Commit Graph
151 Commits
Author SHA1 Message Date
adminandClaude Opus 5.5 0e73fd85b9 Route sounds, songs and buses in one Master routing graph (#351)
game-maker / gate (push) Successful in 4m20s
game-maker / suite (push) Failing after 15m17s
A project bus could only go into one other bus, and a sound, an
instrument, a pattern or a song into exactly one bus. A shared reverb
needs a sound to go there in part and on dry at the same time. Instead
of a list of sends on every asset, the project gets one routing graph
like the routing of a song (#350), one level up.

- Menu -> Master routing opens a tab with one graph per project. It has
  a node per sound, instrument, pattern, song and bus, named after its
  asset, and one Master. Embedded assets have none, their song routes
  them.
- The nodes follow the assets. Every edit settles the routing in the
  same undo step (KeepRoutingCommand), so an undone delete brings the
  links of its asset back. A new asset gets a node and a link to the
  master. What went only into a deleted bus goes to the master.
- New nodes start in three columns: what plays, the buses, and the
  master a row above, so no link into it runs behind a bus.
- Links are sends: -60 to +12 dB, and a Mode on links out of a bus or a
  song. Out of a sound, an instrument or a pattern a link always takes
  the sound after its volume and FX. Loops are refused.
- A node without a link is silent and says "Goes nowhere".
- The rows Bus and Into are gone. Play nodes with a picked bus still
  overrule the routing.
- A click on a node shows its asset in the inspector, the Master shows
  the master volume and sound FX. A double click opens the asset's tab.
- The graph lives in project.routing and is held to the assets when a
  project is read, but nothing is added there. The editor adds the
  nodes when it opens a project. An asset without a node plays on the
  master, so old projects sound as before without a migration.
- In the mix an asset with one link at 0 dB goes straight into that bus.
  A sound or instrument with more gets a bus of its own with the sends.
  A pattern or song sends from the master of its playing, which is
  where its Mode taps. Buses of the mix can have sends now too.

Found:

- The check "the scheduler plans a looped bar" in e2e 34 wobbles. It
  failed in 2 of 5 runs here and in CI task 580 at 490fcba, before
  this change. HEAD passed it 2 of 2. Not fixed here.

Tests:

- Unit master-routing (new, 10), sound-library (2 new), buses
  rewritten onto the routing, 5 lists without the Bus row. 5 counter
  checks red.
- e2e 34 (melektaus-atx): ALL PASSED, 93 checks (3 new, 2 rewritten).
  Counter check with the Sound tab ignoring the routing: the routing
  check red.
- Gate (melektaus-atx): CHECK EXIT 0, 273 files, 4225 tests.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-09 09:18:46 +02:00
adminandClaude Opus 5.5 e721bba120 Let the places of a song share what it keeps to itself (#357)
game-maker / gate (push) Successful in 3m54s
game-maker / suite (push) Successful in 15m58s
With clips, one take serves many items: four items on the four clips
of one take. The rule from #355 gave every further place its own copy
of an embedded asset, so a cut made in the Sound tab reached one item
only. Now the song, its tracks and its items share a sound, a pattern
or a sound FX the song keeps.

- settleEmbedded no longer copies within the song. An embedded asset
  stays as long as one place names it, and goes with the last one.
- A paste from another song still gets a copy, but one copy for all
  the places naming the same asset, with its clip ids mapped.
- The picks in a song offer what the song keeps, under a line after
  the project's assets: the Sound row of an item the song's sounds
  and their clips, the Pattern row its patterns, every sound FX pick
  its sound FX.
- Make project asset and Embed a copy are unchanged.

The project cbbf160bb7e8 was merged by hand afterwards: the items of
Song 2 on three copies of Take 1 now play the same clips of the first
one, and the copies are gone.

Tests:
- Unit embedded, sound-clips, own-fx and the song tests: 66 green,
  3 counter-checks red as expected
- e2e 65 on melektaus-atx: ALL PASSED, 52 checks
- Gate on melektaus-atx: CHECK EXIT 0, 271 files, 4198 tests

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-08 21:13:22 +02:00
adminandClaude Opus 5.5 42a3759d9d Cut a sound into named clips, as a texture into frames (#357)
A sound could only be cut to one stretch, set on the asset itself. To
play two parts of one file, the file had to be imported twice. Now a
sound is always the whole file, and its clips are the stretches of it,
each with a name and a cut of its own.

- A clip is { id, name, props } on SoundAsset.clips. Its id shares the
  generator with asset ids, so a sound field holds a sound or a clip,
  as a texture field holds a texture or a frame.
- props are the cut: start, end, the fades and their curves, reverse.
  Volume, loop, sound FX and bus stay on the sound for all its clips.
  The cut rows left the sound's schema for CLIP_ROWS.
- findSound resolves either id. Play sound, the Sample of an
  instrument, a Reverb impulse, song items, the scheduler and the test
  sound of the Sound FX tab all go through it.
- Every sound field offers each sound followed by its clips, "Kick",
  "Kick / Clip 1". A clip that is gone reads as "<id>?".
- The Sound tab has a strip under the ruler with the clips as bands.
  A drag there makes a clip, a click picks one, beside the bands the
  sound. Only the picked clip has handles. Reverse, fades and curves
  in the header work on it and are greyed out without one.
- Whole file is gone from the header, as the user asked.
- Picking a clip is a selection of its own, like a frame. The
  inspector shows its name, its cut rows and Delete clip. Delete
  takes it away without asking, as with a frame. Ctrl+Z brings it
  back.
- Copies get fresh clip ids: Duplicate, Embed a copy, a song
  duplicated with its takes, and the copy a split or paste makes of
  an embedded take. An item on a clip plays the same clip of its copy.
- A double-click on a song item playing a clip opens its sound with
  the clip picked. The Sound row of an item on an embedded take
  offers the take and its clips, which no list shows.
- A new take stays the whole sound. Clips are made by hand, as the
  user decided.
- No migration in the code, as the user decided. The one project
  with cut sounds was changed over by hand.

Deviation from the issue: deleting a used clip does not warn. A frame
does not warn either, and the issue had promised "like a frame".

Tests:
- Unit sound-clips (new): 11 green, 7 counter-checks each red
- Unit sounds, embedded: adjusted to clips, green
- e2e 34 (melektaus-atx): ALL PASSED, 90 checks (3 new, 12 rewritten)
- Counter-check in the browser: clips left out of the sound fields,
  the picker check fails
- Gate (melektaus-atx): CHECK EXIT 0, 271 files, 4194 tests

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-08 20:18:47 +02:00
adminandClaude Opus 5.5 668682c4ff Give a song, a track or an item a sound FX of its own (#355)
A sound FX for one place in a song had to be a project asset, which
then stood in every picker. Now every pick of a sound FX in a song
can make one the song keeps for that place alone, on top of the
embedded assets of the first two parts.

- Create sound FX sits at the end of the Sound FX row of the song,
  a track and an item, and of the menu of a track's FX button. It
  makes "Sound FX 1", 2, … per song with the Input wired to the
  Output, picks it and opens it. That is one undo step.
- Embed a copy, while the place names a project FX, gives the place
  a copy under the same name.
- While the place names its own, the menu shows it with the mark and
  offers Edit sound FX and Make project asset.
- Picking anything else lets it go in the same undo step. For a
  track or an item `settleEmbedded` did that already. For the song's
  own row `AssetPropertyCommand` now settles the song after `fx`.
- The mark `•` moved to the core. `assetRefLabel` puts it in front of
  an embedded asset, so the inspector rows show it like the timeline.
- `FieldHandlers.moreChoices` lets a caller end a reference menu
  with entries of its own. `embeddedPatternName` is `embeddedName`
  with a base now.

Tests:
- Unit own-fx (new): 6 green, 6 counter-checks each red
- Unit embedded and 80 neighbouring files: 1434 green
- e2e 65 (melektaus-atx): ALL PASSED, 51 checks (2 new)
- Gate (melektaus-atx): CHECK EXIT 0, 270 files, 4178 tests

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-08 11:32:41 +02:00
adminandClaude Opus 5.5 23cb34f781 Keep a recorded take inside its song (#355)
A take used to land in the Assets panel as "Song – Take N", next to
every other sound of the project. Now a song can keep a sound, a
pattern or a sound FX to itself: an ordinary asset whose new field
`owner` names the song. Everything that finds an asset by its id
finds it, so playing, the tabs and the game need no change. Only
the places that list assets leave it out.

- `core/music/embedded.ts` holds the rules: which types can be
  embedded, `listedAssets`, the title "Song › Take 1", and the
  rows an embedded sound hides (Loop, Bus).
- One user per embedded asset. `SetSongCommand` settles the song
  after every edit: a place naming an asset an earlier place
  claimed, or one of another song, gets a copy, and what nothing
  names any more goes in the same undo step. The outcome is worked
  out once, so a redo brings back the same copies.
- A split, Ctrl+D and a paste get that for free. The clipboard
  keeps the embedded assets of the copied items, so a paste after a
  delete brings the take back.
- Deleting a song deletes what it keeps. Duplicating it copies
  that and points the copy at the copies. A duplicate of an
  embedded asset is a project asset.
- The item menu has Open, Make project asset and Embed a copy.
- An embedded item carries a bullet in front of its name. The
  first pick was a diamond, which the gate's font test refused:
  the UI font has no glyph for it.
- A take is "Take 1", 2, … numbered per song. With the song gone
  while it uploads, it becomes a project sound under a free name.
- Left out of: the Assets panel, every asset picker, the sound FX
  menu of a track, the test sound of the Sound FX tab and the
  naming of new assets. Delete on an embedded asset only says
  where it lives. A ping lights up its song.
- Loading drops an `owner` whose song is gone, the asset stays as
  a project asset. No migration: older projects have no owner.

Tests:
- Unit embedded (new): 13 green, 6 counter-checks each red
- Unit take, song, song-edit, asset-fx, asset-ref, serialize and
  neighbours: 116 green
- e2e 65 (melektaus-atx): ALL PASSED, 47 checks (1 new)
- Counter-check e2e 65 (asset panel unfiltered): exactly that
  check red
- Gate (melektaus-atx): CHECK EXIT 0, 269 files, 4167 tests

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-08 10:36:25 +02:00
adminandClaude Opus 5.5 a295d233ed Give sounds, instruments, patterns and songs a sound FX (#352)
A sound FX could only be picked where something was played: on a Play
node, a track or an item of a song. A sound that should always sound
through the same effect needed that pick everywhere it was used. Now
the asset says it once, and every place that plays it follows.

