Goal 0011 design section (research delivered; build gated on wt-lists decision) by alicoding · Pull Request #10 · alicoding/mill

alicoding · 2026-08-12T01:56:50Z

Records the Lists-maturation research verdicts into the goal file. The build is gated on an owner decision: the wt-lists worktree holds a near-complete parallel implementation — rebase-and-land vs rebuild-from-research is theirs to call.

🤖 Generated with Claude Code

https://claude.ai/code/session_01FYwojT8GdUbYSoggbvEFft

go-edlib adopted for fuzzy (with the industry-default-is-exact framing correction); fixed-by-construction output shape; rows without CreatedBy (§3.2.4's verdict); snapshot-per-execution reframed from correctness gap to audit evidence (DBOS checkpointing already guarantees replay stability — Temporal-precedent-confirmed); storage stays settings-store with a named SQLite trigger. Flags the gating owner decision: wt-lists' near-complete parallel implementation. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01FYwojT8GdUbYSoggbvEFft

… self-cancel Owner-mandated framework, researched to the converged standard and found to need NAMING, not building: Scrum 2020's DoD (a formal artifact commitment) + Kanban University's ready-for-delivery pull criteria are the industry vocabulary for gates this repo already runs — consolidated into ONE always-loaded rule (.claude/rules/delivery-discipline.md) instead of four documents. Estimation deliberately excluded (negotiates team capacity a solo+agent loop doesn't have); peer review already settled by ADR-0034. Build-health becomes a ritual, not a wonder: gh pr checks --watch after opening a goal PR; gh run list -b main -L 1 at goal pickup. Tech debt: same queue, same DoR/DoD, never a shadow register. Plus the real config bug the research found: cancel-in-progress applied to main's post-merge runs too, so two quick merges cancelled the earlier one's verification (observed live: merges #1/#10) — now conditional per GitHub's own documented pattern; PR runs still supersede, main runs always complete. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01FYwojT8GdUbYSoggbvEFft

… second pass, ls-lint node_modules ignore (BACKLOG 10/11) (#53) resizable-table.spec.ts (Standing #10, third recurrence post-hardening): structural fix, not another timeout bump. test.describe.configure({ mode: 'serial' }) makes the file's never-interleave requirement explicit; waitForStableBoundingBox polls for the handle's bounding box being IDENTICAL across two consecutive reads (a non-null box mid-reflow was always possible the old check couldn't see); the synthesized drag moved from one batched page.mouse.move(..., { steps: N }) call to discrete, individually-awaited moves (browsers coalesce rapid pointermove events within one CDP command -- real, documented browser behavior via PointerEvent.getCoalescedEvents()); two more expect.polls wait for the drag's actual DOM effect and the localStorage persist to land before the next step depends on them. A keyboard-resize alternative was checked and ruled out -- shared/ResizableTable.tsx has no keyboard path at all, building one would be a feature addition, not a test fix. Verified via 5 separate fresh playwright invocations (10/10) -- an in-process --repeat-each stress loop turned out to be a self-confounding methodology (accumulating same-worker degradation, not representative of a real CI run). golangci-lint second pass (Standing #11b, named future work in goal 0028): gocritic/prealloc/contextcheck/sqlclosecheck enabled. Triaged to zero findings on both build-tag variants -- 14 real findings fixed (1 gocritic assignOp, 13 prealloc slice-capacity hints), zero contextcheck/sqlclosecheck findings, no rule needed a scoped tweak or nolint suppression this pass. .ls-lint.yml (Standing #11c): root ignore: list gains node_modules, with a comment on the root-scoped-recursive-rule interaction that caused the original gap. Claude-Session: https://claude.ai/code/session_018pkViCNAuZp2vBv2K9AbUh Co-authored-by: Claude Fable 5 <noreply@anthropic.com>

… second pass, ls-lint node_modules ignore (BACKLOG 10/11) resizable-table.spec.ts (Standing #10, third recurrence post-hardening): structural fix, not another timeout bump. test.describe.configure({ mode: 'serial' }) makes the file's never-interleave requirement explicit; waitForStableBoundingBox polls for the handle's bounding box being IDENTICAL across two consecutive reads (a non-null box mid-reflow was always possible the old check couldn't see); the synthesized drag moved from one batched page.mouse.move(..., { steps: N }) call to discrete, individually-awaited moves (browsers coalesce rapid pointermove events within one CDP command -- real, documented browser behavior via PointerEvent.getCoalescedEvents()); two more expect.polls wait for the drag's actual DOM effect and the localStorage persist to land before the next step depends on them. A keyboard-resize alternative was checked and ruled out -- shared/ResizableTable.tsx has no keyboard path at all, building one would be a feature addition, not a test fix. Verified via 5 separate fresh playwright invocations (10/10) -- an in-process --repeat-each stress loop turned out to be a self-confounding methodology (accumulating same-worker degradation, not representative of a real CI run). golangci-lint second pass (Standing #11b, named future work in goal 0028): gocritic/prealloc/contextcheck/sqlclosecheck enabled. Triaged to zero findings on both build-tag variants -- 14 real findings fixed (1 gocritic assignOp, 13 prealloc slice-capacity hints), zero contextcheck/sqlclosecheck findings, no rule needed a scoped tweak or nolint suppression this pass. .ls-lint.yml (Standing #11c): root ignore: list gains node_modules, with a comment on the root-scoped-recursive-rule interaction that caused the original gap. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018pkViCNAuZp2vBv2K9AbUh

New capability: CompositionService.ExportWorkflow/ImportWorkflow (compositionservice_export.go) plus a real Composition UI affordance (Export IconButton per workflow row, an Import button + hidden file input near New workflow) -- not backend-only, per explicit direction that this needs to be visibly checkable in the running app, not just git history. Design decision the goal itself required in writing: exportedWorkflow deliberately omits the workflow's own ID and BuiltIn flag from the wire shape. ImportWorkflow always mints a brand-new workflow via the existing CreateWorkflow (same validation bar as a hand-composed one -- ResolveNodeDefaults, ValidateGraph, no import-specific leniency), matching ADR-0013's Duplicate precedent (copying an entity means a new identity, never resurrecting the old one) -- so importing the same file twice, or into a different Mill instance, can never silently collide with or overwrite an existing workflow. Node/Edge IDs inside the graph ARE preserved as-is (relative references, not global identity). Stable/deterministic by construction, not extra bookkeeping: every exported field is already-stored data, never regenerated on save (confirmed by reading idgen.go directly), and Go's encoding/json guarantees deterministic struct-field and sorted-map-key ordering. Two exports of an unchanged workflow are therefore byte-identical -- proven by a real test (TestExportWorkflow_IsDeterministic), not just asserted -- which is the concrete answer to n8n's own documented git-diff-noise problem researched earlier this session. Two files crossed the 500-line limit adding this (CompositionView.tsx, composition.spec.ts) -- split along real seams, not truncated: TestRunDialog extracted to its own file (self-contained, zero shared state with its parent beyond props, same pattern this repo already uses); the three new e2e tests moved to composition-export-import.spec.ts with their own workflowRow helper, matching the existing convention of 5 other spec files each keeping their own copy rather than a shared helpers file that doesn't otherwise exist here. Verified: 11 new Go tests (round-trip, new-ID-never-reused, two independent imports from the same file, determinism, ID/BuiltIn omitted from the wire shape, unknown-ID/invalid-JSON/empty-label/ invalid-graph-shape rejection, Attributes application) plus the full existing suite, all passing with -race. Full frontend check suite (tsc, eslint, boundaries, vitest) clean. Both e2e spec files run twice in a row per .claude/rules/testing.md's explicit discipline for tests creating persisted entities -- 17/17 both times, confirming cleanup actually works, not just once. Regenerated bindings also carry two small doc-comment-only updates trailing from task #3's Go comment edits (configureservice.ts, models.ts) -- no functional change. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01FYwojT8GdUbYSoggbvEFft

…y (task #12) Mill as MCP server for the first time -- §3.1 named three possible MCP roles (server, client, host) back when only client was ever built (§3.6, wrapping external tools as workflow nodes); this closes the server role, already locked as fine ("fits §1.1 cleanly with zero tension") since no LLM/agent loop runs inside Mill here -- an external agent's own host connects and reads, the same shape as httpconnector being an HTTP client. Does not touch or reopen the still-disputed MCP-host question. internal/adapters/mcpserving wraps modelcontextprotocol/go-sdk's server role (the same SDK Mill already depends on for the client role) behind a thin New()/Serve() pair -- new role, not a new dependency. millmcpservice.go registers two URIs per entity type: a plain index resource (mill://workflows, mill://requests, mill://lists, mill://mcpservers) listing every current ID/Label/Description, and a ResourceTemplate (mill://workflows/{id}, etc.) whose read returns the full definition by calling the SAME Export* methods tasks #10/#11 just built for the UI's own Export buttons -- one read-model, reused, not a second one built for this. Secrets stay excluded by the same construction Export* already guarantees, verified independently through this new code path rather than assumed inherited: a real test connects a genuine MCP client over real HTTP, sets a real secret on a real HTTPRequest, reads it back through mill://requests/{id}, and asserts the secret never appears in the wire response. Binds 127.0.0.1:8090 by default (MILL_MCP_ADDR overrides), loopback-only deliberately -- a new unauthenticated local listener, same conservative- default reasoning as the LAN-exposure question elsewhere in this repo. Runs in both desktop and server-mode builds (no build tag); confirmed the second listener doesn't conflict with server mode's existing :8080 via a real Playwright smoke run. A bind failure is logged, not fatal. Explicitly NOT built here, named directly rather than silently dropped: the write side (create/import via MCP Tools). §8's guardrail policy is still OPEN, and programmatic writes are a materially different risk than read-only exposure -- tracked as its own task (#13), not folded into this one. Verified: two new Go tests connect a genuine MCP client (the SDK's own client role, not a mock, not a direct call into the handler functions) over real HTTP -- resources/list returns the real registered set, resources/read against a just-created real workflow returns its real node data, an unknown ID returns a real resource-not-found error, and the secret-exclusion test above. Full existing suite passes with -race. go vet, golangci-lint (0 issues after fixing two real errcheck findings on deferred session.Close() -- fixed using this repo's own existing defer func() { _ = x.Close() }() convention, not suppressed), ls_lint, check-loc.sh all clean. Documented in docs/SPEC.md §3.6. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01FYwojT8GdUbYSoggbvEFft

… self-cancel Owner-mandated framework, researched to the converged standard and found to need NAMING, not building: Scrum 2020's DoD (a formal artifact commitment) + Kanban University's ready-for-delivery pull criteria are the industry vocabulary for gates this repo already runs — consolidated into ONE always-loaded rule (.claude/rules/delivery-discipline.md) instead of four documents. Estimation deliberately excluded (negotiates team capacity a solo+agent loop doesn't have); peer review already settled by ADR-0034. Build-health becomes a ritual, not a wonder: gh pr checks --watch after opening a goal PR; gh run list -b main -L 1 at goal pickup. Tech debt: same queue, same DoR/DoD, never a shadow register. Plus the real config bug the research found: cancel-in-progress applied to main's post-merge runs too, so two quick merges cancelled the earlier one's verification (observed live: merges #1/#10) — now conditional per GitHub's own documented pattern; PR runs still supersede, main runs always complete. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01FYwojT8GdUbYSoggbvEFft

… second pass, ls-lint node_modules ignore (BACKLOG 10/11) (#53) resizable-table.spec.ts (Standing #10, third recurrence post-hardening): structural fix, not another timeout bump. test.describe.configure({ mode: 'serial' }) makes the file's never-interleave requirement explicit; waitForStableBoundingBox polls for the handle's bounding box being IDENTICAL across two consecutive reads (a non-null box mid-reflow was always possible the old check couldn't see); the synthesized drag moved from one batched page.mouse.move(..., { steps: N }) call to discrete, individually-awaited moves (browsers coalesce rapid pointermove events within one CDP command -- real, documented browser behavior via PointerEvent.getCoalescedEvents()); two more expect.polls wait for the drag's actual DOM effect and the localStorage persist to land before the next step depends on them. A keyboard-resize alternative was checked and ruled out -- shared/ResizableTable.tsx has no keyboard path at all, building one would be a feature addition, not a test fix. Verified via 5 separate fresh playwright invocations (10/10) -- an in-process --repeat-each stress loop turned out to be a self-confounding methodology (accumulating same-worker degradation, not representative of a real CI run). golangci-lint second pass (Standing #11b, named future work in goal 0028): gocritic/prealloc/contextcheck/sqlclosecheck enabled. Triaged to zero findings on both build-tag variants -- 14 real findings fixed (1 gocritic assignOp, 13 prealloc slice-capacity hints), zero contextcheck/sqlclosecheck findings, no rule needed a scoped tweak or nolint suppression this pass. .ls-lint.yml (Standing #11c): root ignore: list gains node_modules, with a comment on the root-scoped-recursive-rule interaction that caused the original gap. Claude-Session: https://claude.ai/code/session_018pkViCNAuZp2vBv2K9AbUh Co-authored-by: Claude Fable 5 <noreply@anthropic.com>

alicoding enabled auto-merge August 12, 2026 01:56

alicoding merged commit f86bc27 into main Aug 12, 2026
13 checks passed

alicoding deleted the docs/0011-design branch August 12, 2026 17:44

alicoding mentioned this pull request Aug 13, 2026

chore: hygiene close-out -- resizable-table proper fix, golangci-lint second pass, ls-lint node_modules ignore (BACKLOG 10/11) #53

Merged

5 tasks

Navigation Menu

Goal 0011 design section (research delivered; build gated on wt-lists decision) - #10

Goal 0011 design section (research delivered; build gated on wt-lists decision)#10
alicoding merged 1 commit into
mainfrom
docs/0011-design

alicoding commented Aug 12, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Conversation

alicoding commented Aug 12, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant