The instance inspector drew its own, flatter version of a struct: one row per
field, no bundle head. It now draws the same block as the asset side, because
there is only one implementation left - both call buildStructBlock. What the
instance does not get is everything that touches the shape: renaming a field,
'+ Field', and the per-field '×'. The shape belongs to the asset; a placed
object overrides values. The revert and override marks sit on the head rather
than on a field, since an override covers the whole bundle at once.
Underneath this, editing a single field of an unoverridden struct used to blank
its siblings: structFieldsOf filled the untouched fields from spec.read, which
is object.overrides[key] and therefore undefined until something is overridden.
structOf then handed back the empty shape. PreviewedValue carries 'current' now
- the resolved value the row actually shows - and the fields come from that.
Without it the old path stays, which is right for the asset side, whose prop
bag is dense.
The port editor gets 'From variable' next to '+ Field': pick a struct variable
of the behavior and its fields land as ports. It only adds - a field the node
already declares stays as it is, because it may carry an edge - and the entry
is disabled once there is nothing left to add. A list of structs counts too:
it declares the shape of every item, which is exactly the bundle a Pack builds
for it. A color field arrives as a number port; the graph has no color type.
AddPortsCommand is new underneath: adding five ports had to be one undo step,
and it checks every key before it appends a single one, so a collision does
not leave half the list behind. AddPortCommand is a three-line subclass of it,
which leaves its sixteen call sites untouched.
Ein ausgeschaltetes Objekt zeichnet nichts, verschwand bisher aber auch aus
der Auswahl - im Panel angeklickt blieb der Viewport leer, es gab kein
Anzeichen, wo es liegt. Der Rahmen kommt jetzt, gedimmt auf 45 %, damit er
sich nicht wie eine fehlgeschlagene Zeichnung liest. Ziehen bleibt aus: der
Node ist weiterhin nicht interaktiv.
Dazu __gm.gizmos() als Testhaken - Gizmos waren vorher gar nicht prüfbar. Es
meldet das alpha, das wirklich in den Strich geht, kein daneben abgeleitetes
Flag.
Ein Toggle in der Graph-Kopfzeile lässt gezogene Nodes auf dem 24er-Raster
einrasten – derselben Größe, die den Hintergrund zeichnet. Einschalten bewegt
nichts; die Regel greift erst beim nächsten Zug. Bei Mehrfachauswahl richtet
sich jeder Node für sich aus, statt einem gemeinsamen Versatz zu folgen, der
die alten krummen Abstände mitschleppen würde.
Die Entscheidung liegt als reine Funktion in framing.ts, wo die übrige
Graph-Geometrie schon wohnt, und ist dort wie im Browser gepinnt.
An unconnected list input is a button into the List tab instead of a number
field; a struct field can hold a list, so ListSlot carries an optional field
name and the three owner cases only say what sits at a key; and a list
variable picks its item kind in the inspector (marker doubles as the chooser),
so a list of structs no longer has to be born from a project file.
The declaration keeps the storable shapes finite: a struct field may be a list,
but that list holds primitives only, so list->struct->primitive and
struct->list->primitive are the two deepest forms.
check 674, e2e 483, jump-run 24, Tetris 19.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Next move held four clusters of Set variable in a row, fourteen single nodes
in all; clear-lines read three variables one after the other. They are now
one plural node each, which takes Next move 40 -> 30 nodes and clear-lines
25 -> 23. Every remaining single node was checked and really is single: ok/no
are the same variable in opposite branches, take/keep the same one across a
write, and the three key setters hang on three different events.
Two contracts carry those merges, both pinned as tests. Set variables writes
per port with writeTyped exactly as the singular node does, so a Pack{rot}
still merges instead of overwriting - without it, turning would have erased
the other three fields. And the interpreter collects every input before the
node runs, so no input may come from a variable the same node writes; a data
node is evaluated once per pulse and keeps its answer. That second rule is
invisible in the game, because OnUpdate re-reads each frame and heals it, so
it lives in a unit test rather than in the smoke.
/opt/node22 and /opt/npm are gone from the tree. build.sh ran into the empty
path once they were moved away; the other seven mentions were documentation.
e2e.sh and check.sh replace the bare npm calls. A long run does not survive
the end of its shell - four full e2e runs died mid-suite with oom_kill 0 -
so e2e.sh start hands the run to a transient game-maker-e2e.service, which
does. e2e.sh status reports the running job or the last result, and calls a
log without a closing line ABORTED rather than reporting the passes it got
to. Both scripts end on an explicit exit line, because set -o pipefail inside
a script does not protect the caller: measured, ./check.sh | tail still
reports 0. Logs carry a timestamp and land in logs/, which is ignored.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A list item may now be a struct. The depth is bounded in the declaration
rather than by a comment: FieldKind (number|text|boolean|color) says what a
struct field holds, ListItemKind adds struct on top and says what a list
holds, so list -> struct -> primitive is the deepest storable shape and
sameValue stays finite. A struct item renders inline in the List tab, which
is why it needs no path in the slot address.
Four paths silently assumed a list item was a primitive: sameValue compared
items with ===, textOf would have printed [object Object], asListItem turned
a struct into text on Set item and Fill, and Index of used indexOf, which is
reference equality and could never have found a struct.
Tetris now runs on bundles. shapes is 112 {x, y} structs instead of 224 loose
numbers, piece and the candidate are structs, so Next move hands back a whole
candidate and each strand packs only the field it changes, since mergeStruct
leaves the rest standing. Declared behavior ports drop 49 -> 35, the master's
variables 18 -> 13, its graph 84/94 -> 69/76 nodes/edges.
Get variables and Set variables declare their ports from a list of variable
names, the way Pack and Expand already do, so one node can read or write
several at once. Renaming a variable rewires them, because there the port key
is the variable name.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The value type was already there: StructValue, textOf, sameValue,
writeTyped/mergeStruct and the struct port type came with the first
part, and Get/Set variable carry any-ports, so a bundle already flowed
through the graph unchanged. What was missing was the declaration.
PropertyDef gains kind 'struct' with a list of fields, VARIABLE_KINDS a
fifth kind, variableDef a fifth parameter. A declared field is a
primitive - a list field would need a nested editor and therefore the
path in the slot address, the same step "a list of structs" is waiting
for. coerceValue now returns PropertyValue rather than FieldValue, the
first value type that is not a field; none of its five callers break.
In the inspector a struct variable is a block, not a tab. The list tab
exists because hundreds of rows fit in no inspector row; a struct is the
opposite case - few fields that fit under each other. It is the port
editor Pack already has, one level down and with a value column: type,
key and default value in one place. A placed object sees the same fields
as plain value rows: overriding yes, declaring no.
Adding or removing a field touches the declaration, the default, the
behavior's value and every instance override in one command, so a stored
value can never keep a shape its declaration no longer promises. The row
handlers read the bundle at the moment of the change rather than closing
over a snapshot - a field writes the whole bundle, so with a snapshot
the second edited row revives the first one's old value. defSignature
carries the fields, otherwise a new row appears only at the next
rebuild.
Fixes a fixture fault this uncovered: step 10 pushed its text object
into the scene without instanceEdits, which overrideSites walks with
Object.entries. AddStructFieldCommand is the first caller to reach it
from that state; the step was green alone and failed three checks in the
full suite.
Moves the dynamic-ports note to done - all three of its parts are built.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A short list opened after a long one drew all its rows stacked on one
point. The fault was not in the list tab: rebuild() handed the row stack
to setContent() while it was still empty, so the content was sized to
max(rows, viewport) = the viewport. The rows arrived after that, and for
a short list the same formula yields the same height, so resize() saw no
change and skipped layout(). A long list changes the height and never
showed it.
relayoutContent() now runs whenever the content changed, rather than
leaning on resize()'s change detection - that detection is the wrong
instrument for "the children moved".
The list tab was the only affected caller: the panels use ListView,
which positions its rows itself, and the inspector fills its body
before setContent.
Also drops the comma line under the list. Every section that can carry a
list supplies openList, so the comma fallback in fieldFactory was
unreachable; listToText/listFromText go with it. The list-value test now
covers coerceListItem through coerceValue instead, so the boolean
spelling stays guarded.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A list property in the inspector was a single text field holding a comma
separated line -- unreadable past a handful of items, and it could not
carry anything but text. It becomes a button ("3 items") that opens a
sixth centre tab, where every item is a row built by the very widget the
inspector would build for that item kind.
The renderer was free: a PropertyDef of kind 'list' already carries
`items`, so item i becomes a synthetic def of that kind -- number, text,
boolean and colour rows appear without a new branch.
The way into the tab goes through the state (EditorState.listSlot,
openList(), event 'list'), like every other tab switch here; a tab may
not call showTab itself. ListSlot lives next to Selection rather than in
the tab folder, or state.ts would have to import upwards.
A row's handlers must not close over a snapshot of the list: they outlive
a commit, since the tab only rebuilds when the length changes, and the
second edited row would otherwise revive the value the first replaced.
showTab switched five `visible` lines by hand, where a sixth tab is
silently forgotten; it is a Record<CentreTab, Widget> now.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A rebuild destroys the widget being typed in, and a typed value can change the
very structure that decides whether a rebuild is needed -- a field that says
how many ports a node has. That made such a field unusable: every keystroke
previewed, the preview emitted 'project', refreshValues() saw the signature
change and threw the row away. Dragging the value worked only because a drag
keeps receiving globalpointermove on the destroyed widget.