- New row Sound FX on a sound, an instrument, a pattern and a song,
  right before Bus. A sound goes through it on every voice, an
  instrument on every note after its graph, a pattern on all its
  notes together and a song on its output.
- The fx input of Play sound, Play note, Play pattern and Play song
  and the Sound FX of a song item now read "From the asset" while
  empty. None is a pick of its own, stored as `#none` like `#master`.
  A pick replaces the FX of the asset, it does not come on top.
- A pattern item's Instrument reads "From the asset" instead of None.
  The meaning is unchanged.
- A double-click on a sound item opens the sound, as one on a pattern
  item opens the pattern.
- The tabs play through the FX of their asset. The Pattern and Song
  tab put it on a bus of their own between the output and the bus of
  the asset. The keys of the Pattern tab play there too.
- Notes of a pattern or song go through the FX of their instrument.
  Its chain is loaded with the other needs of the sequence.

Decided without asking:
- Old items and Play nodes with an empty fx now mean "From the
  asset". No asset has an FX yet, so nothing sounds different.
- The label is "From the asset" as on Loop and Bus, not "From asset".
- A sound used as a Sample, as the impulse of a Reverb or as the test
  sound of the Sound FX tab plays without its own FX.

Findings:
- In e2e, `__gm.inspectorField('Sound FX')` and `('Instrument')` find
  the group of the Assets panel first. The check asks for the last
  match.

Tests:
- Unit asset-fx (new): 8 green. Six counter-checks each turned red.
- Field lists of sound, pattern and song: Sound FX added.
- e2e 65 (melektaus-atx): ALL PASSED, 46 checks (3 new). Counter-check
  without the FX bus of the tab: that check red.
- Gate (melektaus-atx): CHECK EXIT 0, 268 files, 4142 tests.
- Not checked: how it sounds, the Sound, Instrument and Pattern tab
  in the browser.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-07 22:02:26 +02:00
adminandClaude Opus 5.5 daeffe0f9d Route the tracks of a song through a graph (#350)
game-maker / gate (push) Successful in 3m47s
game-maker / suite (push) Failing after 14m20s
Every track of a song went straight into its output. Now the Song tab
has a second view, Routing: a graph with a node per track and one
Output, the links between them sends. A track can go into other
tracks, into several at once, and a link has its own volume and the
point of the track it taps.

- The song asset has a graph like an instrument (SongTrack, SongOutput).
  heldSongGraph keeps it in step with the tracks on load and in
  SetSongCommand: a node per track named after it, one output, a new
  track and every track of an old song linked to the output. Links of
  gone tracks, doubled links and links closing a loop are dropped.
- A link carries an optional send (volume in dB, tap). The inspector
  shows Volume and Mode for a picked link of a song: Post volume (Post
  pan), Pre volume (Post FX), Pre volume (Pre FX). SetSendCommand.
- NodeDef has two new flags. `fixed` nodes are never added or deleted
  by hand. Links between `acyclic` nodes may not close a loop, checked
  in canConnect, dropTargetPort and ConnectCommand.
- songBuses gives each track its sends instead of one `into`, and a
  bus `song:<song>:<track>:#items` for its own items. That bus is what
  solo closes on a track that only passes a soloed one on.
- Mute and solo now act on the sends (trackGates). A muted track passes
  nothing on, on any tap. A soloed track is heard with what goes into
  it, through the tracks it goes through, their own items silent.
  Only sends on a way through a soloed track stay open, so a track
  going into it and into the output is heard through it only.
- SoundLibrary: a bus has either `into` or sends. Each send is tap,
  delay, level, target. busLatency takes the latest way (arrivalOf),
  and each send is held back so all ways of a track arrive together
  (sendDelays). Taps are wired again whenever a chain comes or goes.
- The Song tab has Timeline and Routing in its header. The routing
  uses GraphCanvas with an empty palette. A track linked nowhere shows
  "Goes nowhere". Delete takes links, never a track node.
- Game playings route the same way, the output being the master of
  the playing.

Decided without asking: a send ranges -60 to +12 dB like a track,
default 0 dB Post. Mute closes the sends, the track's own level stays.

Found on the way: the inspector built the Mode row with the send as it
was when the rows were made. Setting Volume and then Mode wrote the old
volume back. Both rows now read the current send when they change.

Tests:
- Unit song-routing (new): 13 green. Sound library: 4 new, 94 green.
  11 counter-probes (loop on load, canConnect, solo, solo past the
  soloed track, mute, preFx latency, graph sync, delete guard, delays,
  taps, glide), each red.
- Existing tests adjusted to the new bus names and send shape: song,
  bus-fx, song-play, graph registry, edge description.
- e2e 65 on melektaus-atx: ALL PASSED, 43 checks (5 new). The first
  run caught the inspector bug. Counter-probe with the old closure:
  2 FAILED.
- Gate on melektaus-atx: typecheck, lint, format clean, 2 tests red
  (registry rule, edge shape), fixed and rerun locally: 56 green.
- Not rerun after the per-send solo fix: e2e 65 and the gate.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-07 20:18:09 +02:00
adminandClaude Opus 5.5 3a52f2290f Mix sounds through buses and a master (#325)
game-maker / gate (push) Successful in 3m36s
game-maker / suite (push) Failing after 14m30s
Until now every sound went straight to the speakers. There was no
point where all of them came together, so nothing could turn the
music down against the effects, and a compressor or a shared echo
over everything had nowhere to sit. The bus technique of #346 was
there already, only the end of the chain was missing.

- Bus asset (Audio): Volume, Sound FX and Into. Into is empty for
  the master and never offers the bus itself nor a bus that already
  goes into it. The list is worked out when the menu opens, so a
  change made meanwhile counts. A loop that is in a project anyway
  sends the buses on it into the master.
- The master is no asset. Its volume and sound FX are two new rows
  under Master in Project settings, one undo step each. A project
  written before has 1 and none, which is how it sounded.
- Sound, instrument, pattern and song have a row Bus, empty for the
  master. Play sound, Play note, Play pattern and Play song have an
  input bus: From the asset, Master or a bus. Inside a song or a
  pattern only the bus of the song or pattern counts.
- Set bus volume (bus, volume, time, easing) and Bus volume. An
  empty bus is the master. The volume set holds for the whole run,
  across scenes, over that of the asset. GameRun resets it at start
  and end. Set sound FX parameters without a voice now reaches the
  sound FX of the bus picked in its new bus input.
- SoundLibrary: the master is the bus "mix", each bus "mix:<id>".
  Every bus, the playing masters of #346 included, ends there, and
  only the master is heard. setMix sets them up once a context
  exists, so an editor nobody listened to opens no audio device. The
  same mix again changes nothing, which lets the editor hand it over
  on every change of the project.
- The tabs play as the game does: Sound, Sound FX, Instrument,
  Pattern and Song tab through the bus of their asset and the master.
- Pause game does nothing to the buses (the user's answer 4b). A
  game that wants quiet in its pause menu sets a bus volume itself.
- Streaming long music files is #349, split off as asked.

Finding on the way: the Into menu first worked from a list made when
the inspector row was built. After another bus was sent into the
open one, e2e 34 saw it still offered. It is now asked on open.

The SoundLibrary tests expected voices to end at the destination.
They now end in the master, which adds its input, level and pan to
every wiring and two gains to every list. Those expectations are
updated, nothing else in them changed.

Tests:
- Unit buses (new): 15 green. Ten counter-checks, each red on its
  own: loop check, bus picked, reset at run start, bus of Set sound
  FX parameters, master volume in the settings, and in SoundLibrary
  default into the master, voices into the master, run volume over
  the mix, no context before a play, a playing on its bus.
- Unit sound-library: 90 green (5 new). Node catalog: 416 green
  (9 new cases).
- e2e 34 (melektaus-atx): ALL PASSED, 75 checks (3 new). The voice
  of the Sound tab was heard on mix:<bus>, that bus into the other,
  that one at 0.5 into the master. Master volume 0.6 via the dialog,
  undo back to 1.
- Gate (melektaus-atx): CHECK EXIT 0, 262 files, 4072 tests.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-07 19:13:32 +02:00
adminandClaude Opus 5.5 b519cd192f Send a track, a pattern item or a whole song through a sound FX (#346)
A sound FX could only be put on a single sound. Now a track, a pattern
item, a Play pattern and a Play song send everything they play through
one chain of it, as an insert in a DAW. A Play note gets one of its own.

- Track: an FX box in the head next to M and S opens a menu of the
  project's sound FX, and the track's inspector has the row too.
  Order: items, track FX, volume, pan.
- Pattern item: Sound FX in its inspector. Its notes play on a bus of
  their own into the track's bus.
- Play pattern and Play song take fx and fx params. The chain sits on
  the master of the playing, after the tracks and before the level and
  the fades. Set sound FX parameters reaches it by the playback.
- Play note takes fx and fx params next to the instrument's params.
  The FX parameters go by "fx:" so the two sets stay apart. Set sound
  FX parameters on a note changes a name both have in both.
- A Pitch on a bus is a new PitchShifter: the phase vocoder of the
  PitchPlayer, fed while it plays. It cannot look ahead and is exactly
  one frame late at every pitch, 2048 samples.
- That latency is made up. The other way of a Mix is held back by a
  delay of the same length. Every note and clip starts as much earlier
  as the buses it passes are late. The scheduler plans that much
  further ahead, and the first note waits that much after Play.
- The Song tab waits for the chains before it plays and lets its buses
  go a moment after Stop. A playing in a game lets them go after its
  echoes and what the shifters still hold.
- songBuses moved to its own file. Kept in song.ts it pulled the FX
  compiler into the node registry through an import cycle.

Decisions of the user: one chain for the whole (1a), FX everywhere
asked for (2a), Set sound FX parameters on playbacks (3a), and a
pitch shifter of its own with exact latency compensation for a Pitch
on a bus.

Tests:
- Unit pitch-shifter (new): 11 green. At 0 semitones the output is
  the input 2048 samples later, off by at most 1.2e-7.
- Unit sound-library: 84 green, 8 new. Counter-checks (no early
  start, no Mix delay, chain made every time, tune misses the master)
  each turned their test red.
- Unit bus-fx (new): 6 green, 4 counter-checks red. Scheduler ahead: 1
  new, its counter-check red.
- e2e 65 (melektaus-atx): ALL PASSED, 35 checks, 4 new, first run.
  Chrome at 44.1 kHz: the bus was 0.046440 s late, the note started
  exactly that much early. Offline the shifter moved 1 kHz to 2000 Hz
  and passed a click 2048 samples later.
- Gate (melektaus-atx): CHECK EXIT 0, 259 files, 4021 tests.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-07 13:15:52 +02:00
adminandClaude Opus 5.5 f46b2316f2 Make 8-bit sounds without a file with a Chip node in instruments (#329)
game-maker / suite (push) Failing after 36s
game-maker / gate (push) Successful in 3m33s
An instrument could only play sound files so far. The Chip makes the
sound of a note itself, the way old game consoles and sfxr did: a raw
wave and what its frequency does over the note. A laser, a jump or a
pickup is now an instrument and ends by itself.

- Chip node with wave (square, triangle, saw, sine, noise and the
  periodic noise of the NES), follow note, frequency, duty, duty
  sweep, slide, delta slide, min frequency, vibrato, arpeggio and
  repeat. It counts in Hz, semitones and seconds, not sfxr's 0 to 1.
- It runs as a third processor in the FX worklet. The math is a plain
  class, ChipPlayer, tested in Node. Every sample is worked out 8
  times and averaged, as sfxr does.
- Frequency and duty are AudioParams, so an Envelope can move them
  and a parameter can tune them while the note plays. A Pitch behind
  a Chip moves its frequency directly, without the phase vocoder.
- Below its min frequency a Chip ends and tells the page. Without the
  worklet it is silent and its note ends at once.
- An Envelope with sustain 0 now ends its note once the decay is over,
  for every source under it, Samples too. Before, a note with
  duration 0 ran on silently until a Stop sound.
- A note made of Chips alone needs no file. playNote and the wait for
  the bytes no longer drop it for having none.
- Inputs that do nothing as things stand are grey in the inspector.
  idleParams is told which inputs are connected, so a connected one
  always counts.

The credit names sfxr (Tomas Pettersson, MIT) and jsfxr, which turned
out to be public domain, not MIT as the plan in #329 said.

Tests:
- Unit chip (new) 12, sound-library +4. All of tests/core,
  tests/viewport, tests/editor/graph and tests/editor/inspector:
  178 files, 2953 green.
- Counter-checks: semitones read as /24 turn 4 chip tests red. Without
  the sustain rule and the file check 4 library tests fail.
- e2e 34 (melektaus-atx): ALL PASSED, 60 checks (2 new). Offline in
  Chrome a square at 220 Hz plus 12 semitones and duty 0.25 gave 329
  periods in 0.75 s, up 0.249. 64 noise voices for 2 s rendered in
  2.0 s, about 63 at once. Live a 0.2 s note with sustain 0 was gone
  after 0.26 s.
- Gate (melektaus-atx): CHECK EXIT 0, 255 files, 3962 tests.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-06 22:54:44 +02:00
adminandClaude Opus 5.5 f1f4a57589 Cut a sound to a stretch of its file in its own Sound tab (#343)
game-maker / gate (push) Successful in 3m32s
game-maker / suite (push) Failing after 18m20s
A sound asset now opens in a tab that shows its file as a waveform.
Two handles set where the sound starts and ends in the file, and that
stretch is the sample: Play sound, a song item and an instrument's
Sample all play it and nothing else. The file itself is never changed.

- The asset has three new rows: Start (s), End (s) and Edge fade (ms).
  End 0 is the end of the file, the fade is 10 ms by default. Older
  assets have no values and play the whole file as before.
- Every way of playing asks playedSoundOf() for a name instead of the
  file: the file, or <file>#<start>~<end>~<fade> for a cut.
- SoundLibrary makes the cut from the decoded file the first time that
  name plays and ramps each edge inside the file over the fade. Length
  and peaks of a cut are read from the file without a copy.
- Cut buffers can be made again at any time, so the library lets the
  oldest go past 64 Mi samples. A voice holds its buffers itself.
- The tab draws min/max per pixel from 256-sample block peaks, and the
  samples as a line when zoomed in far. Wheel zooms, the middle button
  pans, as in the music tabs. Wave, cut and ruler sit under a mask,
  so a handle panned out of view does not draw over the panels beside.
- Dragging a handle moves that edge, dragging elsewhere spans a new
  stretch. One drag is one undo step for both edges. Whole file puts
  the handles back, Play or Space plays the stretch with a playhead.
- The inspector preview and the Sound FX preview play the cut too, and
  the song lanes draw and repeat it.

Tests:
- Unit: sounds 20, wave-view 8 (new), sound-library 68, plus 258 in
  the touched areas (instrument, song, pattern, inspector, tabs)
- e2e 34 on melektaus-atx: ALL PASSED, 57 checks (7 new)
- Gate on melektaus-atx: CHECK EXIT 0, 253 files, 3936 tests

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-06 21:42:21 +02:00
adminandClaude Opus 5.5 32051b5097 Wrap node descriptions and notes in the inspector
game-maker / gate (push) Successful in 3m23s
game-maker / suite (push) Successful in 16m24s
A node's description and every note in the inspector were one line
cut with an ellipsis. Stop pattern says what it does with no pattern
picked, but only past the cut, so nobody could read it.

- Label takes a wrap option. It breaks its text at the width it gets
  and keeps the margin a single line has, above and below all lines.
- Widget.heightFor(width) gives the height at a width. It is
  measure().h everywhere except where text wraps.
- VStack and SectionStack lay a column out with it, at the width
  each row will get.
- ScrollContainer measures again at the narrower width when its
  scrollbar shows, because the text then takes more lines.
- addNote and the empty inspector's hint wrap.
- The file list's own heightFor(count) is now heightOfRows, so the
  layout does not hand it a width as a row count.

Tests:
- Unit scroll-container: 7 green, 2 new (height for width, again
  beside the scrollbar)
- Unit ui + inspector: 17 files, 336 green (before the 2 new ones)
- e2e 02 (stein2, atx busy with suite 545): ALL PASSED, 29 checks
- Screenshot: Stop pattern shows its description in six lines

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-06 19:54:43 +02:00
adminandClaude Opus 5.5 aa097bd076 Zoom with the wheel and set the loop on its own strip in music tabs
The ruler of the Pattern and Song tabs did two things with one drag,
and the wheel did three things with Alt and Shift. Both now do one
thing each, and the Song tab gets a few edits it was missing.

- The ruler has a loop strip along its top. A drag there sets the
  loop, a press or drag below it moves only where playing starts.
- The wheel only zooms: the time over the grid, lanes and ruler, the
  rows over the keys, the tracks over their heads. Alt and Shift no
  longer change it. A middle-drag still pans.
- Tracks can be zoomed from 28 to 160 px. Below 46 px a head shows
  only its name, M and S. An old saved view gets the old height.
- A click on nothing in the lanes picks the song itself.
- A right-click on an item opens a menu with Delete instead of
  deleting it. The inspector has no delete buttons any more.
- A sound item goes up to +12 dB, a track up to +12 dB. A pattern
  item stays at 0 dB, since a note is held to unity.
- An end of an item under the pointer lights up and shows a resize
  cursor. A long item marks every point where it starts over.
- The Pattern tab has an Audition switch in its header, off at first.
  Only with it on does a note sound when drawn, grabbed or moved.
  The keys down the left sound either way.
- A looped sound item draws every round alike. Each round is drawn
  from its own start, and each column shows the loudest part of the
  sound it covers instead of one sampled bucket.
- A song is at least one bar long instead of four. An empty song
  fits that bar across, and Play song ends after the last item.
- The ghost notes held back in #334 got their own issue, #339.

Finding: suite 542 for f97961c failed 5 game checks in e2e 64 and 65
while my runs on the same Windows machine overlapped it. Modules
60-65 alone without load passed, so this push's suite has to show it.

Tests:
- Unit: song-edit 11, piano-roll 17, song 14, sound-library 67.
  Four of them are new, one in song changed.
- e2e 64+65 on melektaus-atx: ALL PASSED - 46 checks (+3).
- e2e 60-65 on melektaus-atx: ALL PASSED - 86 checks.
- Gate on melektaus-atx: 251 files, 3907 tests, CHECK EXIT 0.
- Audition, after the gate: e2e 64 alone ALL PASSED - 25 checks
  (+1), transitions and piano-roll 30 tests, typecheck and ESLint
  clean.
- Waveform and one bar, after the gate: song, song-edit, song-play
  and the node catalog 438 tests. e2e 65 alone ALL PASSED - 25
  checks (+1). A probe shot of a looped pop sound: all six rounds
  pixel for pixel alike.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-06 18:25:49 +02:00
adminandClaude Opus 5.5 8f9c250eb6 Lay patterns and sounds out on tracks and play them as a song #334
game-maker / gate (push) Successful in 3m29s
game-maker / suite (push) Successful in 16m29s
Patterns could only be heard one at a time, in their own tab or from
a behavior. A song lays patterns and sounds out on tracks along one
time axis and plays them through the scheduler of #332, the third
piece of the music tools of #330.

- New asset type song under the heading Songs, with four empty tracks,
  120 BPM in 4/4 and a grid of a bar. Tempo, meter and grid are rows
  in the inspector. Tracks, items and the loop are the asset's data,
  held to their rules on load and in every command.
- A track has a name, volume, pan, mute and solo. Each is heard on a
  bus of its own, so a change reaches what is sounding at once.
- A pattern item repeats its pattern for as long as it is, with a
  transpose and another instrument if wanted. A sound item plays its
  sound as a clip at its own speed, round and round, through a sound
  FX if wanted. Both have a volume on top of their track's.
- Cut back at the left end or split, an item starts as far into what
  it plays as the cut is: ticks into a pattern, seconds into a sound.
- The tab shows the heads with M, S, volume and pan, the ruler of the
  Pattern tab and the items, a pattern with its notes, a sound with
  its waveform, a missing asset red. Patterns and sounds are dragged
  in from the Assets panel. Body, right and left end move, lengthen
  and cut back, Alt leaves the grid. Ctrl+C/V/D/A, Ctrl+E splits
  where playing starts, Delete, arrows. A double-click opens a
  pattern item in its tab.
- What is picked is a new selection of its own, so the inspector
  shows the item or the track, with a button to delete it.
- The scheduler plans clips beside notes. Where the playhead starts
  or jumps back inside a clip, it is taken up that far into its
  sound. A clip fades out at the end of its item or of the loop.
- The editor plays one piece at a time: a pattern started stops the
  song and the other way round.
- The rules are pure functions in core/music/song, editor/song/
  songView and songEdit.

Findings:
- The suites of 4ea752f and 2a3b90b were red at e2e 54: the Pattern
  pushed Struct below the fold of the + menu. The module now scrolls
  to it, as e2e 47 does for Vector art. It also left the menu open,
  which made e2e 55 fail behind it.
- The Assets panel only dragged assets a scene can place. Patterns
  and sounds are dragged now as well, and a scene says why it does
  not take one instead of calling every asset a scene.

Tests:
- song.test.ts (new, 14) and song-edit.test.ts (new, 9): model,
  round trip, sequence of patterns and clips, mute and solo, cuts,
  clips on the scheduler, moving, sizing, trimming, splitting,
  duplicating, hits, head parts. sound-library.test.ts +5: clip
  offset, end fade, pitched clip, track bus, peaks.
- e2e 65 (new) on melektaus-atx: ALL PASSED, 18 checks. e2e 54 and
  55 with the fix: ALL PASSED, 33 checks. e2e 4 with 30 asset
  types: ALL PASSED, 62 checks.
- Gate on melektaus-atx: CHECK EXIT 0, 250 test files, 3884 tests.
- Counter-checks: no taking up of a clip, no pattern offset, no clip
  end each fail 3, 1 and 1 tests.
- Probe on stein2: the tab with four tracks, before and while
  playing, looked at in two screenshots.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-06 16:28:25 +02:00
adminandClaude Opus 5.5 8a4e4a6762 Play sounds as notes of an instrument asset #331
game-maker / suite (push) Failing after 2m40s
game-maker / gate (push) Successful in 3m14s
A sound could only be played as it was, one at a time. An instrument
is a node graph that makes a note a sound: Notes nodes bring the note
in, Samples play sound assets for it, the effects of a sound FX shape
it and the Output sums the layers. Its tab plays it from a keyboard,
a behavior with Play note. It is what patterns and songs (#330) will
play.

- New asset type instrument, under the heading Music, not placeable.
  A new one starts as Notes into a Sample into the Output.
- Notes are counted as MIDI does: 60 is C4, names like C#4, sharps
  only, the lowest C-1. Notes has From and To, outputs any, note,
  frequency and velocity, and with Single notes a trigger per note,
  each with an optional name for a drum kit.
- New port type trigger, only into the trigger of a source. Inputs
  can take any number of edges now (PortDef.many): a trigger, and
  the Output, which plays everything into it side by side.
- Sample plays a sound asset. Follow note shifts it by note - root
  through the pitch worklet of #317, so it keeps its length. Gate
  fades it out at the end of the note, One-shot plays through. Loop
  runs between Loop start and Loop end, also in the worklet.
- Parameters declares the instrument's own parameters with a
  default. Sound FX sends the sound through a sound FX asset, its
  parameters as inputs, the sound set in where its Input was.
- compileInstrument walks the graph once per note. A trigger is on
  only for the notes it stands for, a silent source silences what it
  alone feeds, a Mix with a silent side is the other side. velocity
  and the parameters stay open for the voice. Builds are cached per
  note until the graph or a sound or FX it names changes.
- The player takes a note as one voice of several layers, each way
  with its own file and volume. release fades the gated ways out,
  one-shots play on. playNote can release after a duration.
- Play note plays one note with velocity, volume, duration and
  detached, and hands out voice for Set sound FX parameters and
  Stop sound. Velocity is always loudness. Instrument parameters
  picks values for the instrument's parameters like Sound FX
  parameters does for an FX, renames follow.
- The tab has the graph, a header field per parameter, Stop, and
  128 keys below. Keys without sound are grey. A key plays while it
  is held. Focused, the computer keyboard plays as in a tracker by
  KeyboardEvent.code, the arrows move the octave.

Findings:
- A label of the keys set fill on the shared small text style and
  darkened every small label of the editor. Caught in a screenshot
  before the commit, the keys have a style of their own now.
- Suite task 523 on cf6799a failed two checks of e2e 34 that timed
  audio by the page's clock: a 2.95 s echo tail was gone after 1 s,
  and Quad in was further along than 0.7 s allow. Both time on the
  audio clock now through a new hook __gm.sound.clock().

Tests:
- instrument.test.ts (new, 22): notes, asset, palette, round trip,
  compile per note, drum kit, zones, Mix, velocity and parameters,
  Sound FX node with rename, problems, cache, Play note, the keys
- sound-library +7 (layers, release, duration, early release, loop
  region in buffer and worklet), pitch-player +2 (loop region),
  node-catalog +4, graph adjusted. 582 tests in the touched files.
- e2e 4 and 34 on melektaus-atx: ALL PASSED, 34 with 44 checks (+7)
- Gate on melektaus-atx: CHECK EXIT 0, 243 test files, 3750 tests
- Counter-checks: any always on, release ending one-shots, Sound FX
  inputs ignored, loop region ignored. 3, 2, 1 and 1 tests failed.
- Browser probe on stein2: keys, held key, Single notes names and
  the header looked at in screenshots.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-06 10:07:40 +02:00
adminandClaude Opus 5.5 297fca3e94 Show Direction and Curve of a node only where they matter
An animation key shows Direction only for one of the ten curves and
the Curve row only for Curve. Lerp and Set sound FX parameters hid
them on the node body, but the inspector still offered both rows for
every easing, so a linear Lerp had a Direction to pick that did
nothing.

- A node definition can name `idleParams`: settings that do nothing
  at the current values. The node body and the inspector both leave
  them out.
- `hiddenParams` stays as it was. It only trims the node body, the
  inspector keeps those rows (the Order of an event, the mode of
  Find), so it could not take this job.
- Lerp and Set sound FX parameters use `idleEasing` (formerly
  `hiddenEasing`) for it.
- The idle settings are part of the node signature of the inspector.
  Picking an easing that needs Direction or Curve rebuilds the
  section, and the row appears.

Tests:
- easing-curves: 14 tests (+1), node-catalog: 381, green
- e2e 06 (149 checks, green) checks that a new linear Lerp has no
  Direction and no Curve row in the inspector. Picking Cubic brings
  the Direction row, and Out is picked there as before.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-05 22:15:15 +02:00
adminandClaude Opus 5.5 183f356c99 Let a game hand values to the parameters of a sound FX #311
game-maker / gate (push) Successful in 3m10s
game-maker / suite (push) Failing after 14m15s
A sound FX played the same every time: its numbers were settings of
the nodes. Now the Input declares parameters with a default, every
number of Gain, Pitch and EQ is an input they can feed, and a Play
sound hands in values for them.

- Input declares number parameters in the inspector, each with a
  Default (DeclaredPort.default, SetPortDefaultCommand). They are
  outputs of the Input. Input and Output count as terminals now, one
  of each per graph.
- Gain, Pitch and EQ take their numbers at inputs under the keys the
  settings had, so saved graphs sound as before without a migration.
  Each value is held to its range however it arrives.
- The compiler gives an FxPlan with the parameters still open.
  fxChainOf fills them in when a voice starts. The player still only
  sees an FxChain.
- New behavior node Sound FX parameters: names an FX and picks its
  parameters like a Variables node picks variables. Each picked one is
  a number input that starts at the FX's default. Its output has the
  new port type fxparams, which only fits Play sound's fx params.
- Play sound warns when the parameters name another FX than the one
  it plays. Names the FX does not know are left out.
- Renaming a parameter at the Input renames it in every Sound FX
  parameters node naming that FX, edges and values included.
- The Sound FX tab has a number field per parameter in its header.
  Play uses those values, they are never written into the asset.
- A value counts from the start of a voice. Changing a playing voice
  is #315, a Mix node is #316.

Tests:
- soundfx.test.ts 22 (+11): parameters in the plan, held values, a
  0-by-default stage, EQ band, save with defaults, wrong port types,
  default command with undo, Sound FX parameters into Play sound,
  defaults of picked ones, node problems, mismatch, rename with undo
- node-catalog 368 (+3): Play sound fx params, Sound FX parameters
- e2e 34: ALL PASSED - 28 checks (+2: header field plays depth 5,
  a Sound FX parameters node gives the game's voice depth 7)
- Gate on melektaus-atx: typecheck, lint, prettier clean, 240 test
  files, 3636 tests green
- Counter-check: with fxChainOf ignoring the values and the rename
  not followed, 5 of the new tests fail.
- Browser probe: header fields, inspector Default and the picked
  parameter looked at in two screenshots.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-05 17:26:20 +02:00
adminandClaude Opus 5.5 a5184b3368 Let the rows of a Variables node be dragged into a new order
game-maker / gate (push) Failing after 3m4s
game-maker / suite (push) Successful in 14m57s
The Variables and Properties nodes listed their picked rows in the
inspector, but unlike Start, End and struct rows they could not be
reordered. The drag dots also sat right on the edge of the row and
touched the kind mark.

- buildPickedPortSection gathers each row with its warning line as one
  block and hands the blocks to sortBlocks. A move reorders the ports
  on the node and is one undo step. Properties uses the same section
  and gains the same reordering.
- Grip draws the dots 3 px in from the left edge and moves the kind
  mark 3 px clear of them (DOTS_PAD). The column keeps its width, so
  the name fields do not shift. Rows without dots are unchanged.

Tests:
- block-drag, reorder, variable-choices: 3 files, 13 tests, all green.
  block-drag now expects the mark 11 px in with dots, 0 without.
- e2e 23 on melektaus-atx: ALL PASSED, 32 checks (+3 for dragging vy
  before vx and undoing it). It ran before the padding change, which
  touches no checks there.
- The gate did not run before this commit.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-05 16:08:59 +02:00
adminandClaude Opus 5.5 d80d5cce26 Let a Start input or a variable name the state machine it carries
A sub-behavior that gets its machine through a Start input had no
state ports on Is in state, Set state and the other state nodes.
Their ports came only from an inline pick or a State machine literal,
so a machine handed in from outside left them empty.

- A statemachine port of a Start node can now name a machine
  (DeclaredPort.machine). The inspector shows the pick next to the
  port name. SetPortMachineCommand sets it with undo.
- machineIdOf follows an edge from such a port to that machine, so
  the state nodes it feeds offer its states.
- A Run behavior node has the machine picked in advance on that input.
- A different pick there is allowed, but the node warns, because the
  called graph asks for the states of its own machine.
- A statemachine variable of a Variables node can name a machine
  the same way, picked next to the variable's row.
- The inspector signature of a node now includes the picked machine.
  Before, the pick changed the ports but the dropdown kept its text
  until the node was selected again.
- A local machine handed in still starts in no state every call.
  That rule from #229 is unchanged.

Tests:
- tests/core/start-machine.test.ts (7): ports with and without a
  machine and undo, save and load, the preset value moving a shared
  machine through the call, the warning, a port of another kind,
  the inspector signature, a Variables node.
- Counter checks: without the new branch in machineIdOf 2 tests
  fail, without the machine in the signature 1 test fails.
- Gate on melektaus-atx: 239 files, 3599 tests, CHECK EXIT 0.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-04 18:15:56 +02:00
adminandClaude Opus 5.5 3f12a77aea Preview a sprite at the end of its inspector, placed or as an asset
A sprite's inspector listed its rows but never showed what they add up
to. The picture, its frame, nine-slice, tint and shader could only be
judged in the scene, often far from the object being edited.

- SpritePreview draws the sprite through syncSpriteMesh, the function
  the scene draws it with. Slices, padding, tint and shader come out
  the same without a second path. syncSpriteMesh now takes only the
  three fields of a DrawItem it reads.
- The subject is resolved by previewPicture from the Shader tab, so a
  placed sprite shows its own overrides.
- previewFit is the rule for size and place: whole device pixels per
  unit from 1 on, at most 8, and the corner on a whole device pixel.
  Pixel art stays sharp. A sprite larger than the group shrinks freely.
- A shader with a Time node keeps the preview redrawing once a frame
  while it is on screen. A folded group or a hidden panel stops it.
- New test hook __gm.inspectorSpritePreview().

Finding: the preview redrew on a model change in a microtask at first.
A drawn picture whose revision moved on is retried once, in a
microtask queued by the scene's synchronous redraw. That retry ran
before the preview had let go, and the old pixel art sheet stayed
waiting (e2e 48l-iii read waiting [1,1,1]). The preview now redraws
at once on a model change, as the scene does.

Tests:
- unit: 238 files, 3592 tests (+5 in preview-fit.test.ts), and the
  group list of a placed sprite in sections.test.ts ends in PREVIEW
- e2e 48: ALL PASSED, 47 checks (+2: 48l-i asset preview with
  the frame, 48l-iii placed preview with its shader values)
- counter-checks: without whole-pixel zoom 2 unit tests fail; with
  the microtask redraw 48l-iii fails with waiting [1,1,1]

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-04 14:27:35 +02:00
adminandClaude Opus 5.5 c74b6b1cab Show where an inspector row can be dragged and answer the pointer
game-maker / gate (push) Successful in 3m10s
game-maker / suite (push) Failing after 16m38s
After #310 rows could be dragged by their mark, but nothing said which
marks take a drag, and the pointer on one changed nothing but the
cursor.

- A mark that drags is now a Grip: the same widget wrapped, the same
  size, so the value column stays where it was. Once sortBlocks takes
  it, it carries six faint dots at its left edge.
- Marks that cannot drag stay plain: the variables of a placed object
  and the fields of a typed struct show no dots.
- With the pointer on a grip its dots light up and the whole block that
  would move is washed in the accent colour, a struct with its fields.
  While the block is carried the wash gives way to the dimming, and the
  grip stays lit.
- While a block is carried the other grips let the pointer pass: no
  lit dots, no wash. Every grip of the column shows grabbing instead of
  grab until the drop.
- The wash first used controlHover like list rows. In the inspector it
  is darker than the panel and vanished next to the dark fields, so it
  became the accent at 15 percent, the colour of the drop line.
- __gm.inspectorMark also reports the dots: none, idle or lit.

Tests:
- Gate on melektaus-atx: 236 files, 3576 tests (+4 in
  block-drag.test.ts, one of them a drag passing over another grip).
- e2e 22: 40 checks (+1, sum 1740). 22f-iii reads idle dots, moves the
  mouse onto the mark and reads them lit.
- Two counter-checks bite: no wash on hover fails 1 test, grips never
  made grabbable fail 3. A third: hover ignoring the drag fails 1.
- Screenshots of a lit single row and a lit struct block looked at.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-04 00:11:41 +02:00
adminandClaude Opus 5.5 2c19c89878 Let variables, fields, ports and list items be dragged into order (#310)
game-maker / gate (push) Successful in 3m10s
game-maker / suite (push) Failing after 19m34s
Variables of a behavior, the fields of a struct, the ports of a start
or end node and the items of a list could only be added and removed.
Changing their order meant removing and adding them again. Now each of
them is dragged by the mark in front of its row.

- sortBlocks (ui/blockDrag.ts) makes the rows of one column sortable.
  A block is every row it covers, so a struct moves with its fields and
  a list item with its nested list. The block is dimmed while it is
  carried, an accent line shows where it lands, and a scrolled panel
  follows the pointer at its edge. Less than 4 px stays a click.
- The arithmetic is pure and lives in ui/sortable.ts: blockGap,
  gapLineY, dropSlot and edgeScroll. dropIndex of the tile strip moved
  there too, so the reorder maths sits in one file.
- MoveVariableCommand, MoveStructFieldCommand and MovePortCommand
  build on the existing MoveCommand. Undo moves back.
- Lists need no command. InlineList replaces the list as a whole, as
  it does for every other edit of its items.
- The grip is the kind mark, for a list item its number. Moving only
  happens among siblings: a field stays in its struct.
- On a Struct asset the move runs through WithStructTypes. Typed
  variables, Pack/Expand ports and split outputs follow the new order.
- Values and edges hang on names, so only the order changes. Nodes
  running a behavior read their ports from the start node, the new
  order reaches them without a copy.
- Not sortable by design: the variables of a placed object, the fields
  of a typed struct (moved on the Struct asset) and the picked ports
  of Variables/Properties nodes.
- Test hook __gm.inspectorMark(label) finds the mark of a row.

Tests:
- Gate on melektaus-atx: 235 files, 3572 tests, CHECK EXIT 0 (+9).
- tests/core/reorder.test.ts: a variable, a field two levels deep, a
  Struct asset reaching a typed variable and Pack, and start ports
  reaching the calling node with edges and result kept.
- e2e 22: ALL PASSED, 39 checks (+4, sum 1739). 22f-iii drags a
  variable past a struct of several rows, then a field inside it.
- Counter-probes: dropSlot by top edge instead of middle and a field
  move that ignores its path fail 3 unit tests. With sortBlocks
  switched off, the three order checks of 22f-iii fail.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-03 21:45:04 +02:00
adminandClaude Opus 5.5 3e05cbbed5 Let pixel art and vector art be nine-sliced and sorted (#308)
Nine-slice and the sort line hung on the texture frame alone. Drawing
and sorting already read both off the TextureRef every picture hands
out, but the generated pictures always said "no slice, no line". So
the feature moves from the frame to the picture, without a second
copy of any rule.

- PictureCut (border, fill, sortLine) is one type. TextureFrame
  extends it, PixelArtAsset and VectorArtAsset carry it as a whole.
  The texture file format does not change, and no migration is needed.
- The rules in texture.ts take a cut and a size: cutSortLine,
  setCutSlice, rawCutSlice and the rest replace the frame* versions.
  serialize.ts reads all three through one readCut.
- cutTarget.ts resolves a target to cut and size: a texture frame, a
  pixel frame's size or the vector canvas. It lives apart from
  texture.ts because vectorart.ts reaches texture.ts itself.
- SetSliceCommand and SetSortLineCommand take that target instead of
  (assetId, frameId). They refuse an asset that draws no picture.
- generatedRef fills slice and sortLine from the drawing. The whole
  strip of a pixel art asset with several frames gets none, it is no
  picture of that size. With one frame asset and frame are the same.
- Nine-slice scale is one shared schema row, now also on pixel art and
  vector art.
- cutSection.ts holds the NINE-SLICE and SORTING rows once. The frame
  inspector and the drawing inspector both build them.
- SheetCanvas draws a drawing's lines that are set, dashed as in the
  Texture tab. They are not dragged there, the painting tools own the
  canvas (as proposed in the issue).
- drawOrder falls back to the picture's height when a ref has no
  rect, or a sort line on vector art would have been ignored.

Tests:
- Gate: 232 files, 3542 tests (+8: drawing-cut.test.ts, and vector
  art in sort-by-y.test.ts). Three counter-proofs bite: no cut on a
  drawing, a cut on the whole strip, no height fallback.
- e2e 48: 44 checks (+1, sum 1732). 48l-ii reads red and blue where
  12-pixel borders at scale 2 put them. Without borders both spots
  read green.
- Looked at: the Pixel art tab with the lines, and the sliced sprite.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-03 19:53:51 +02:00
adminandClaude Opus 5.5 808e1eefc2 Clean up duplicates and dead code, and fix the + menu check (#85)
game-maker / gate (push) Successful in 3m2s
game-maker / suite (push) Successful in 15m29s
Refactoring round 9 covers everything since round 8 (14972d0): 62
commits with about 29 000 new lines in src, api, tools and tests.
Behaviour is unchanged. The one fix is in a check: e2e 04 counted 25
asset types, and the shader function from #303 makes 26, so the run
of a526e1b failed there.

Duplicate code (jscpd over src, 50 tokens, without the two licence
texts: 35 clones / 289 lines before, 29 / 232 after):

- ShaderPicture (new, viewport/shaderPicture.ts) is the picture a
  camera's shader and a particle field's shader are laid over. It
  keeps the render texture while its size holds and frees the
  replaced one once the shader holds the new one. CameraShade and
  ParticleField.shadeField drew it the same way, line for line.
- CameraShade builds its mesh on the first draw. Before, every redraw
  of a view without a camera shader built a mesh and threw it away.
- emptyHost in inspector/host.ts builds the empty host the Inspector,
  the Variables tab and the Tilemap panel start from.
- paddedQuad(null) is the plain unit quad. symbols.ts and
  shaderPasses.ts each wrote it out by hand.
- In graph/nested.ts, calls of shader functions and On behavior
  nodes are followed by one walk, `reaches`, over ids `namedBy` a
  node type.

Dead code:

- ShaderEditor.colorFieldOf and the getter CurveCanvas.snapping had
  no reader.
- 22 values used only in their own file lose their `export`. Types
  and the node catalogues of shaders and particles stay exported.
- Kept on purpose: PhysicsSystem.holdsJoint is read by tests only,
  but it is their one view of whether a joint holds.
- A pass with stricter lint rules over every file changed since
  round 8 found no bug. PixelArtEditor.started is now readonly.

Comments and docs:

- Ten comments told history ("since #190", "used to", "any more").
  They now describe what is. The bundle is md5-identical before and
  after, and the four e2e files minify to the same bytes.
- README: joint and shader function in the asset list, joints on the
  Physics row, what the Shaders page covers, two CI runners.
- e2e 04 keeps the number of asset types once, as ASSET_TYPE_COUNT.
- screenshot.png in the README is a new picture of the editor
  (3840x2160 instead of 1825x1140), taken by the user.

Tests:
- Gate on melektaus-atx: 229 files, 3514 tests, CHECK EXIT 0
  (+3 for emptyHost, +1 for the unit quad).
- e2e 1-4, 31, 63 on melektaus-atx: ALL PASSED, 195 checks. That is
  the + menu of 04, the field shader of 31 and the camera shader of
  63, both through ShaderPicture.
- No new e2e check, so the sum stays 1728.
- Counter-checks, each put back and compared with cmp: reaches that
  never goes on fails 1 test, listensTo reading the wrong parameter
  fails 1, emptyHost dropping what a panel hands in fails 1, folding
  without a relayout fails 1.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-03 17:38:39 +02:00
adminandClaude Opus 5.5 a526e1b40f Use a shader graph as a node in other shaders (#303)
game-maker / gate (push) Successful in 3m16s
game-maker / suite (push) Failing after 14m37s
A calculation needed in several shaders (a greyscale, a noise, an
outline) had to be built again in every one of them. A shader
function keeps it as an asset of its own, called like a behavior is
since #16.

