than at the end of
  the plan in docs/pre-started-container-pool.md.

  What main has that this branch does not is the part worth having: the
  start-point test kit (#307-#311), the deploy-by-hand workflows
  (#302-#304), the base-image bump (#305), the deadline and its readings on
  one clock (#301), and the container stderr fix (#306).

  Where the two designs collide this branch wins, because a test-run here
  execs into a container that already exists, and main still creates one
  per run. So image_config plus exec_config stay, create_config goes, and
  the daemon keeps containers_named, rename_container, create_exec and
  start_exec.

  Two of main's fixes land inside files this branch rewrote, and neither
  survives taking a side wholesale. #301's clock is kept: read_payload
  builds its DeadlineReader with max_seconds and a clock, and the clock
  delegator it needs is kept with it. #306's stderr test is kept as c9Gf27,
  both sides having called their newest test c9Gf21.

  Two merges that git resolved without a conflict were wrong, both by
  splicing one side's body onto the other's declaration. DockerDaemonStub
  got a start_container reading @start_code and @start_body that its
  initialize never sets. CyberDojoShRunner got a second docker delegator.
  The delegators now sit together at the end of the class.

  c9Gf10 asserted a :ran_the_kata phase that PHASES no longer names, which
  is main's name for a run that does its work in the container's own Cmd.

  One failure was older than this merge. 340be96 lowered fsize from 256MB
  to 16MB and left 3A8D99 expecting the old value, so the client suite has
  been red on all three OSs since then. The test now expects what the
  config sets.
Merge branch 'main' into allowlist-the-spare-pool
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant