Drive the fleet cycle off my-pipeline's declarative graph - #13
Merged
Conversation
fleet_cycle no longer hand-codes the 10-step order: it ticks mypipeline.plan.build_plan() and binds each graph `stage` to concrete argv via a RESOLVERS registry lifted out of the old _run_cycle. my-pipeline owns what runs and in what order; my-fleet owns the runtime binding it can't (accounts, per-repo fan-out, live gh queries, missing-clone skips) plus whether it can afford to tick (halt gate, account pool, backoff -- all unchanged). mypipeline sync is now a cycle stage (the ledger->issue handoff axis, wired into the cycle it used to sit beside). tester/changelogger now run grouped (all testers, then all changeloggers) rather than interleaved per repo, since they're independent per-repo nodes -- behaviour-equivalent. Adds my-pipeline to the CI sibling installs; its @main must carry build_plan before this repo's CI is green (land the my-pipeline PR first). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
lorenzoliuzzo
marked this pull request as ready for review
July 22, 2026 09:29
4 tasks
…coder swap Both this branch's DAG-wiring work and main's now-merged mycoder-dispatch swap (#15) added an install step + a sibling-repos comment to the same two lines of ci.yml/pyproject.toml. Combined both: CI now installs my-pipeline and my-coder; the comment lists all five siblings. No overlap in the actual source (fleet_cycle.py only imports DISPATCH_LEDGER/HALT_MARKER/ _critical_halt_issues from fleet_dispatch.py, none of which changed), so src/tests merged cleanly with no conflicts to resolve there.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
fleet_cycleno longer hand-codes the 10-step order. It ticksmypipeline.plan.build_plan()and binds each graphstageto concrete argv viaa
RESOLVERSregistry lifted out of the old_run_cycleblocks.binding it can't hold (accounts, per-repo fan-out, live
ghqueries,missing-clone skips) and whether it can afford to tick (halt gate, account
pool, backoff — all unchanged).
mypipeline syncis now a cycle stage: the ledger→issue handoff axis, wiredinto the cycle it used to sit beside (previously not run by the cycle at all).
Behaviour change (intentional)
tester/changelogger now run grouped (all testers, then all changeloggers) rather
than interleaved per repo — they're independent per-repo nodes, so
behaviour-equivalent.
Merge ordering
CI installs
my-pipeline @ main, so this PR'stestcheck stays red untilMyThingsLab/my-pipeline#3 merges. Merge that first.
Readiness
ruff check src testscleanpytestgreen (241 passed) — added: handoff stage runs, order follows the plan, unknown-stage skip is non-fatal; every existing gate/fan-out/skip/halt/loop test still passes🤖 Generated with Claude Code