Three bundled fixes to slice 7's port flow:
1. Port-pointerdown branches deterministically:
- cable-draw in progress → finish / cancel
- no tool, no draw → select port (inspector opens)
- cable tool → start a draw from this port
- any other tool armed → bubble (so +Port can place a new port even
when the click lands on top of an existing one)
2. Port circles now fill *and* stroke with the cable_type colour so the
port reads as obviously coloured against the device rect. Selection
adds a drop-shadow halo.
3. Port inspector — clicking a port (no other tool armed) selects it
and shows a panel with cable-type swatch, label input, edge selector
(Top / Right / Bottom / Left), and Delete. Changing the edge PATCHes
x_offset / y_offset to the centre of the chosen side.
snapToDeviceEdge already picks the nearest of the four edges, so
placement on +Port lands correctly without further changes.
Adds 5 built-in device_types (project_id NULL, built_in=1):
- Multi-plug 3/4/5/6 (kind=hub, 🔌) — Power × N+1 (1 in + N out)
- Wifi-plug (kind=accessory, 📶) — Power × 2 pass-through outlet
The solver treats every Power port identically regardless of in/out
direction; m knows which end is which from the physical setup.
Tests:
- TestSeed_BuiltInDeviceTypes: built-in count rises from 16 → 21.
- TestSeed_PortProfiles: new entries' port totals.
- TestSeed_PowerCatalog (new, table-driven): asserts kind, icon, and
the single Power port row for each of the 5 new types.
mxdrw expects HTTP Basic Auth (BASIC_AUTH_USER + BASIC_AUTH_PASS on the
server side). Replace MEXDRAW_TOKEN with MEXDRAW_USER + MEXDRAW_PASS,
use req.SetBasicAuth on the export PUT.
Updated docker-compose.yml comment and README env table to match.
Roundtrip verified locally against mxdrw.msbls.de.
picasso shipped (2 commits): internal/exporter pure BuildScene +
Generate21 (crypto/rand base62 IDs), internal/db/excalidraw_ids.go
idempotent persistence, internal/server/export.go POST handler with
bearer auth + 10s timeout, frontend Export button + toast.
6 new exporter tests + 60+ existing all green with -race. Hand-test
roundtrip vs mxdrw confirmed: 20 elements per spec, IDs stable across
re-exports.
Deploy to mDock blocked on MEXDRAW_TOKEN — picasso correctly refused
to fake the secret. m to drop value into /home/m/secrets/mcables/.env
on mdock, then redeploy.
Export button is no longer disabled. On click it POSTs to the export
endpoint and shows a toast next to the button:
✓ Exported · open in mxdrw (with viewer URL)
✗ Export failed — <detail>
apply-template now auto-solves by default (?solve=0 opt-out for power
users) and returns combined {template_apply, solve} response.
Frontend reloads via activateProject() after Apply, so devices +
cables render immediately without manual Solve click.
Verified: TEST-AUTO project + Living Room template → 3 devices +
2 HDMI cables visible in one round-trip.
Two changes to close the UX hole m hit on slice 6 — Apply Template
appeared to do nothing because (a) the canvas wasn't refreshed cleanly
and (b) the cables hadn't been computed yet.
Backend (internal/server/solver.go applyTemplate handler):
- After ApplyTemplate succeeds, run Solve(false) inside the same
request. Combined response shape:
{ template_apply: <ApplyTemplateResult>, solve: <SolveResult> }
- Opt out with ?solve=0 for power-users who want to inspect the
seeded devices/requirements before the solver runs. Response in that
case is { template_apply: ... } only.
- If Solve fails after a successful apply, return
{ template_apply, solve_error: "..." } so the frontend can recover
(devices are still there; m can hit Solve manually).
Frontend (web/static/main.js apply-template modal submit):
- Replaced the bare re-snapshot with a call to activateProject(pid).
That's the canonical project-load path — it re-hydrates ALL
collections (frames, devices, ports, io_markers, cables, bundles,
requirements, cable_types, device_types), clears state.selection
so a stale pre-apply selection can't linger, and routes through the
same render() the URL-state hydration uses on initial page load.
- The slice-6 inlined re-snapshot missed the device_types refresh +
selection reset, which I suspect was what made the canvas look
stuck — render()ing with state.selection.kind="cable_type" or
"requirement" pointing at a not-yet-loaded row.
Hand-test (local): Living Room + auto-solve produces 4 devices + 3
requirements + 3 cables; ?solve=0 leaves cables empty. Snapshot
includes the cables on auto-solve path.
+Port (device inspector):
- New button on the device inspector arms a port-placement tool with
the device + currently-active cable type pre-selected.
- Click anywhere on the canvas: snapToDeviceEdge() finds the closest
edge of the selected device, clamps the perpendicular coord, POSTs a
new port. The new port renders immediately (state.ports.push +
render()).
- Per-port × delete button in the inspector ports grid.
Manual cable draw:
- Port circles are now clickable (slice 4 had pointer-events:none).
- Click a port → starts a cable draw with that port as the source
(state.cableDrawFromPortID, port highlighted via .cable-from class).
- Click another port → POSTs a cable with from_port_id + to_port_id,
type derived from source port, auto=false. If the target port's type
differs, confirm-prompt warns m before committing.
- Shift+click target port → binds to the target's parent device
(to_device_id) instead of the port.
- Click an IO marker mid-draw → terminates the cable with to_io_id.
- Esc cancels the draw + clears state.cableDrawFromPortID.
- "Draw cable" toolbar button is now enabled (data-tool=cable, keyboard
is implicit via port-click). armTool() teardown clears the source-port
state.
Cable inspector tweak (slice 6 callback):
- "driver" row now renders as a clickable button showing the
requirement's "FromName ↔ ToName" instead of the raw id; click jumps
the inspector to that requirement.
CSS:
- tool-port + tool-cable add the same crosshair cursor as the other
tools (descendant-targeted with !important to beat svg-draggable's
grab cursor — same fix-pattern as slice 3's cursor-cache pass).
- .port-circle.cable-from gives the source port a glow.
- .btn-link styles for inspector inline buttons.
The schema has ON DELETE SET NULL on cables.from_port_id /
cables.to_port_id, but the cables CHECK constraint requires exactly one
of (port/device/io) to be non-null per side. Setting both refs to NULL
on a port-delete violates the CHECK, blowing up the DELETE with a 500.
DeletePort now opens a tx, deletes any cable that referenced the port
on either side, then deletes the port. Same observable effect from m's
POV: cables that point at a deleted port are gone (he can re-draw with
the manual cable tool if he still wants them).
New store methods on internal/db/ports.go:
- CreatePort / GetPort / UpdatePort / DeletePort (all project-scoped)
- ListPortsForDevice for the inspector's per-device list
New handlers (internal/server/ports.go):
- GET /api/projects/:pid/devices/:id/ports
- POST /api/projects/:pid/devices/:id/ports ← {type_id, label?, x_offset, y_offset}
- PATCH /api/projects/:pid/ports/:id ← partial
- DELETE /api/projects/:pid/ports/:id (cables ref → ON DELETE SET NULL)
Lets slice 7's +Port tool add/remove instance ports without going
through the type-seeded auto-creation path from slice 4.
Header gains a Solve button (keyboard S) + Apply template button.
Canvas:
- Cables render as straight lines port→port (or device-centre when the
endpoint is a whole device, or io-marker centre). Auto-cables get a
dashed stroke; manual cables (auto=0) solid. Stroke colour = cable_type.
- Click a cable to select it → inspector pane updates.
Solve preview-diff modal:
- Calls POST .../solve?preview=1 on open.
- Renders cables_added, cables_removed, bundles_added in colour-coded
lists. Unsatisfied entries get a class="unmet" badge + one-click
quick-fix:
* "no free <type> port" → "+ Add <type> port to <device> and re-solve"
fires POST .../devices/:id/ports-and-resolve in one round-trip and
re-renders the preview.
* "ambiguous cable type" → "Specify cable type…" re-opens the
requirement modal.
* "no compatible cable type" with a preferred type → "+ Add port…"
quick-fix on the from-side device.
- Apply → POST .../solve (no preview) → re-snapshot to pick up new
cable ids + bundle assignments.
Cable inspector (kind=cable):
- Shows type, from-endpoint, to-endpoint labels.
- For solver-owned cables, shows the driving requirement (best-effort
match by unordered device pair + type) and a "Promote to manual"
button (PATCH with `promote: true` flips auto→0).
- Delete button on both auto and manual cables.
Apply-template flow:
- "Apply template…" header button opens a wide modal with a template
dropdown (Living Room / Home Office / Server Rack) + a preview panel
showing each device row (skip checkbox + editable name input) and
the template's requirements.
- Submit → POST .../apply-template with name_overrides + skip_devices,
then re-snapshot.
State + snapshot:
- state.cables, state.bundles, state.setupTemplates added.
- activateProject pulls them from the snapshot; teardown on switch.
Migration 004:
- setup_templates + setup_template_devices + setup_template_requirements
- 3 built-in templates seeded: Living Room (TV+Soundbar+ChromeCast,
2× HDMI), Home Office (PC+Screen+Keyboard+Mouse, 1× HDMI + 2× USB),
Server Rack (NAS+Switch+fritz, 2× RJ45).
Cables store (cables.go):
- CRUD with endpoint validation (port|device|io exactly-one, project-
scoped). Tx-aware: validateEndpointEx + assertCableTypeEx avoid
deadlocks when the solver Apply tx holds the MaxOpenConns(1) connection.
Bundles store (bundles.go):
- CRUD with cable_ids replacement on PATCH. createBundle(ex, …, ownTx)
inherits the caller's tx for solver-internal use; returns a locally-
constructed Bundle when ownTx=false (re-fetching via s.db would
deadlock).
Solver (solver.go) implements design v4.1 §5b.2 exactly:
- Pre-fetch devices/ports/cables/requirements/bundles.
- Reserve ports used by manual cables (auto=0) so the solver can't
reuse them.
- For each requirement (must_connect DESC, id ASC):
* Resolve cable type: preferred, or T = port-types(from) ∩
port-types(to). |T|==0 → unsatisfied "no compat type"; |T|>1 →
"ambiguous"; |T|==1 → that one.
* Pick lowest-id free port on each side. None → unsatisfied with
WhichSide hint + cable-type name.
- Endpoint-pair bundle: ≥2 staged cables between the same device pair
→ auto bundle.
- Diff against existing auto cables by (type_id, MIN(from,to), MAX(from,to))
signature. Matched = kept; new = added; orphans = removed.
- Preview returns the diff without writing; Apply runs in a single tx
that wipes auto bundles, deletes orphan auto cables, inserts new
ones, and rebuilds bundles.
- PortsAndResolve: combo helper for the inspector quick-fix —
inserts a port + re-runs Solve.
Setup-templates store (setup_templates.go):
- List/Get with hydrated devices + requirements.
- ApplyTemplate(projectID, templateID, opts) seeds devices + requirements
in one tx. Per-device name overrides + opt-out. Name collisions skip
the device (skipped_devices); requirements whose endpoints both fail
are also skipped (requirements_skipped). UNIQUE-collision on an
existing requirement is non-fatal; logged in requirements_skipped.
Snapshot: cables + bundles fields tightened to []Cable / []Bundle and
populated from the store.
11 new tests (solver_test.go), all green with -race:
- Basic NAS↔Switch (RJ45) → 1 cable, auto=true
- Ambiguous cable type → unsatisfied
- No free port → unsatisfied with side hint
- Preview doesn't write
- Apply then re-apply → idempotent (kept=N, added=0)
- Manual cable reserves its port → solver can't claim it
- ApplyTemplate Living Room → 3 devices + 2 requirements + 7 ports
(from the device-type port seeder)
- Home Office template then Solve → 3 cables, 0 unsatisfied
- Name-collision pre-existing device → skipped + req-pair skipped
Snapshot now carries connection_requirements; state.requirements is
populated on project switch.
Sidebar:
- New "Requirements" section between Cable types and Tools.
- Each row shows "A ↔ B · cable-type" plus a must/nice badge. Clicking
a row selects the requirement (inspector pane updates).
+ Requirement modal:
- Device-pair pickers (autocompletes from the project's current devices).
- Cable-type picker with "— solver picks —" as the first option (saves
preferred_cable_type_id as null on the wire).
- "Must connect" checkbox (default on); notes textarea.
- POSTs to /api/projects/:pid/connection-requirements. 409 collisions
(reversed-pair duplicates) surface as inline form errors.
Drag-from-A-to-B gesture:
- New tool `req` (keyboard R + "Drag req A→B" button). Arming the tool
+ pointerdown on a device starts a dashed-line preview. Pointerup on
another device opens the modal with from/to pre-filled. Anywhere
else cancels. Crosshair cursor while armed.
Inspector:
- Device pane gains a "Requirements" section listing every requirement
involving the selected device, sorted by the other device's name.
Each row is clickable → inspector jumps to that requirement.
- New `requirement` selection kind with its own inspector renderer
showing from/to, cable type, must/nice toggle button, debounced
notes textarea, "Edit" (re-opens modal), and Delete.
Delete of a device cleans up its requirements in local state (server
already CASCADEs the rows).
Migration 003 adds the solver's per-project input table + the auto flag
that slice 6 will use to distinguish solver-owned cables from m's
hand-drawn ones.
connection_requirements:
- (from_device_id, to_device_id, preferred_cable_type_id) with
preferred_cable_type_id nullable ("solver picks if exactly one type
matches both ends").
- (pair_lo, pair_hi) is the order-normalised MIN/MAX of (from, to),
stored alongside the m-facing from/to so the UI doesn't have to
denormalise.
- UNIQUE (project_id, pair_lo, pair_hi, preferred_cable_type_id) →
(A,B,T) and (B,A,T) collide; (A,B,Power) + (A,B,RJ45) coexist.
- CHECK (from != to). FK CASCADE from devices → requirement vanishes
if either endpoint device is deleted.
Store + 11 new tests:
- pair normalisation rejects the reversed-direction duplicate
- different cable types on the same pair coexist
- self-loop rejected (ErrInvalidInput)
- cross-project device reference rejected
- two null-cable-type reqs on the same pair both succeed (SQLite NULL
!= NULL in UNIQUE — semantically "solver picks both times", second
wins)
- partial PATCH: preferred_cable_type_id tri-state (leave/set/clear),
must_connect bool, notes string
- device delete cascades to its requirements
- snapshot.connection_requirements is non-nil and populated
cables.auto:
- ALTER TABLE cables ADD COLUMN auto INTEGER NOT NULL DEFAULT 0 CHECK
(auto IN (0,1)). Slice 6 sets 1 from the solver; slice 7's manual
cable POST keeps the default 0.
picasso shipped (4 commits @ 7f0b6e4):
- migration 002: device_types + device_type_ports + devices.type_id,
seeded with 16 built-ins (NAS PC Mac Notebook TV Soundbar Switch
fritz ChromeCast SteamLink IOx-3/6/8 Screen Keyboard Mouse)
- store: type-aware device POST seeds ports transactionally with
even-spread layout along the configured edge
- handlers: /api/device-types (built-ins) + /api/projects/:pid/device-types
(merged with project-custom), 403 on built-in mutations
- frontend: +Dev becomes a type-dropdown grouped by kind + name input
pre-fill, port rendering as SVG circles colour-stroked by cable type
Modal-driven +Dev (replaces the v3 inline namer):
- Tool armed → click on canvas captures the click position + frame_id
from frameAt(p), then opens a #modal-new-device dialog.
- Dialog has a <select> grouped by `kind` for built-ins, then
project-custom rows, then "Custom (no type)" at the bottom.
- Default selection is the first built-in (NAS). Name input is
auto-pre-filled to <type-name>, bumping to <type-name>-N if a name
collision is detected in the current device list.
- Submit POSTs name + type_id + x/y/w/h + frame_id. Server seeds the
ports in the same transaction; we re-snapshot to pick them up.
Canvas:
- After each device's <rect> + label, render the device's ports as
white-filled <circle>s with stroke = the port's cable_type colour.
- Position: (device.x + port.x_offset, device.y + port.y_offset). The
seeder's "evenly along the edge" layout means ports already sit on
the device's bottom edge by default and follow the device on drag
(because they re-render from the same x/y on every renderCanvas).
- Ports themselves are `pointer-events: none` for slice 4 — selection
remains device-level. Per-port click semantics ship in slice 7
(manual cable draw).
Inspector device pane:
- New "type" row showing the type name + a "(custom)" badge for
project-custom types, or "Custom (no type)" for freeform.
- New "Ports" section with one row per seeded port: cable-type-colour
swatch, label, "unconnected" placeholder. Label falls back to the
cable type's name when the seeded label_prefix was blank.
State + snapshot:
- state.ports populated from snap.ports; cleared on project switch /
404.
- state.deviceTypes hydrated from GET /api/projects/:pid/device-types
after the snapshot loads. Failure of that fetch is non-fatal — the
+Dev modal just shows "Custom (no type)" only.
- Delete-device cleans up its ports from state.ports too (server-side
CASCADE already handles persistence).
- GET /api/device-types — built-ins only (read-only).
- GET /api/projects/:pid/device-types — built-ins + project-custom merged.
- POST/PATCH/DELETE /api/projects/:pid/device-types — project-custom only.
Mutating a built-in row returns 403 via the new ErrForbidden → 403 map
in writeError.
- devicePatch / deviceCreate JSON shapes accept type_id (tri-state for
PATCH via the existing parseFrameRef helper applied to type_id too).
- POST /api/projects/:pid/devices with type_id seeds ports in one tx
server-side; response carries the device row + the snapshot will then
carry the new ports.
Catalog: 11 built-ins from §2.2 + the v4.1 trio (Screen, Keyboard, Mouse)
seeded in migration 002, totalling 16 built-in types.
Store layer:
- internal/db/device_types.go — CRUD for device_types. Built-ins
(project_id NULL) reject PATCH/DELETE with new ErrForbidden sentinel
(handler maps to HTTP 403). Project-custom types accept full CRUD;
cross-project access returns ErrNotFound. Replacing the port profile
on UPDATE is one transaction.
- internal/db/ports.go — ListPortsForProject for the snapshot loader +
seedPortsFromType(tx, …) used by CreateDevice. Layout is "evenly spaced
along the configured edge", per-edge group ordering by sort_order +
id. Labels are "<prefix>" for count==1 and "<prefix> N" 1-indexed for
count>1.
- Device gains a nullable TypeID + tri-state on UpdateDevice. CreateDevice
validates the type is built-in or a project-custom row of the same
project, then seeds the device's ports in the same transaction.
Snapshot now populates Ports from the store; field type tightened to
[]Port.
Tests (15 new, all green with -race):
- 16 built-ins seeded with correct names + project_id=NULL + built_in=1
- Port-profile totals match the §2.2 table for every built-in type
- Project-custom create + name-collision-with-built-in → 409 (ErrConflict)
- Per-project name UNIQUE — same custom name across projects is fine
- PATCH/DELETE built-in → ErrForbidden
- Cross-project custom PATCH → ErrNotFound
- CreateDevice with NAS type → 2 ports along bottom edge, evenly spaced,
labels set
- CreateDevice with PC type → 5 ports incl. "USB 1" + "USB 2"
- CreateDevice without type_id → 0 ports (freeform fallback)
- Cross-project custom type on CreateDevice → ErrInvalidInput
- Snapshot includes the seeded ports
Slice 3 frontend.
+ IO tool (keyboard `I`):
- Single-click on canvas places a 30x30 diamond (rotated <rect>) at the
point, with the Power-cable_type colour fill (red-ish).
- Inline namer prompts for a label; empty → server defaults to "IO".
- Drop-point determines initial frame_id via the existing frameAt()
point-in-rect logic, same as devices.
Render:
- io_markers come from snap.io_markers in the snapshot loader. Each
renders as a <rect> with rotate(45) around its centre + a small text
label below the diamond. Selection halo on stroke-width.
- Drag is the same pointer-event flow as devices; on pointerup, PATCH
x,y + recompute frame_id from the new centre. Cross-frame moves
update frame_id with explicit null on the wire when leaving all frames.
- Frame-drag now also relocates contained IO markers (mirrors the
device-cascade pattern). Single PATCH per IO marker on release.
Cable-type inspector:
- Clicking a legend row now sets state.selection = {kind:"cable_type", id}
in addition to toggling activeTypeId. The inspector renders the cable
type's details (name + colour, both editable, with the
"shared across projects" banner from v3 §7), a used-by counter (0
until slice 7 ships cables), and a Delete button that surfaces the
RESTRICT in_use_by_cables count from the server.
- Debounced rename via the existing bindDebouncedRename helper.
Inspector frame view picks up an "IO" count alongside the device count.
Background click + Esc clear the selection (existing behaviour, now
covers cable_type too).
Hand-tested via the API equivalents: 3 IO markers created (free, in
frame, default-label), PATCH x,y + frame_id-to-null all work, cross-
project frame_id rejected with 400, DELETE 9999 returns 404. Snapshot
shape post-slice-3: {frames, devices, io_markers, cable_types} all
populated, ports/cables/bundles still [].
Schema already in 001_init.sql; this is just the Go store layer.
IO markers are project-scoped wall-outlet terminators (a cable's
"this end plugs into a wall socket outside the diagram" endpoint).
Power-by-convention; no schema-level type enforcement.
- CreateIOMarker validates frame_id is in the same project (cross-project
ref → ErrInvalidInput), defaults label to "IO" when blank.
- GetIOMarker is project-scoped — wrong-project read returns ErrNotFound.
- UpdateIOMarker uses the FrameRef tri-state for frame_id (same as
DeviceUpdate) so callers can clear it explicitly.
- DeleteIOMarker is direct delete — ON DELETE SET NULL from the schema
drops the io_markers.frame_id ref cleanly when the frame is deleted
(verified by TestDeleteFrame_SetsIOMarkerFrameIDToNull).
Snapshot now populates IOMarkers from the store; field type tightened
from []any to []IOMarker.
7 new table-driven tests, all green with -race.
m's review of v4 locked 6 answers. Tight doc pass:
- Schematic-only bundling: dropped trunk-segment/frame-edge/cable-tray
language. v3 endpoint-pair rule is the only bundle rule.
- Setup templates folded in (not post-MVP): migration 004 with 3
built-ins (Living Room, Home Office, Server Rack) + 3 new device
types (Screen, Keyboard, Mouse) + apply-template API in slice 6.
- Unmet-requirement quick-fix: combo endpoint adds a missing port and
re-solves in one server roundtrip.
- Solver still button-only, catalog still SQL-seeded, promote still
explicit on cable inspector.
All 9 §9 questions resolved.
Tight pass on m's review of v4 (single commit per head's instruction).
Six locked answers integrated:
1. mCables is a schematic, not a physical-routing tool. Stripped
'trunk', 'frame-edge corridor', 'cable tray', 'path optimisation'
from §5b.1, §5b.2, §7, §8, §9. Bundling reduces to the v3 endpoint-
pair rule: ≥2 cables between the same A↔B endpoint pair → group as
one bundle. Anything path-shaped is "out of scope, period" (§8).
2. Solver button-only for v0 (no change). Live-solve parked at 9+.
3. Unmet-requirement quick-fix: red badge on the affected device in the
inspector with a single "+ Add <type> port to <device> and re-solve"
button per §5b.4. New endpoint
POST /api/projects/:pid/devices/:id/ports-and-resolve chains the
port insert + the solve re-run in one transaction.
4. Setup templates fold INTO v4.1. New §2.4 with the schema for
setup_templates + setup_template_devices + setup_template_requirements
(migration 004), 3 built-in templates seeded (Living Room, Home
Office, Server Rack). New API: GET /api/setup-templates,
POST /api/projects/:pid/apply-template. New UI flow: "or start from
a template" section in the New Project modal + an "Apply template"
action on empty projects. Built-in catalog grows to 14 types
(adds Screen, Keyboard, Mouse).
5. Catalog SQL seed in migration 002 (no change).
6. Promote-to-manual: explicit button on cable inspector (no change).
§8 slice 6 absorbs the templates work alongside the solver MVP.
§9 closes all six v4 questions; no open design questions remain.