- New asset type `shaderfunction` ("Shader function"), edited in the
  Shader tab. The panel beside the graph shows no picture, the
  readout counts its inputs and outputs.
- Two new nodes: `Function input` and `Function output`. Their
  declared ports, numbers or colours only, are the ports of every
  call. One of each per function, like Start and End.
- The call is node `ShaderCall`, named after the function. The
  palette lists every function, it can be dragged in from the Assets
  panel, and a double-click opens it.
- The compiler sets the function in where it is called
  (`inlineFunctions`). It renames the copied nodes `<call>/<node>`
  and turns an unconnected port into a constant. Passes, Time, the
  Edge cut and the corner walk then see a plain graph.
- A problem inside a function is reported at the call, in the
  function's words. The function's own tab checks it on its own
  (`checkShaderFunction`).
- `ShaderLibrary` keys a shader on its graph and on every function it
  calls, however deep, so a change reaches every caller.
- The port commands mirror a new, renamed or removed port into every
  call, as for a behavior's Start and End. `NodeDef.portKinds` holds
  the two ends to Number and Color.
- As agreed in the issue: no `Property`, `Color property`, Output or
  Vertex output in a function, and `Time` is the caller's clock.
- Own decision: a `Pass` may stand in a function. It becomes a pass
  of the calling shader, so a wide blur can be a function.
- A function calling itself, also through another, says so at the
  node and is not set in. The palette does not offer such a call.
- The Shader tab showed only the compiler's problems. A call nothing
  reaches was never compiled and looked healthy, so a call now also
  shows its own problem there, as in a behavior.

Tests:
- Gate: 228 files, 3510 tests, CHECK EXIT 0 (+13: 11 in
  tests/core/shader-functions.test.ts, 2 for the new icon)
- e2e 28: 100 checks, all passed (+3 in 28l-viii). Called and in
  place both #606060/#a2a2a2, halved in the function #303030/#515151.
- Counter-check: a fingerprint without the called functions fails 1
  unit test, and in the browser the change check (#606060 stays).
- The first e2e runs found two faults, both fixed: a double-click
  only selected the function, and an unreached call of itself
  showed no problem.
- e2e total 1728 is computed (1725 + 3).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-03 17:09:53 +02:00
adminandClaude Opus 5.5 2e55071fa5 Lay a shader over a camera's whole picture (#299)
game-maker / gate (push) Successful in 3m2s
game-maker / suite (push) Successful in 14m59s
A shader so far worked only on its own wearer. Effects over the whole
game picture, such as a vignette, grey, scanlines or a red flash on a
hit, had no place to live. A camera now wears one, drawn the way a
field shader of a particle source is (#170).

- The camera gets a Shader row and, while one is picked, a Shader
  resolution row. The shader's properties follow as camera rows. They
  load, animate and appear in a Properties node under Type Camera.
- Every view keeps masks and objects in one container. CameraShade
  renders it into the camera's picture, cleared to the background,
  and draws the shader over the frame. Game tab and play.html take
  the container off the stage while the picture stands in for it.
- Screen pixels gives the picture as many pixels as it covers on
  screen. Camera pixels uses View width x View height and scales up
  without smoothing.
- A pixel counts as a camera pixel in both (the user's choice). The
  uniform uPixelScale makes Texture size read View width x View
  height, so a Blur keeps its reach at any game scale.
- Each camera wears its own shader, overlay cameras included. A red
  flash on the game camera leaves the HUD of an overlay camera alone.
- The scene view's eye has a Shader switch on the Camera row, off by
  default. With it on, the main camera's shader lies over its frame,
  below symbols and gizmos, and the world around it stays drawn.
- New tool tools/perf-camera-shader.mjs measures the cost.

Measured on stein2 (HD 630, Game tab at Fit 1310x679): a camera
shader that hands the picture through stays inside the scene's own
spread of 3 to 5 fps between runs. A Blur of 4 cost 4 to 8 fps.

Tests:
- Gate on melektaus-atx: 227 files, 3497 tests (3487 + 10).
- e2e 63 alone: 11 checks (6 + 5), e2e sum 1725 counted.
- 63e-63i: pass-through is byte-equal and upright at 100%, flat red
  reaches the corners while the overlay keeps its green, Texture
  size reads 640 x 360 in both resolutions at Fit, no shader draws
  as before, and the scene view shows it only with the switch on.
- Counter-checks: a camera left out of the shader wearers fails 1
  unit test. Counting screen pixels fails 63g with #ab6000.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-03 16:03:09 +02:00
adminandClaude Opus 5.5 b2326085ed Show each camera only its layers and lay overlay cameras on top (#302)
game-maker / gate (push) Successful in 3m2s
game-maker / suite (push) Successful in 14m26s
A camera could only show the whole scene, so a HUD drawn in the world
would run through a camera shader (#299) with everything else. Render
layers say which cameras see an object, and overlay cameras draw their
own picture over the main one.

- The project has a list of up to 16 render layers, edited in Project
  settings next to the collision layers. It is a list of its own, read
  by the same rules (`layerNamesOf`).
- Sprites, texts, tilemaps, particle sources and containers carry a row
  `Render layer` under their switches. Its default `Inherit` takes the
  parent's layer, at the top the first one.
- A camera has `Sees` (all layers by default) and `Overlay`. The main
  camera is the first one switched on without Overlay, so two plain
  cameras still draw through the first one only.
- Every view through a camera keeps only what that camera sees and lays
  one see-through view per overlay camera over itself. That view is a
  `WorldViewport` of its own, handed its view and frame from below.
- An overlay is fitted into the main camera's whole picture without
  distortion. One of the same size and zoom covers it exactly.
- `z` sorts within one camera only. The pictures lie on top of each
  other as a whole, which is what a shader per camera (#299) needs.
- The Game tab's preview now collects the scene once per redraw. Before
  it collected it twice, with the cameras it would have been five times.
- `SetCollisionLayersCommand` became `SetLayerNamesCommand`, which
  takes the list it sets.

Follow-ups: #304 (the mouse through a chosen camera) and #305 (change
the render layer at run time).

Tests:
- Gate: 226 files, 3487 tests (3480 + 7)
- e2e 63 (new): 4 checks. A green overlay sprite at z -10 covers a
  red main sprite at z 1000, in the preview and in the running game.
  A `UI` sprite in front of the main camera and a `Default` one in
  front of the overlay stay away. Without Overlay the second camera
  draws nothing. The settings dialog adds a render layer at the end.
- Counter-checks: inheriting from parents switched off fails 1 unit
  test. Overlays never drawn fails the two overlay checks in e2e 63.
- e2e total 1720 (1716 + 4), computed.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-03 15:32:52 +02:00
adminandClaude Opus 5.5 963f8a152a Open a scene framed, and thin out two inspector sections (#295)
Three small things asked for in one issue. A scene came up wherever
the view happened to stand, and the inspector carried a button and a
list that are better reached somewhere else.

- A scene comes up framed as if Fit had been clicked: after a
  project is loaded and when another scene is opened. From then on
  the view belongs to the user.
- An empty scene has nothing to frame and stands at the origin, one
  pixel per unit. A scene the running game loads is not framed anew.
- The Container tab is left out. It keeps the view of each container
  as before (#171).
- The Unpack button and the two lines under it are gone from the
  inspector of a placed container. Unpack container stays in the
  right-click menu of the hierarchy and of the scene view.
- The corners of a polygon collider are no row in the inspector any
  more. They are edited on the outline in the scene view. The new
  schema option noRow keeps a property out of the inspector while it
  stays a property.

Tests:
- Gate on melektaus-atx: 219 files, 3401 tests, CHECK EXIT 0.
- One new unit test in tests/core/polygon.test.ts: the corners are
  a property of the collider and no row.
- Three new e2e checks: a loaded project is framed like Fit (module
  03), an opened scene is (04), and after a real reload a click on
  Fit moves nothing (09).
- Module 09 unpacks through the right-click menu now and checks that
  the inspector offers no button.
- Full suite on melektaus-atx: 1695 checks, all pass. Smokes 24, 19
  and 21, all pass.
- No counter-probe evaluated, and nothing looked at after the build.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-02 23:52:49 +02:00
adminandClaude Opus 5.5 7dd3ae1131 Tie two objects with a hinge, a spring, a rope or for good (#154)
game-maker / gate (push) Successful in 3m19s
game-maker / suite (push) Failing after 11m41s
Two bodies could not be connected: no door on a hinge, no lamp on
a rope, no wheel on an axle. This is the first step of the issue,
the four joints without a motor. The slider and the motors follow
in a second step under the same issue.

- Joint is a new asset type, placed as a child of the object it
  holds, like a collider or a rigidbody. "Add joint…" in the
  object menu makes one or places one the project has. An object
  takes as many as it needs.
- Where the joint sits is the point it holds its object at. A
  hinge and a fixed joint have that one point only. A spring and a
  rope end at a connected anchor on the other object, counted in
  that object's own units, or in the scene's without one.
- The rows follow the type: length, stiffness and damping for a
  spring, length for a rope, an angle limit for a hinge. Length 0
  takes the distance the two ends were placed at. The limit counts
  from how the two were placed.
- The connected object is an object reference in the schema of the
  type. Pasting, containers and removing treat it as they treat
  every reference since #258.
- shownFor is a new schema option: a row is shown while an enum of
  the same bag holds one of the named values.
- The object needs a rigidbody, its own or one further up. Without
  one the joint holds nothing and the console says so once.
- A connected object without a rigidbody is a moving anchor: the
  joint has a kinematic body of its own that follows it.
- A connected object destroyed while playing makes the joint let go
  with a hint. Removed in the editor the reference empties, so the
  joint holds on to the scene. The issue asked for "inactive" there.
- "Add joint" ties two objects while the game runs, from a joint
  asset as the template. Destroy on the joint lets go, there is no
  Break joint node. Destroy with keep children takes the joints
  along with the object.
- The scene view draws a line from each joint to what it holds on
  to and rings the held point. A selected spring or rope has a
  round handle on its far end, and dragging it is one undo step.
- Joints hold in Play only. The editor run simulates nothing
  (#148), and the inspector of a placed joint says so.
- Rapier counts a revolute joint's limits absolutely, not from the
  start, so the placed angle is added. Measured with a probe.
- The + menu has 25 types now and no longer fits under its button
  on a canvas 800 high. It scrolls there, and e2e 04 scrolls to
  the last entry instead of expecting it in view.

Tests:
- Gate on melektaus-atx for this state: 218 files, 3390 tests
  (3336 + 54), exit 0.
- e2e 1-16,24,39-46,51,52,57,58 on melektaus-atx: 831 checks, all
  pass. Module 41 alone: 25 checks (21 + 4).
- 54 new unit tests: 40 in joints.test.ts (rope, spring, pendulum,
  limit, fixed, the other end, the node, containers, paste), 6 for
  the line the scene view draws, 4 catalog cases of the node. The
  remaining 4 come from tests that make a case per asset type.
- Four new e2e checks in module 41: Add joint from the menu, the
  rope to the hook with its line and handle, dragging the handle
  with one undo, the bob hanging in Play next to one that falls.
- e2e 04, 40 and 46 count the new type in their lists.
- 21 of 22 unit counter-probes fail their tests. The other one is
  equivalent: Rapier ignores removing a joint that left with its
  body, the guard stays.
- Three browser counter-probes fail their checks (2, 1 and 2).
- One screenshot looked at: icon, line, handle and the rope rows.
- The e2e sum of 1691 is counted (1687 + 4, 1687 measured by CI
  on e066613). The full suite and the smokes did not run.
- Not seen in the browser: spring, hinge with limit and fixed
  joint, the node Add joint, a connected object that moves, joints
  inside a container, the released game. Unit tests cover them
  but the last. Limits under a mirrored parent are not handled.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-02 21:43:12 +02:00
adminandClaude Opus 5.5 46c620ae3b Give a behavior more ways in and out, and hear it as an event (#294)
game-maker / gate (push) Successful in 2m48s
game-maker / suite (push) Failing after 13m34s
A behavior run by another one had a single flow port at its Start
and at its End, so it could not branch for its caller. And it could
not be used as an event: a call needs a pulse at its flow input,
and the events of a called graph never fire. An input like
"Confirm" (Space or Enter) had to be built again in every graph.