So refreshValues() now postpones a due rebuild while the focus sits inside the
inspector, and runs it as soon as the field is left. Values are safe in the
meantime: TextField.setValue ignores an edit in progress.
Two handles for that in the toolkit. FocusManager.isWithin(node) also replaces
the parent chain refreshLive() walked by hand, and changeHandlers lets a panel
notice that the focus has gone -- there was no signal for that before.
Nothing in the editor can trigger this today, since the Sequence setting that
did is gone. It is the ground for the List tab, where a row that adds or
deletes an item changes the structure while you type. The e2e check therefore
builds the case out of real parts: type into a field, set the object's parent
from outside -- the parent is part of the structure signature -- and confirm
the input survives and the rebuild follows on Escape.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Ports that vary per node already existed, but only for the three nested-behavior
types, and the rule for them sat as a switch on type names in nested.ts – the
one place in the project that knew node types by name. It dissolves into two
optional contracts on NodeDef, so a type carries its own rule:
declaresPorts the node holds an editable list, shown as a section in the
inspector – behavior start/end, and now Pack/Expand
derivePorts the node computes them from its settings – Sequence from a
count, Run behavior from the behavior it names
Sequence gains an Outputs setting (2 to 26). Its keys stay a, b, c and count on
with d, e, so the edges of the sequences already out there survive.
A param can now take a port away, and the edge on it would stay: it draws as
nothing yet survives every save, and comes back to life if the port returns.
SetNodeParamCommand takes those edges with it, which also closes the same leak
that switching a Run behavior node's target has always had.
The port editor addresses its node by id rather than by type, so two Pack nodes
can live in one graph, and it calls the rows Fields when they are not a
behavior's signature.
Steps 18 and 20 both hand-rolled the same object-plus-behavior scaffold, down
to lane offsets that nothing kept distinct. Step 19 claimed in a comment that it
cleaned up after itself but never checked it – the only one of the four that
did not.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
GM_STEPS=20, 1-12 or 6,20 takes a subset of the suite, which is seconds
instead of two minutes while one step is being written. Steps live off
what the earlier ones leave behind, so a subset that skips a prerequisite
fails on purpose. The filter had been written as a throwaway script three
times; it belongs in the runner.
The README stops calling a remote browser the way this suite is meant to
be run: with the GPU in use a local run costs ~124 s and about 0.8 of the
four cores here, so GM_REMOTE is a convenience, not a requirement.
A held key fired once and then nothing, so sliding a Tetris piece sideways or
walking a menu meant building a timer next to every key node. Two settings on
the node do it now: Repeat after and Repeat interval, both in milliseconds so
the usual values stay whole numbers (300, 50) instead of 0.3 and 0.05. A delay
of 0 keeps the node exactly as it was, an interval of 0 means every frame.
The browser's own auto-repeat is deliberately not used - Input.press has always
dropped it - because its delay and rate are an operating-system setting of the
player's, not a game's. The repeat is built from the measured hold time instead,
per node, so two nodes on one key keep their own clocks. The rule itself is a
pure function in keyRepeat.ts and is checked without a world.
Both rows stay off the node body while the delay is 0: two lines of "0" on every
key node in every graph would be noise, and the screenshot said so. NodeDef gains
hiddenParams for that; the inspector still shows them.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A link can be folded away from the canvas: instead of a curve running clear
across the graph, each end grows a short stub with a mark naming the node at
the far side. The curve comes back while either of its two nodes is selected,
so following a connection is a click rather than a search, and a link carrying
a live pulse is always drawn so the trace keeps its thread.
Marks are collected per port, not per link, because several hidden links on one
port used to stack stubs, dots and labels on identical pixels. A bundle reads
"Move +1" and its dot selects every node the port feeds, as one undo step.
The whole thing is one optional flag on the edge; the interpreter never sees it.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The board is a list of 200 numbers now instead of 200 characters, so a cell is
Item at and a landing is Set item instead of seven nodes of string surgery.
The bigger change is that the one 219-node graph is gone. Each step of a frame is
a behavior of its own - Spawn piece, Next move, Fits, Draw piece, Lock piece,
Clear lines, Show lines, Draw board, plus Piece cell for where one of the four
cells of a piece sits. None of them holds state: what they need arrives through
their inputs, every answer leaves through their outputs. That is what lets Fits
answer for turning, moving and falling alike, and Piece cell serve the fit test,
the drawing and the landing instead of being the same twenty nodes three times.