Trailer changes to "DESIGN v4.1 READY FOR REVIEW".
CLAUDE.md mirrors: schematic-only framing, 14-type catalog, setup
templates as a first-class feature, quick-fix UX note.
Big rescope driven by m's product-vision clarification: mCables is a
cable-management framework with a solver as its core value prop, not a
manual draw-and-click editor. m declares devices + required connections
between them; the solver emits the cable plan + bundle recommendations,
optimising for maximum bundling.
Schema additions (migrations 002 + 003):
- device_types (catalog) — built-ins (project_id NULL) + project-custom
(project_id non-null). 11 built-in types seeded with default port
profiles (NAS, PC, Mac, TV, Soundbar, Switch, fritz, ChromeCast,
SteamLink, IOx-3/6/8, Notebook).
- device_type_ports (profile rows: cable_type × count × edge).
- devices.type_id (nullable). Picking a type seeds ports once;
instance-owned thereafter (no retroactive re-seed).
- connection_requirements (per-project, from/to device + preferred type
+ must_connect flag, with order-normalised pair_lo/pair_hi for
duplicate prevention).
- cables.auto (slice 5.5 migration) — distinguishes solver-owned cables
from user-drawn ones.
API additions:
- GET /api/device-types (built-ins only, read-only) and
GET /api/projects/:pid/device-types (built-ins + project-custom merged)
- POST/PATCH/DELETE under /api/projects/:pid/device-types (project-custom
only; built-ins are 403)
- /api/projects/:pid/connection-requirements full CRUD
- POST /api/projects/:pid/solve with ?preview=1 — pure-function solver
(greedy port allocation, endpoint-pair bundling for v0); returns
add[], remove[], bundles_added[], unsatisfied[], warnings[]
Solver algorithm (§5b):
- Read project devices + ports + connection_requirements + manual cables
- Assign each requirement a (port_a, port_b) using the preferred cable
type (or auto-pick if exactly one type matches both ends)
- Bundle by endpoint-pair (v3 rule, applied to auto cables only)
- Surface unsatisfied requirements per class (no compat type / ambiguous
type / no free port) — does NOT auto-add ports; UI quick-fix instead
- ?preview=1 returns the diff without writing; default applies in a tx
UI additions:
- Device-create modal: type dropdown (built-ins grouped by kind, then
project-custom, then "Custom (no type)" for the v3 freeform fallback)
- Left-sidebar Requirements section with + Requirement button
- Header Solve button (S keybinding) → preview modal → Apply
- Inspector for selected device: type, ports grid, unmet requirements
with red badges + quick-fix actions
- Inspector for selected auto cable: driving requirement, parent bundle,
Promote-to-manual button
Slice reshape (§8):
- Slices 1, 2 shipped. v4 inserts: 4 = catalog + type-aware device create,
4.5 = catalog management, 5 = requirements CRUD + UI, 6 = solver MVP +
Solve button. Old "manual port + manual cable draw" slides to slice 7
as a tweak path on solver output. Export becomes slice 8.
Six new open questions (§9) for m to gate before slice 4:
1. Path source (auto-route through frame edges / user cable-trays /
Steiner-tree)?
2. Live-solve vs. button-only?
3. UX when solver has no compatible port pair?
4. Setup templates in v4 or post-MVP?
5. Catalog as code seed or JSON file?
6. Auto-promote vs. explicit Promote-to-manual on solver cable edits?
CLAUDE.md updated to reflect the solver-core framing, hybrid catalog,
connection-requirements model, and auto/manual cable distinction.
Trailer changes to "DESIGN v4 READY FOR REVIEW".
startDrag set state.selection but never re-rendered. One render() call
after the assignment fixes it. Now selecting a device or frame
populates the inspector with name/dims/delete-button as designed.
startDrag set state.selection but didn't render until pointerup's onUp
ran — and onUp can throw on `e.currentTarget.classList.remove` if the
event reference is stale after pointer capture release, which leaves
the inspector stuck on 'Nothing selected.'
One-line fix: call render() right after state.selection assignment so
the inspector + halo update from pointerdown, independent of whether
onUp completes cleanly. The drag-completion render at the end of onUp
stays — when both fire it's idempotent (renders are pure functions of
state).
Primary fix: e.preventDefault() on the pointerdown for both armed-tool
branches in onCanvasPointerDown. Without it, the browser's default
mousedown action blurs the freshly-focused input in promptInline
(the SVG click target isn't focusable), and the blur handler calls
done(null) before m can type.
Secondary fix: clear activeNamer before fo.remove() in done(), to
prevent a re-entrant pageerror when Enter triggers blur synchronously.
Root cause traced by sherlock with Playwright (docs/sherlock-+dev-bug.md
on the sherlock branch). The previous routing fix at 94869f3 was
necessary but not sufficient: placeDeviceAt() now reaches promptInline()
correctly, but the synchronous input.focus() is undone ~6ms later by
the browser's compatibility-mousedown default — which blurs the active
element when the mousedown landed on a non-focusable target (SVG rect /
SVG root). The blur listener then ran done(null) and ripped the
<foreignObject> out before m could type a name.
