Commit Graph
3 Commits
Author SHA1 Message Date
adminandClaude Opus 5 d7a2521ca7 A port can take its type from what is plugged into it (#207)
game-maker / gate (push) Canceled after 1m8s
game-maker / suite (push) Canceled after 1m8s
Until now every port of "First not empty" was grey and typeless, so the
editor could not tell you what actually travels through the node. The
output is now typed as soon as both inputs agree, and every any-input is
painted in the colour of its own edge. Nothing changes at runtime.

- A NodeDef can declare which of its ports carry the same value
  (`sharedType: { inputs, outputs }`). The rule itself lives as a pure
  function `agreedType` in `src/core/graph/portTypes.ts`.
- Only the output inherits. It becomes the shared type when every input
  of the group is connected and they all carry the same type, otherwise
  it stays `any`.
- The inputs stay `any` on purpose. `canConnect` checks against the port
  type and a busy input is replaced by the new edge, so a typed `a`
  could never be re-plugged with a text without disconnecting first.
- Chains do not propagate. When looking up what feeds an input, a node
  of the same kind reports the `any` it is registered with, so two of
  them in a row stay untyped and a data cycle cannot ask itself in a
  circle.
- An edge that the new type makes invalid stays where it is and turns
  red. `EdgeLayer` already paints a mismatching edge with `danger`,
  nothing is disconnected on its own.
- New `PortDef.paintType`: colour only, never truth. It is set in
  `drawnDef` and read in `NodeView.redraw`, so `canConnect`, the
  interpreter and the compiler keep seeing `any`. Every connected
  any-input is painted, not just members of a `sharedType` group; ghost
  ports keep their hollow red look.
- The pin now matches the edge that ends in it, because both take the
  colour from the resolved output type of the neighbour. Mixed sources
  therefore show two colours, which explains why the output stays grey.
- New hooks `__gm.graph.portTypes(id)`, `__gm.graph.edgeColor(id)` and
  `__gm.themeColor(name)`. `portTypes` reports type and paint colour
  separately so the type checks do not silently become colour checks.

Findings:

- `resolveDef` only appends derived ports to the registered ones, so a
  fixed port cannot be retyped that way -- it would appear twice. The
  shared type is therefore resolved before the ports are built.
- Answer 5 of the issue (inline field follows the type) has no case left
  under variant A: an unconnected input stays `any` and keeps its text
  field.
- `Item at`, `Join` and `Equals` do not profit: `Item at` has no
  any-input, the other two have fixed outputs. The next candidates are
  the shader arithmetic nodes, which were left alone.

Tests:

- ./check.sh -> 143 files, 2127 tests, exit 0 (was 142 / 2113)
- GM_STEPS=1-6 fresh -> ALL PASSED, 245 checks (was 242), 9 of them new
- 14 new unit tests in tests/core/shared-port-type.test.ts
- 9 new e2e checks in block 6z, one of them a pixel check straight on
  the pin via pixelAt + themeColor
- Counter-probes all bite: filtering unconnected inputs out of
  `agreedType` fails 3 unit tests, ignoring `inherit` fails 1, killing
  the derivation in the bundle fails exactly the 2 expected e2e checks,
  and never setting `paintType` fails 4 unit tests plus the 3 pin
  checks (#bdbdbd instead of #e6a8ff / #6fcf97)
- Screenshot read: `a` pink, `b` green, output `value` grey, edges in
  the same colours
- Check total 1302 (1299 + 3), 1293 was measured by CI at 1a3babf

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-22 21:15:21 +02:00
adminandClaude Opus 5 1a3babf9c5 First not empty gets a setting for what counts as empty (#206)
game-maker / gate (push) Successful in 2m46s
game-maker / suite (push) Successful in 8m46s
The node treated 0, false and an empty text alike, which is right for a
missing behavior parameter but wrong the moment a real 0 travels through
the graph. A setting now says how much counts as empty, so the same node
works for both cases.

- The new parameter "Empty means" has three steps: "Any empty value"
  (the default, the behaviour so far), "Nothing there (0 and false
  count)" and "Only a missing value".
- The default keeps every existing graph as it is, including the common
  setup where an unset animation parameter arrives as an empty text.
- The rule lives in isEmptyValue(value, mode) in values.ts; the node only
  passes the step through. An unknown setting behaves like the default,
  the same way Compare does.
- The node description is shorter because the old one was cut off in the
  inspector.

Findings:

- In the "Nothing there" step a struct is empty only when its fields are
  empty under the same step, so a position (0, 0) stays a real value.

Tests:

- ./check.sh: 142 files, 2113 tests, exit 0 (before 2104).
- 9 new cases in the node catalog table: the edge cases per step (0,
  false, empty text, empty list, position (0, 0), unconnected a) plus an
  unknown setting.
- Two counterproofs hold: letting "Nothing there" act like "Any" drops
  the 3 cases with 0, false and the position; removing the "Only a
  missing value" step drops the 2 cases with the empty text and list.
- Seen in the browser: the node shows "Empty means: Any empty value" and
  the inspector dropdown lists all three steps.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-22 20:24:01 +02:00
adminandClaude Opus 5 d98c2898b1 A new node takes the first of two values that is not empty (#205)
game-maker / gate (push) Successful in 2m46s
game-maker / suite (push) Successful in 8m38s
A graph that reads an optional value needed a chain of Equals and
Branch just to fall back on a default. The new node does that in one
step: it passes a through unless a is empty, in which case b comes
out. Both ports are any, so it works for numbers, text, lists,
structs and objects alike.

- `First not empty` sits in the data category with two any inputs
  `a` and `b` and one any output `value`. Are both empty, b comes
  out, so the second input doubles as the default.
- `isEmptyValue` in values.ts holds the rule as a pure function:
  nothing connected, 0, an empty text, false, an empty list, and a
  struct whose fields are all empty.
- The struct case is what covers a position: behavior graphs have no
  vector type of their own, so (0, 0) arrives as a struct. `{x: 0,
  y: 2}` therefore counts as set, `{x: 0, y: 0}` as empty.

Findings:

- A real 0 or false on a is passed over. That follows from counting
  0 as empty and is written down in the issue.
- The output stays any. Ports that inherit their type from an input
  do not exist in the graph; `Item at` has the same shape.

Tests:

- ./check.sh: 142 files, 2104 tests, exit 0 (was 2094).
- 10 new catalog cases: a set, a unconnected, 0, "", false, an empty
  list, a position (0, 0), a struct with one field set, an object,
  and both empty.
- Counter-check: with structs never empty and 0 a real value exactly
  those two cases fail, the rest stays green. Originals restored and
  compared with cmp, then rebuilt.
- No e2e check: single data nodes have no e2e coverage here, the
  palette fills itself from BUILT_IN_NODES.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-22 19:58:30 +02:00