- A declared port of a Start or an End node can be of the kind
  Flow. It sits above the values and shares their names, exec stays
  reserved. Only these two nodes declare one, a struct does not.
- Run behavior shows a flow port of the Start as a further flow
  input. A call arriving through it leaves the Start through the
  output of that name and not through exec.
- Run behavior shows a flow port of the End as a further flow
  output. The call goes on through the one the End was reached by
  last, and then through exec. exec fires always, as before.
- On behavior is a new event node without a flow input. The
  behavior it names runs at the node for as long as the behavior
  around it runs, with all of its own events and one set of
  variables per node.
- Its outputs are those of the End. It fires once the inner pulse
  is through, by the output the End was reached by last, exec only
  for End.exec. A pulse that never reaches the End is not seen.
- The value ports of the Start are its inputs, read anew for each
  inner pulse. The Start itself never fires there.
- It nests to any depth. Under a Run behavior it stays silent. A
  behavior hearing itself, however far round, is not started and
  the node says "A behavior cannot listen to itself".
- The palette offers every other behavior with an End once more as
  "On <name>". A node naming one without End says so.
- An On node removed while the game runs ends its behavior with On
  disable and On destroy. The sounds of the entity end only with
  the behavior placed on it.
- Renaming and deleting a declared port follows the edges at On
  behavior nodes as it does at Run behavior nodes.
- The flow outputs of the Start of a behavior placed on an object
  never fire, like its exec.
- The engine prompt of the benchmark names flow ports and On
  behavior.

Tests:
- Gate on melektaus-atx right before this commit: 215 files, 3319
  tests (3278 + 41), exit 0.
- e2e 19 on melektaus-atx: 39 checks (+4), all pass. e2e
  1-6,17-24,26,36,51: 628 checks, all pass.
- 19 unit tests in flow-ports.test.ts, 19 in on-behavior.test.ts
  on a real world, the rest in the node catalog.
- Four new e2e checks in module 19: a flow port declared by mouse
  at the End, On Confirm in the graph, the call leaving through the
  further output while playing, Enter and Space on a real keyboard.
- 19 unit counter-probes all fail their tests. One did not at first
  (a removed node kept its instance), a test was added. Two browser
  counter-probes fail their checks.
- One screenshot looked at: On Confirm with exec and pressed, the
  second flow pin at the calling node.
- The e2e sum of 1680 is counted (1676 + 4), the full suite and
  the smokes did not run.
- Not seen in the browser: flow inputs at the Start, nesting, the
  ring, values through On behavior, the palette entry. No test
  holds that the sounds of the entity go on when a heard behavior
  ends.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-02 19:33:34 +02:00
adminandClaude Opus 5.5 c7dc0e4733 Leave hiding and deleting a link to its right-click menu
game-maker / gate (push) Successful in 2m51s
game-maker / suite (push) Successful in 14m57s
The inspector of a picked link carried the buttons Hide link and
Delete link below its facts. Both actions are in the right-click
menu of the line since #289, and Delete removes a picked link as
well. The user asked for the buttons to go.

- buildEdgeSection in editor/inspector/edgeSection.ts only
  describes now: the two ends, the type, and for a transition its
  name field.
- The button under several picked links and Delete transition in a
  state machine go with them. They came from the same line of
  code, and the menu offers both.
- The note under several links stays. It speaks of the Delete key.

Tests:
- Gate on melektaus-atx: 212 files, 3260 tests, exit 0.
- e2e 4,6,26 on melektaus-atx: 275 checks, all pass.
- No full run and no smokes. The sum of e2e checks stays at 1676,
  which is reckoned.
- Module 06 now checks that the inspector of a picked link shows
  neither button. Its menu check finds Hide link and Delete link
  only in the menu, where it had to take the one drawn last before.
- Module 26 checks that a picked transition shows no Delete
  transition.
- No unit test names the buttons. No counter-probe was run. The
  old checks found both texts in the inspector, which is the state
  the new ones would fail on.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-02 16:50:51 +02:00
adminandClaude Opus 5.5 9e869df063 Free a redrawn picture only once no shader names it (#293)
game-maker / gate (push) Successful in 2m48s
game-maker / suite (push) Successful in 13m44s
Editing a vector drawing made Pixi warn "destroyed while still bound
to a shader", twice with every change, once it had begun. It began
where an object wears the drawing through a shader asset. Each
change draws the picture anew and destroyed the one before it, but
the mesh of that object keeps its picture bound until it syncs. The
issue also asks for the button "Add shape" to go: a shape without
points cannot be given any, the Pen draws a new one.

- GeneratedPictures takes two more callbacks. inUse says that a
  picture cannot be freed yet, soon runs a retry once whoever asked
  is done.
- A replaced picture that is still in use waits in a list of its
  own. It is freed right after the redraw that asked, as a rule. On
  a later ask where its wearer was not part of that redraw.
- TextureLibrary tells a bound picture by the listeners for change
  on its source. Under WebGL the bind groups are the only ones that
  listen there. The retry is a microtask.
- The rule is asked for every drawn picture alike, whoever wears it.
- The inspector of a drawing and of a shape has no "Add shape" any
  more. "Delete shapes" for several selected shapes stays.

Decided here:

- Under the list of shapes a second note says "The Pen draws a new
  shape." while only the asset is selected.

Found on the way:

- Reproduced on a throwaway copy of the project the report came
  from: 6 warnings for 3 changes with the shader override on the
  object, none without it. After the change none, and no replaced
  picture is left waiting.
- The fixture project of the tests never showed the warning. No
  object there wore a drawing through a shader.

Tests:
- Gate on melektaus-atx: 211 files, 3219 tests (3209 + 10), exit 0
- e2e 47,48 on melektaus-atx: 113 checks, all pass
- e2e 28,30,31,39,40,47,48: 350 checks, all pass, run before the
  wording of the new note changed
- The first gate was red: an existing test read the wording of the
  note that had changed. Only three test files had been run after
  the change.
- No full run and no smokes. The sum of 1668 (1667 + 1) is reckoned,
  1667 is what CI counted for ec2fd90.
- Ten unit tests, five of them against the real bind group of Pixi
  in tests/viewport/bound-pictures.test.ts
- One e2e check in block 47b-ii of module 47: a drawing worn through
  a shader, three changes, no warning. The check on "Add shape" now
  reads that there is no such button.
- Nine unit counter-probes and two browser counter-probes all fail
  their tests. Without the change the new check counts 6 warnings.
- Not checked: a tilemap or a particle emitter holding a picture
  bound the same way. Nothing was looked at after the build, the
  new note included.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-02 13:03:37 +02:00
adminandClaude Opus 5.5 ec2fd9084f Say what a button without words does when the pointer rests (#209)
game-maker / gate (push) Successful in 2m50s
game-maker / suite (push) Successful in 15m34s
The top bar, the magnet, the eye and many marks of the inspector are
pictures without a caption. Whoever did not know one could only try
it. The toolkit had no hint on hover, which also forced the point
menu of #195 to be words. The answers in the issue settle the open
points: a cut tab name shows in full, shortcuts are part of the
text, the point menu stays as it is.

- Button takes a tooltip, a text or a function asked each time. It
  shows on a small sign once the pointer has rested for half a
  second.
- There is one sign for the page, a TooltipLayer on the overlay
  above the menus. A control reports each pointer move and the
  pointer leaving, the layer does the waiting.
- A press, a key, the wheel or a blur of the window takes the sign
  away. initUI listens for them on the window.
- placeTooltip in ui/tooltip.ts is the rule for the place as a pure
  function: under the control and flush left, flush right at the
  right border, above the control at the bottom.
- toolbarTip says what the five buttons of the top bar do right now.
  Undo and redo name their step from CommandStack.undoLabel and
  redoLabel, as in "Undo Paint tiles (Ctrl+Z)".
- A SortableList reports its own marks with a key per mark: the eye,
  the duplicate mark, the cross and the plus.
- The magnet, both fullscreen buttons, the eye of the scene tab, the
  overflow of the tab strip, the plus of the assets panel, the marks
  of the inspector rows and the nine anchors of the resize dialog
  carry a text.

Decided here:

- Every button whose text is cut short shows the whole text, not
  only a tab.
- A row action of the inspector names the one thing a click does. A
  button that opens a menu says only Reset, Apply or Size from the
  picture.
- A button just clicked stays quiet until the pointer has left it.
- A greyed out button shows no sign.
- The switches of the eye popup, the cross of a document tab and the
  tools of pixel art have none. The tools carry words, other than
  the issue assumes.
- The marks of a list say Show or Hide, Duplicate, Remove and Add,
  whatever the list holds.

Found on the way:

- Signing out empties the overlay and would have destroyed a sign on
  screen with it. The editor takes the tooltip away first. This has
  no test of its own.
- The text hooks locate and has skip the sign, or it would answer
  for the button under it.

Tests:
- Gate on melektaus-atx: 210 files, 3209 tests (3176 + 33), exit 0
- e2e 46 on melektaus-atx: 20 checks, all pass
- e2e 1-14 on melektaus-atx: 508 checks, all pass
- e2e 1-9,14,39,40,44-48: 685 pass, 2 fail, both in module 14. It
  measures tabs that modules 10 to 13 open, and the chain 1-14 is
  green. The cut chain was not run on the state before.
- No full run and no smokes. The sum of 1667 (1664 + 3) is reckoned,
  1664 is what CI counted for 3452899.
- 33 unit tests in tests/ui/tooltip.test.ts
- Three e2e checks in block 46h of module 46: the redo button after
  the wait, taking the sign away, turning at the right border
- 50 unit counter-probes and four browser counter-probes all fail
  their tests. Three of the unit ones did not at first and got
  sharper tests.
- Not checked in the browser: the sign of a cut tab, the marks of
  the lists, turning upwards at the bottom, each other button alone

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-02 12:02:57 +02:00
adminandClaude Opus 5.5 345289982b Scale, turn and mirror the shapes of a vector drawing (#291)
game-maker / gate (push) Successful in 2m57s
game-maker / suite (push) Successful in 14m25s
A shape could only be moved and reshaped point by point. Whoever
wanted it larger or turned pulled every point alone, and for several
shapes or a group there was no way at all. The answers in the issue
settle the rest: the change is baked into the points, nothing of it
is remembered, and the line width stays.