Primary fix: e.preventDefault() at the top of both armed-tool branches
in onCanvasPointerDown. Suppresses the focus-shifting default so the
input keeps focus.
+Frame is also wrapped for symmetry. It wasn't strictly affected (its
namer runs from pointerup, not pointerdown) but preventDefault avoids
a subtle text-selection side effect during rubber-band drag.
Secondary fix in promptInline.done(): clear activeNamer *before*
fo.remove(). Enter-key triggers a synchronous blur listener which
re-enters done() — if remove() ran first, the re-entry hit a
"node no longer a child" pageerror. Reordering makes the re-entry a
no-op (activeNamer is already null).
Verified locally: served /main.js shows e.preventDefault() inside both
tool branches and the reordered done() body. go test -race ./... still
green.
- CSS: .canvas-wrap.tool-{frame,device} #canvas, #canvas * { cursor:
crosshair !important } so frame/device rects don't display grab while
a tool is armed
- Server: Cache-Control: no-cache on embedded static handler so browsers
revalidate via ETag instead of serving stale main.js after redeploy
Issue 1 — cursor lies about armed tool. .svg-draggable { cursor: grab }
on frame/device rects beat the .canvas-wrap.tool-device #canvas {
cursor: crosshair } rule because element-level wins over descendant.
m saw "grab" hovering a frame with +Dev armed and thought the tool was
broken even though clicks routed correctly after the previous fix. Add
a descendant rule with !important so tool-armed wraps any child cursor:
.canvas-wrap.tool-frame #canvas *,
.canvas-wrap.tool-device #canvas * { cursor: crosshair !important; }
Issue 2 — stale browser cache after each redeploy. http.FileServerFS
served embedded assets with no Cache-Control header, so browsers held
on to the previous main.js/style.css until hard-reload. New noCache
middleware on the static handler emits Cache-Control: no-cache. Note:
embedded FS files have zero ModTime, so http.FileServer suppresses
Last-Modified — every fetch is a fresh 200 rather than a 304. Fine at
~30KB of JS+CSS, and fixes the staleness problem completely.
Middleware is wrapped only around the static handler. /api/* responses
write their own headers and aren't touched.
Verified locally:
curl -I /main.js → Cache-Control: no-cache
curl -I /style.css → Cache-Control: no-cache + contains the new rule
curl -I /api/healthz → unaffected (no Cache-Control from us)
go test -race ./... still green.
Move the [data-frame-id]/[data-device-id] early-return below the
tool-armed branches in onCanvasPointerDown. With a tool armed,
the canvas-level handler always wins; without a tool, the original
behaviour (frame/device pointerdown handlers capture for drag/select)
is restored.
onCanvasPointerDown returned early whenever the click landed on a
[data-device-id] or [data-frame-id] element so the per-element drag
handlers wouldn't get hijacked. Problem: this early-return fired BEFORE
the tool check, so clicking +Dev inside an existing frame never reached
placeDeviceAt().
Reordered: tool-armed branches run first and short-circuit. Only when
no tool is armed does the "click started on a child element — leave it
alone" guard kick in. End behaviour:
- +Dev anywhere (incl. inside a frame) drops a device. frame_id
auto-resolves via the existing frameAt() point-in-rect.
- +Frm anywhere (incl. inside an existing frame) starts a rubber-band;
rare but not harmful.
- No tool armed: clicking a device/frame still goes to its own handler
(drag / select). Clicking empty canvas still clears selection.
Hand-tested via the served /main.js + the equivalent backend POST/PATCH
sequence: device-in-frame, device-outside, device-drag, frame-drag with
cascaded device patches — all work.
Renders the slice-2 backend on the empty canvas from slice 1.
Canvas:
- Frames render as dashed-stroke rects with top-left label, slightly
tinted fill. Devices render as solid-stroke rects with centred label
in device.color.
- Selection halo via .selected class (stroke-width bump).
- Empty-state hint disappears once any geometry exists.
Tools (left sidebar + keyboard):
- F / + Frame — rubber-band rect on the canvas. <80×60 cancels. On
release, inline foreignObject namer → POST /api/projects/:pid/frames.
- D / + Device — single click places a 100×35 device centred at the
click. Inline namer → POST devices. Drop-point determines initial
frame_id via point-in-rect against all frames (smallest bbox wins).
- Esc cancels active tool / inline namer / clears selection.
Drag (pointer events + svg getScreenCTM):
- Devices: drag updates x/y live via transform, persists via
PATCH .../devices/:id on pointerup. Also recomputes frame_id from
drop point and includes "frame_id": null|<id> if it changed.
- Frames: dragging a frame moves its contained devices visually too;
on pointerup, single PATCH for the frame + one PATCH per moved device.
Children-batch is computed at pointerdown and only sent on release —
no per-pointermove network traffic.