Tetris itself is down to 84 nodes and only says in which order the steps run and
which variable each answer lands in. The smoke test holds it to that: it fails if
a loop or a list access reappears in the master graph.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A variable could only be a number, a text or a yes/no, so every set of values had
to be smuggled through a string - which is where Tetris got its unreadable
"03066326048763714" colours from. A list is now a value of its own, with an item
kind (number, text, yes/no, colour) that decides how the inspector reads the one
comma-separated line it is typed as, and eight nodes to work with: List, Fill,
Count, Item at, Set item, Sub list, Join lists, Index of.
Two traps came with it. An array is typeof 'object', so every place that told a
value from an entity had to learn the difference, or a list would have been
passed around as a game object. And a value was compared by identity in three
places - with lists that marks everything as an overridden property after each
reload, so those comparisons now compare contents.
There is no list of objects: a list is written to the project file and an object
is a live handle that no reload survives. A set of objects stays a list of names.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A behavior node that cannot run looked exactly like a healthy one: no
start node, no behavior picked, a deleted behavior or a self-call were
all silently skipped. The call still had to be forgiving at runtime -- a
dangling reference must not abort a game -- so the fix is diagnosis, not
an exception: describeProblem() names the one reason, the node draws it
in red under its title, and the inspector repeats it where you pick the
behavior. A missing end node is deliberately not a problem: a call is
always synchronous, and the end node only decides which values come back.
Two ways in: a behavior dragged from the asset panel lands in the graph
under the pointer, and double-clicking a call node opens the graph it
runs. Both reuse machinery that already existed -- the asset panel
already starts the drag, and list views already report double clicks.
The e2e module needed its assets at the head of the asset list: after 18
modules the list is long enough that a newly appended row scrolls out of
view, and the press then lands on the panel behind it, starting no drag
at all. That looked exactly like a rejected drop and cost two full runs.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A behavior that has a Start and an End node can now be dropped into any
other graph as a node of its own. What is declared on the start node
becomes the inputs of that node, what hangs on the end node its outputs;
the port list is derived from the sub graph, not copied, so declaring a
port in one graph grows the node in the other without touching it. The
node carries the behavior's name and the palette lists behaviors by
name, so a call is one grab instead of a node plus a dropdown.
An embedded behavior has no object and therefore no place for
variables: every call starts on the asset defaults and communicates
only through its ports. That is the conservative direction - state can
still be added later, and shared state already has a home in a behavior
object in the scene. Node ids are unique per graph only, so each call
gets its own pulse cache; without that, two graphs would collide.
Renaming a port used to silently drop the value typed on the calling
node. It is carried along now, project wide, like the edges.
A Tetris example project needs to draw 200 board cells and scan rows
for a full line. Neither is expressible without iteration, so the node
system gets a Repeat node plus the missing text side: Substring, Find
text and To number. The interpreter drops its data cache once per
round -- otherwise every round would recompute with the first round's
values.
Object by name takes its name as an input now, not just as a setting;
addressing "c" + index is impossible otherwise. An unconnected input
reads the stored setting, so saved projects are untouched.
The board itself is a 200-character text, "." for empty. That makes
both operations of Tetris a single expression instead of a loop, and
each frame attempts exactly one move -- rotate beats sideways beats
fall -- so the fit check exists once in the whole graph.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Pause is a sub-state of play, not a third mode: the toolbar button and
Ctrl+Enter stop ticking the world, the clock included. A held frame still
drops the input edges, so resuming does not fire a pause worth of keys in
one go.
The editor run missed the case it is most wanted for. hasEditorRun only
looked at the stored scene objects, so a behavior inside a placed
container never started one; the world now builds from resolveObjects and
writes back through objectRef, keeping the move as an instance edit
instead of baking the instance out.
Keys and clicks go to the run while the Game tab is on screen, where no
key means anything to the editor, and are handed back on leaving it -
held keys released, so nothing keeps walking unpressed.
A behavior object carries a "Run in editor" switch next to Enabled. When
it is on, the behavior ticks outside of Play and works on the saved model
directly - Unity semantics: what it moves is a real change, autosaved and
not undoable.
Input stays with the editor, so On key never fires while typing. The
editor run never bakes placed containers into the model, since that would
silently drop their instanceEdits.
syncToModel reports whether it wrote anything, and the dirty flag is only
raised on change - otherwise a behavior moving something every frame
would trigger ~60 saves a second.
The game tab held a "Press Play" hint where the picture belongs. It now
previews the scene through the active camera, and says "No active camera"
when none is switched on.
Camera choice and its world position moved into activeCamera.ts so both
are testable without Pixi. Entity.x/.y are parent-relative; the viewport
used them as the camera's spot, which framed a parented camera at the
wrong place.
The game view took whatever camera came first in the scene, ignoring the
switch, so `enabled = false` on a camera did nothing and a second camera
never got its turn. `World.activeCamera()` now decides: of the cameras
switched on – the whole parent chain counted, as for any object – the
first in hierarchy order, and none on renders nothing at all. It is asked
per frame, so a graph switching cameras hands the view over immediately.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The pixel-art check falls over sporadically, and its follow-up hid it. That one only
asked whether some pixel across the border is blended – but in the failing case the
`nearest` reading is already blurred, so the `linear` one is too, and the check passed
while saying nothing about whether the switch had any effect at all.
It now also demands that the picture actually changed against the `nearest` reading:
the summed colour distance over the five sample points must exceed 150. Measured on the
readings recorded in the note, a working switch drifts 425 and a hung one 0 – and 30
even if every channel were off by two, so the threshold is not brittle against noise.
Exact equality was the first draft and would have gone green on a single wobbling bit.
Both checks now carry `__gm.textureFilter(file)` in their message, so a failure says by
itself whether the model never switched, the switch missed the GPU sampler, or the
pixels were measured wrong.
The cause of the sporadic failure is still unknown; four hypotheses are measured and
ruled out in the note, which is filed with that stated plainly.
"+ → Texture" opened the machine's file dialog, so a picture that was uploaded once
but lost its asset – an undone import, a deleted asset – was unreachable from the
editor while still sitting on the server. Nothing collects those files, by design.
It now opens a grid of every picture the server holds, with thumbnails: pick one and
"Use image" makes a texture asset on that file, "Upload…" is the way in for bytes the
server does not have yet and lands in the same import as a drop. A tile captioned with
an asset name is one this project already wears – picking it selects that asset instead
of making a second one, since two assets on one file would fight over `filter`.
- `GET ?r=textures` now reports `width`/`height`; a texture asset holds the pixel size,
and reading it off the GPU texture would tie the asset's correctness to load timing.
- `TextureLibrary.load(file)` loads without touching `scaleMode` – a thumbnail must not
restage the filter of a picture the scene is drawing with.
- The grid geometry lives in `texture/imageGrid.ts` as pure functions; inside the widget
no test could reach it.
- The part after an upload is now one function shared by the drop import and the picker,
so "one file = one asset" is stated in a single place.
12 new E2E checks; the strongest one undoes an import, leaving the file on the server
with no asset naming it, and picks exactly that file back out of the dialog.
The inspector already marked an asset value that differs from its schema
default, but the Assets panel gave no hint, so a changed asset looked
like an untouched one until you opened it.
The rule sits next to the existing objectDeviates as a pure function:
assetDeviates asks whether any marked property differs, markedAssetSchema
says which rows the inspector marks at all. buildAssetSection now builds
its rows from that same list, replacing the behavior/rest split it used
to carry, so the list cannot claim a deviation the inspector then fails
to show.
A behavior's variables are deliberately not in it. Their default is
declared on the variable itself, which means the asset value *is* that
default; counting them would paint the fixture's Mover gold while the
inspector beside it marks nothing.
Three things the graph view did differently from the scene view, or not
at all:
Panning moves to the middle button, which works from anywhere – a node
under the pointer included, since NodeView now lets that one button
bubble up. Shift-drag on empty canvas pans too. The left button pulls a
rubber band instead and no longer pans.
Nodes can be selected several at a time: by band, by ctrl/shift-click,
moved together, deleted together, each as one undo step. The selection
model gains `nodeIds` alongside the objects' `objectIds`.
Fit really fits. It used to pin the zoom at 1 and merely centre, so a
graph larger than the view always left nodes hanging out, and it sized
every node with one guessed height although real ones vary with their
port count. Both sums now live in editor/graph/framing.ts as pure
functions, which is what made the defect testable at all.
The band selects any node it overlaps rather than one whose position it
contains, unlike the scene view: nodes are large rectangles, and a band
drawn visibly across one has to catch it.
Pixi only fills `detail` on the `click` event it synthesises; on
`pointerdown` it is always 0, so every `e.detail === 2` branch was dead
code. Asset rows, hierarchy folding and select-all in text fields never
responded to a double-click.
A shared click counter in src/ui/clicks.ts replaces the three readings of
`e.detail`. A finished double-click ends the run, so two in a row count
2 and 2 rather than 3 and 4.
The harness could only send `clickCount: 1`, which is why no test caught
this: every "activate" check went through Enter. It can send real
double-clicks now, and five checks drive the three paths separately.
The context menu entry is gone; dragging replaces it entirely. The drop
target is the view under the pointer, not the active tab, so the same
gesture works for scenes and for container templates.
While dragging, an outline shows the footprint the object will have plus
a crosshair on its origin, turning red when the drop would be refused.
Preview and command share one function (placementRejection), so the
outline cannot promise something the drop then rejects.
A left-press used to select the list row immediately. Since the container
tab follows the selection, it switched to the very container being dragged
and every drop would have nested it inside itself. The list now selects on
release without a drag: picking something up is not the same as choosing it.
Releasing an edge over empty space opened the node palette but dropped the edge
drawing, so the connection being made was invisible exactly while its target was
being chosen. The finished drag is now held in `pendingDrop` and drawn like a
running one until the palette closes, by pick or by Escape.
The e2e check for the Escape path exposed a real defect underneath: `Popup.open()`
blurs the focused widget, `close()` did not. A field inside the popup therefore kept
the keyboard bridge's DOM focus after the popup was gone, and every later Delete or
Ctrl+Z went into that invisible input. The suite caught it one check later ("Delete
removes the node and its edge" found the node still there). `close()` now blurs,
mirroring `open()`.
Verified remotely on the Windows host: 287 editor checks and 24 jump-and-run smoke
checks, no failures.
Dragging an edge now rings the port it would connect to and snaps the line
there. A drop on a node's body takes that node's first free compatible port,
or the last compatible one when none is free, replacing its link. A drop into
empty space opens the palette filtered to the node types the edge could reach
at all, and the picked node arrives wired - node and link as one undo step,
so a single undo takes back the whole gesture.
The decision of what fits what lives in core/graph/dropTarget.ts as pure
functions, tested directly; the canvas only draws the answer. The palette
takes an optional filter rather than knowing about drags.
Verified remote on the Windows box (GPU): ALL PASSED, 284 checks.
An inspector label whose value deviates from its base was drawn bold, which is
easy to miss next to the regular labels and says nothing in the hierarchy, where
a folded subtree could hide a change entirely.
The colour now lives as a theme token (`override`, #f2c84b) used by both the
inspector labels and the hierarchy rows. Whether a row deviates is decided by
`src/core/properties/deviation.ts`: `objectDeviates` compares an object against
what it follows -- the container template for content, the asset otherwise --
and `deviationMarks` carries the mark up to every ancestor, so a folded subtree
still shows that something inside was changed.
`deviationMarks` takes the stored holder, not the resolved one: in the resolved
view container content looks like an ordinary object and `instanceTargetOf`
would compare against the asset, marking every instance.
Covered by 3 unit tests and 5 e2e checks, including one that strips an object's
own overrides to prove the row stays gold for a deviating child and restores
them byte for byte. Full suite remote on the RTX 4070: 275 checks, all passed.
`decimals` only reached `formatNumber`, so it changed the label and nothing else: typing 640.5 into
a field with `decimals: 0` stored 640.5 and displayed 641. The two existing users got away with it
because rounding happened outside the widget (`normalizeFrameRect`, the slice dialog), which their
comments had to point out.
`NumberField.clampValue` now rounds before it clamps, and every path that sets a value goes through
it (typing, scrubbing, arrow keys, the constructor), so the stored number matches the shown one and
`min`/`max` stay exact. The rounding is `roundTo` in math.ts, which `snapToStep` already did inline.
With that safe, camera `View width`/`View height` and text `Font size` become integers – they were
showing 640.000 and 32.000. Everything else that looked integral is not: positions, sprite and
collider sizes and zoom are all scrubbed, and `z` is a summed coordinate (walk.ts), not a layer.
Verified remotely against the RTX 4070 host: 270 checks, all passed (step 10 gained a check that the
field text has no decimal point and survives a reopen). `npm run check` green, 407 tests.
Both shutdown paths ran Get-Process chrome,msedge | Stop-Process -Force,
which is every browser on the machine. The remote host is somebody's
desktop: a test run must not close the windows they are working in. It
only stayed harmless because the test account cannot touch another
user's processes.
The launcher now writes its browser pid to %TEMP%\gm-browser-<port>.pid
and kills that pid; the job object still takes the renderer children.
Measured against the Windows desktop: five msedge processes belonging to
the logged-in user before the run, five after, and our chrome gone.
24 smoke checks ALL PASSED.
The leash that ties the Windows browser to the run never held. Three
findings, each measured:
- stdin is not a signal: only the first chunk a caller writes reaches
the remote command, so the stdin heartbeat looked like a dead caller
and killed a run at check 44.
- a dropped SSH connection does not kill the remote command; Windows
OpenSSH leaves it running. Six orphaned chrome.exe survived a
SIGKILLed tunnel -- this is what locked up the desktop.
- Get-NetTCPConnection goes through WMI, which is denied to the
non-admin test account; the broken CIM session took a browser down
mid-run.
The launcher now watches the forwarded CDP port with .NET sockets: a
connected client means the run is alive, 45 s without one ends it. A job
object with KILL_ON_JOB_CLOSE and a 30-minute cap catch the rest, and
the caller shoots the browser on the way out instead of waiting for the
idle window.
Measured: survives 80 s of an idle held connection, gone 44 s after
kill -9 on node, immediately on an ordered end. Full suite 267 checks in
128 s and 24 smoke checks in 16 s against an RTX 4070, with this machine
at load 0.04.
The harness forced software WebGL (--disable-gpu --use-gl=swiftshader) on every host.
Measured here on an idle editor page over 20 s: that costs 373% of a core at 8 fps,
while the Intel HD 630 via ANGLE/Vulkan costs 69% at 56 fps. The full suite now runs
in ~124 s. Note the app is not rendering 60 fps as previously assumed - in software
it only reaches 8, and a single frame costs more than any throttle interval would save,
which is why capping the framerate had measured no effect.
--use-angle=gl and --use-gl=egl look cheaper still but get no WebGL context at all,
so the app never renders. They are not usable and are named in the comment as such.
There is deliberately no software fallback: a host without a usable GPU should fail
loudly rather than drag the machine to a standstill. None of the 267 checks reads the
rendered image anyway - they read window.__gm.locate() and the debug hook, and the
screenshots assert nothing - but Pixi needs WebGL2 to run at all.
The browser launch moves out of harness.mjs into tools/lib/browser.mjs, which also
supports GM_REMOTE=user@host via an SSH-forwarded CDP port. That path is unfinished:
the tunnel degraded until the suite crawled at ~25 s per check, and forcing software
WebGL on a desktop with a real card brought it to its knees. README says so plainly
rather than advertising it.
Also: the jump-run HUD check asserted a fixed "Coins 1/10", which fails on a faster
host where the same walk sweeps up two coins. It now asserts the HUD against the count
the run actually reached - the real guarantee is that HUD and state agree.
Verified: npm run check green (404 tests), 267 e2e ALL PASSED and 24 jump-run checks
ALL PASSED against the GPU flags. The fallback removal leaves the flag set on this host
identical (same flags, different order), so it was not re-run.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
typeText fired Input.insertText per character and never looked back, so a
partially typed value was indistinguishable from a wrong one: a run committed
x=3 for "300" and the failing check blamed the value rather than the three lost
characters.
Trying to reproduce that failed over 21 rounds across four variants (no pause,
400ms between characters, with the preceding scrub, and both commit paths), and
the code rules out the obvious suspects: typing lands in a hidden input, each
character is only a preview, objectSignature omits position so no rebuild is
triggered, and NumberField.setNumber guards write-back with isEditing. Cause
unknown, so this makes the symptom name itself instead of hardening against a
mechanism that may not exist.
Typing stays per character - that exercises the live-preview and search-filter
paths - but now asserts the text is in the focused input, reporting focus and
actual contents when it is not.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A run whose chromium died stopped mid-step and exited with code 0, silently
reporting success for checks it never ran: CDP.send() left its promise pending
when the socket closed, and an unsettled promise does not keep node alive.
Waiting on the server had the mirror problem – a bare fetch keeps its socket
open, so a stalled response hung the run forever.
Every wait can now time out, the socket rejects what it can no longer answer,
and reaching process exit without a summary reports RUN DIED with exit 1.
Chromium's stderr and exit signal are kept, so "killed by SIGKILL" is visible
rather than guessed at.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The aspect button now carries two offers, so it becomes a menu: fit one
side to the texture ratio, or set both sides to the exact pixel size as a
single undo step.
Dropping a row on the outer quarter of another row inserts it into that
gap instead of parenting into it, reordering among the siblings there.
Order is the stored order, so ReorderObjectCommand is a splice; equal z
falls back to tree order, which makes this the way to draw one object in
front of another. Rejected on container content like the other structural
edits.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Two regressions from the template-resolution rework, plus the harness gap
that hid them:
- `instanceTargetOf` decided from the shape of the id. Unpacking keeps the
derived ids, so unpacked content still counted as template content. It now
asks whether the object is spelled out in the holder.
- The parent picker captured its holder when the row was built. Since
`viewOf()` returns a derived copy, the menu listed the objects of that
moment (measured: 40 in the model, 17 in the menu). It resolves on open.
- `evaluate()` swallowed page exceptions, so a mutate that died half-way let
every later check pass or fail for the wrong reason.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A scene now stores only the container object; its content is resolved from
the template on every read, with deviations kept as sparse instanceEdits on
the container object. This drops the propagation bookkeeping entirely --
templateSync.ts is gone and the three structural gaps (add, remove, reparent)
cannot diverge any more.