Inspector:
- Frame selection: name (debounced rename), x/y/w/h, device count,
Delete button (confirm prompt — devices keep existing, frame_id → NULL
via the schema's ON DELETE SET NULL).
- Device selection: name (debounced rename), colour picker
(change-event PATCH, no debounce), x/y/w/h, current frame, Delete.
- Background click clears selection.
devicePatch wire format uses tri-state frame_id: key absent = leave,
key:null = clear, key:<int> = move. Frontend uses `null` explicitly
when a device drops outside all frames.
All 8 endpoints (list, create, patch, delete) for both resources. Path
params parsed via Go 1.22 ServeMux PathValue.
devicePatch uses json.RawMessage for frame_id so the wire format
distinguishes:
- key absent → leave as-is
- "frame_id": null → clear (device leaves all frames)
- "frame_id": 42 → move to that frame
parseFrameRef translates that into the store's db.FrameRef tri-state.
Sentinel-error mapping unchanged (writeError covers ErrInvalidInput,
ErrConflict, ErrNotFound, etc.). Cross-project frame_id refs surface as
400.
Snapshot now populates frames + devices from the DB (slice 1 left them as
empty arrays).
Frame store:
- CreateFrame requires positive width/height; rejects empty name; UNIQUE
(project_id, name) collisions surface as ErrConflict via mapWriteErr.
- GetFrame is project-scoped — wrong-project read returns ErrNotFound.
- UpdateFrame applies a partial; project_id is not exposed (moving a
frame across projects would orphan its devices).
- DeleteFrame relies on the schema's ON DELETE SET NULL to drop
devices' frame_id refs cleanly; verified by test.
Device store:
- CreateDevice defaults color to #1e1e1e if blank; rejects empty name,
non-positive size; validates frame_id is in the same project (returns
ErrInvalidInput on cross-project ref).
- UpdateDevice uses a FrameRef tri-state for frame_id so callers can
distinguish "leave alone" from "clear to NULL" from "move to frame X".
- Cross-project frame_id on PATCH is rejected with ErrInvalidInput.
- ListDevices supports an optional frame_id filter.
13 new table-driven tests, all green with -race.
mAi got admin on m/mCables but Gitea container packages are
user-namespace-scoped — repo-collab perm is insufficient. Pushed
once using m's ~/.netrc token, deleted the mAi/mcables stub.
Compose now references mgit.msbls.de/m/mcables:latest.
m granted mAi admin on m/mCables, but Gitea's container registry is
user-namespace-scoped (not repo-collab-scoped) so the push had to go
through m's own credentials for this one administrative move:
docker login mgit.msbls.de -u m -p <m's token>
docker push mgit.msbls.de/m/mcables:latest
Image digest sha256:76624f17… is identical to the one previously living
at mgit.msbls.de/mai/mcables:latest — same build, just retagged.
Drops the workaround comment from the compose file. The mai/mcables
package will be deleted via API after the deploy verifies.
picasso shipped (commit 8a31f0a on mai/picasso/deploy-mdock):
- Dockerfile: multi-stage golang:1.23-alpine -> distroless/static
- docker-compose.yml at repo root (raw-docker pattern, not Dokploy)
- .dockerignore
- README deploy section
Live: http://mdock:7777 (image sha256:76624f17, 12.2MB).
Persistence verified across compose restart.
Note: mAi lacks write on m/ in Gitea, so image lives at
mgit.msbls.de/mAi/mcables:latest. m can retag once mAi gets write
on m/mCables (see docker-compose.yml comment).
Pulls the deploy infra forward from §10 so m can see slice 1 on his LAN.
- Dockerfile: multi-stage golang:1.25-alpine → distroless/static-debian12.
CGO_ENABLED=0 (modernc.org/sqlite is pure Go). USER 1000:1000 so the
bind-mount on mDock (owned by m:m) is writable without chowning the
host dir. -trimpath + -s -w; 12.2MB final image.
- docker-compose.yml: matches the mDock convention surveyed earlier
(container_name explicit, restart: unless-stopped, env_file in
/home/m/secrets/mcables/.env, bind-mount /home/m/stacks/mcables/data,
port 7777 exposed on LAN). Image temporarily under the mai/ namespace
on mgit.msbls.de because mAi doesn't have write access to m/* today —
documented in a comment so retagging is one line when permissions land.
- .dockerignore: keeps .git, .worktrees, .m, data/, docs/, *.md,
editor cruft out of the build context.
Manual deploy verified end-to-end:
- docker build → image sha256:76624f17 (12.2MB)
- mAi-authenticated push to mgit.msbls.de/mai/mcables:latest
- ssh mdock anonymous pull works (registry allows public reads on this
namespace)
- POST /api/projects {"name":"LOFT"} returns the row, GET /api/projects
shows it; docker compose restart preserves it on disk; second GET
still shows LOFT.
Gitea Actions auto-deploy left for a follow-up task per the head's
instruction — gets us the moving parts right first.