Values stay editable on an instance, structure does not; Unpack writes the
content into the scene, keeping ids. Play mode bakes containers into the
throwaway copy before entities are built. No migration: formatVersion is 2
and the seeders write the new format.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The Container tab reuses SceneViewport against any object holder, so a
container template is edited exactly like a scene. Right-click no longer
changes the selection, which previously made nesting impossible: opening
the menu on a container loaded it, and "Place in container" then targeted
itself.
Template edits (move, rename, override) now reach already-placed instances
per slot, through a shared templateSync used by both the apply path and the
plain object commands. Copies the user has touched keep their own value.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Add ↔ next to Width and ↕ next to Height in the inspector. Each sets its
field so the sprite matches the ratio of its assigned texture or frame,
reading the frame's own rect rather than the sheet's.
The buttons reuse the existing FieldActions convention, so they simply
do not appear when no texture is assigned, when the sprite is already
proportional, or on assets whose schema has no texture property.
Comparison is a cross product with a relative tolerance: 3135 of the
40000 size pairs up to 200x200 do not survive a round trip through a
division, so an exact check would leave the button standing after its
own click.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Remove comments that only restate the code, keeping the section banners
and the notes that carry a reason, a measured fact or an external
contract.
ButtonOptions.width claimed to be a fixed width but only set minWidth,
so the Graph and Texture pickers grew with the name they showed. It is
now what it promises, with a separate minWidth for the buttons that have
two forms. Narrow buttons shrink their padding, otherwise a 24px icon
button had 8px left for its glyph and truncate dropped it entirely.
__gm.labelsDropped() guards that class: the overflow guard only saw text
that was too wide, never a label truncated to nothing.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
- NumberField takes a decimals option (default 3); frame rects and the
slice dialog use 0, matching their integer model.
- Button truncates its label to the drawn width; Dropdown reserves the
chevron slot. NodeView reuses the shared truncate().
- Setting scaleMode leaves the style's shared resource id in place, so
the old GPU sampler stayed in use until a reload. Call style.update().
- Graph gets a header bar like the Texture tab: GraphEditor is now the
frame, the former editor is GraphCanvas. A selection the tab does not
concern no longer unloads the behavior.
- New e2e step 14 guards text overflow across all widgets.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A new Texture tab next to Scene/Game/Graph shows a texture sheet and lets
frames be drawn on it: drag to create, move, resize via eight handles, edit
name and rect in the inspector, Delete to remove. Every change is a command,
so one gesture is one undo step. "Slice..." cuts a grid from columns and rows
with optional offset and spacing, previewing the resulting frame count and
cell size live.
Sprites pick a frame instead of the whole texture through the same `texture`
property: it now holds either a texture id or a frame id. Frame ids come from
the same generator and are document-wide unique, so one picker, one property
and one set of instance overrides cover both cases, and every sprite written
before this still means the same thing.
A Graphics fill ignores a Pixi sub-texture's frame and draws the whole image,
because it derives its uvs from the shape bounds and the inverted style
matrix. The cutout is therefore a matrix (`viewport/textureFill.ts`), which
keeps a sheet at a single GPU upload no matter how many of its frames are on
screen at once.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Images are uploaded to a content-addressed blob endpoint and stored as
data/textures/<sha256>.<ext>; the project only keeps the file name. The
sprite points at a texture asset through an ordinary property of the new
kind `asset`, so instance overrides, "Apply to container", "Reset to
asset", undo and `Set property` work without any extra code. The new kind
also replaces the ad-hoc asset picker the graph nodes faked as an enum.
Import goes through + -> Texture, drag & drop onto the canvas or Ctrl+V.
Images wider than 2048px are downscaled client-side and re-encoded to
WebP if needed, because nginx caps uploads at 1MB; an image that already
fits is uploaded byte for byte so pixel art is never resampled.
The sprite colour default changes from red to white, since colour tints
the texture multiplicatively and a tint of white shows an image as it is.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Menu entries are left-aligned and live in a ScrollContainer capped at 320 px, so a
long list (the parent picker holds every object of the scene) scrolls instead of
running off the screen. `Menu(items, { searchable: true })` puts a filter box above
the scroll area; Enter takes the first match. Filtering forces a relayout because
two queries with the same hit count leave the height unchanged, which would keep
the stale row positions.
Fields that open a menu are a new `Dropdown` widget: the value sits flush left and
the chevron is drawn flush right instead of being written into the label. A menu
opened below a field is never narrower than that field.
New debug hooks `overlayTexts()` / `overlayRect()` make the popup contents
assertable; e2e step 11 covers the cap, the scrolling and the search box.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
root