- With Select the selection stands in a frame of eight squares and a
  knob, drawn like the one of the scene view. It holds one shape,
  several shapes or a group, the same selection that is dragged as
  one.
- A corner square scales from the opposite corner, a square on an
  edge in its own direction. Shift keeps the proportions, Alt scales
  about the middle.
- The knob turns round the middle of the frame, with Shift in steps
  of 15 degrees. After the pull the frame stands upright again.
- The inspector has a band TRANSFORM with X, Y, Width, Height and
  Rotate by, and the buttons Flip horizontal and Flip vertical.
- Points, handles, gradients and texture tiles are taken through the
  same matrix. The rule is a pure function in vectorTransform.ts,
  transformShapesCommand makes one undo step of any number of
  shapes.
- A press goes to a point or a handle first, then to the frame, then
  to a line. shapeFrame.ts holds where the handles are and what a
  pull of each means.

Decided here:

- The frame stands 15 pixels off the shapes. The points of a shape
  lie on its bounds, so a square on the bounds would cover them.
- The knob only turns. In the scene view it scales as well.
- A side cannot be pulled through the opposite one, the frame keeps
  at least 1 pixel.
- Coordinates are rounded to a thousandth of a pixel, so a quarter
  turn is exact.
- A texture tile stays upright under a turn and grows to the bounds
  of the turned tile. The bands of a gradient stay square to its
  axis under an uneven scale.
- The switch Scale stroke of the proposal is not built.

Found on the way:

- The inspector of a shape is now taller than the window. Two checks
  of module 47 no longer found their row or button on the screen and
  use the new hook inspectorButton, which scrolls there first.
- Not solved: with a shape at the top edge of the view the knob lies
  outside the tab. The wiki names it under the limits.

Taken over from #217: the e2e check for two selected shapes without
a group, and unit tests for the snap of the common offset and for
the gradients of the shapes dragged along.

Tests:
- Gate on melektaus-atx: 209 files, 3176 tests (3121 + 55), exit 0
- e2e 47,48 on melektaus-atx: 112 checks, all pass
- No full run and no smokes. The sum of 1664 (1660 + 4) is reckoned,
  1660 is what CI counted for 6d79cf5.
- 55 unit tests: 25 in tests/core/vector-transform.test.ts and 30 in
  tests/editor/shape-frame.test.ts
- Four e2e checks in block 47k of module 47: pull a corner, turn
  with Shift, type width and rotation, drag two shapes without a
  group
- 47 unit counter-probes and four browser counter-probes all fail
  their tests. Four of the unit ones did not at first: two showed
  dead code, which is gone, two got tests.
- Not checked in the browser: the edge squares, Alt, the cursors,
  the flip buttons and the rows X, Y and Height

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-02 11:20:55 +02:00
adminandClaude Opus 5.5 2ff3a78607 Put the shapes of a vector drawing on layers and in groups (#190)
game-maker / gate (push) Successful in 2m41s
game-maker / suite (push) Failing after 7h24m17s
A vector drawing was one flat list of shapes. With many shapes the
list got hard to read, and several shapes could not be kept or moved
together. Now the shapes stand on layers, as in pixel art, and can be
put into groups that nest.

- A drawing has layers, each with a name and an eye. A hidden layer
  is neither in the tab nor in the picture a sprite wears, and its
  shapes cannot be pressed. One layer is the one worked on: the pen
  and Add shape draw onto it, and SHAPES lists what is on it.
- The inspector has a list LAYERS over the shapes. Its plus adds a
  layer, the cross on a row removes it with all on it, the last one
  stays. A double-click renames, a drag sorts.
- Group (Ctrl+G) puts the selected shapes into a group, Ungroup
  (Ctrl+Shift+G) takes it away. A group stands over its rows in the
  list, the rows one step in, and a row is dragged into a group or
  out of it.
- In the tab a press on a shape takes its outermost group, and a drag
  moves the group as one undo step.
- The model has layers with a tree of shapes and groups instead of a
  flat shapes array. The rules are pure functions in
  core/model/vectorTree.ts. A drawing saved before opens with all its
  shapes on one layer.
- SortableList takes rows with a depth and drops into a tree. The
  flat lists of tilemap and pixel art run as before.

Decided here:

- A group has no transformation of its own and cannot be hidden or
  locked. Rotating and scaling is the follow-up #291.
- A group is never empty: the one that loses its last shape goes
  with it.
- Several selected shapes are dragged together, grouped or not, and
  a click on a shape of the selection keeps the selection. This is
  what #217 asked for.
- Shift+click selects like Ctrl+click. With Select, Shift pans only
  beside the shapes, as in the scene view. The middle button pans
  anywhere, and with the pen and Paint Shift pans as before.
- A row of the list picks one shape of a group alone.
- Shapes change their layer through a row Layer in the inspector,
  which is there from two layers on. Rows are not dragged between
  layers.
- A group cannot span two layers, the button is off then.
- The pen does not draw onto a hidden layer, the hint says so.
- Groups are always shown unfolded.
- The four order buttons work inside the own group or layer.

Tests:
- Gate: 204 files, 3067 tests (3008 + 59), exit 0
- e2e 47,48: 108 checks. e2e 39,40,47,48 before the Shift change:
  209 checks. No full run, the sum of 1645 (1639 + 6) is reckoned
  and CI measures it.
- 59 unit tests: 36 in tests/core/vector-tree.test.ts, 23 in
  tests/editor/vector-layers.test.ts
- Six e2e checks with a real mouse in block 47j
- 47 unit counter-probes for layers and groups and three for the
  Shift click all fail tests, run on stein2
- Three browser counter-probes fail their checks
- Two screenshots looked at (group in the list, hidden layer)
- Only unit tested: the row Layer, sorting and removing layers,
  copying a group, the Shift pan beside the shapes

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-01 23:13:43 +02:00
adminandClaude Opus 5.5 393caadc9d Head the groups of the inspector with bands that fold (#290)
game-maker / gate (push) Successful in 2m39s
game-maker / suite (push) Successful in 13m10s
The groups of the inspector had a small grey label as their title, in
the same grid as a property row, so with many rows one group hardly
stood apart from the next. Now every group has a band as its head,
from edge to edge of the panel, and a click on it folds the group.

- A band runs across the padding of its column in the header colour,
  with the title and a mark that says whether the group is open. The
  titles stay as they are, as answered in the issue.
- A click on a band folds its group: the rows go, the band stays and
  the next band moves up against it. A control in the band (Open,
  + Variable, + Property) does its own job and folds nothing.
- A placed object has a group SWITCHES under its name, asset and
  parent: Enabled, Keep on scene load and the switches of a behavior.
  An asset has SWITCHES and PROPERTIES, and a group without rows is
  left out.
- The same bands head the tilemap panel, the Variables tab, the side
  of the Pixel art tab and the shape list of vector art. A list with a
  title (Layers, Frames, Shapes) folds with a click on its header,
  clear of its plus.
- ui/sections.ts holds the rules as pure functions, SectionStack is
  the column and SectionHeader the band. Every title of the inspector
  goes through addTitle or sectionHeader in inspector/host.ts.

Decided here:

- A fold is kept for good, in localStorage of this browser. The issue
  asked whether for the session or for good and the answer left it
  open.
- A fold is remembered by title, per panel. PROPERTIES folded once is
  folded on every object, asset and node. A number in a title does
  not count.
- A behavior asset has all its rows under SWITCHES. A Variables asset
  and a state machine get a group SCOPE.
- A greyed out control in a band takes the click too. Without that the
  click fell through to the band and folded the group.
- The Curve tab has no column of this kind and is unchanged.

Found on the way:

- The inspector of a sprite is taller than the 800 pixel window of
  the checks now. The hooks that hand out a row scroll it into view
  first, and module 28 scrolls before it looks for a row by its text.

Tests:
- Gate: 202 files, 3008 tests (2982 + 26), exit 0
- e2e: no complete run for this state. Modules 28 to 61 passed alone
  (749 checks). A full run was cut off in the accounts module at 1492
  passes without a failure, the modules after it and the smokes are
  missing. The sum of 1639 (1635 + 4) is reckoned, CI measures it.
- 26 unit tests in tests/editor/sections.test.ts
- Four e2e checks with a real mouse, three in module 02 and one in
  module 48
- 25 unit counter-probes all fail tests, run on stein2
- Screenshots of the bands looked at before the last build, none
  after it

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-01 22:08:28 +02:00
adminandClaude Opus 5.5 c5ba21533d Pick a link on its line and delete it from there (#289)
game-maker / gate (push) Successful in 2m37s
game-maker / suite (push) Successful in 13m10s
A link could only be removed through the menu of its port, and the
line itself took no input. A transition of a state machine could only
be deleted in the list of the inspector. Now the line is the handle,
in a behavior's graph and between states alike.

- A click on a line selects that link and draws a halo behind it. A
  press counts within 6 screen pixels of the line, and where two cross
  the one drawn last wins.
- Ctrl-click or Shift-click adds a link or takes it away. The rubber
  band takes every line it crosses, into the same selection as the
  nodes and boxes it caught.
- Delete or Backspace removes what is picked as one undo step. A
  right-click on a line offers Hide link and Delete link.
- The inspector shows a picked link: the node and port at each end and
  the type it carries, with Hide link and Delete link below. For
  several it says how many.
- The arrow of a transition is picked the same way, on its name too.
  Delete or a right-click with Delete transition removes it.
- The menu of a port is unchanged, Disconnect included, as answered in
  the issue.
- editor/graph/edgeHit.ts holds the hit test as pure functions, and
  editor/inspector/edgeSection.ts builds the section. The selection
  has a new kind edge, and node and annotation carry edgeIds.

Decided here:

- A hidden link has no line and is never picked this way, also while
  it is shown on loan. Its mark at the port stays its handle.
- The menu of a line speaks about the whole selection, like the menu
  of a box. With nodes or boxes held alongside it reads Delete N
  items.
- Hiding picked links lets go of them, since no line is left.
- A link wins over the box it crosses, and a mark wins over a link.
- A transition has no Hide entry.
- With nodes or boxes in the selection the inspector keeps showing
  the node or box, not the link.

Tests:
- Gate: 201 files, 2982 tests (2967 + 15), exit 0
- e2e, full run: 1635 checks passed (1629 + 6), exit 0
- Smokes: 24 / 19 / 21, exit 0
- The full run was done for exactly this state.
- 15 unit tests in tests/editor/edge-selection.test.ts
- Six e2e checks with a real mouse, block 6z-xi in module 06 and 26e7
  in module 26
- 12 unit counter-probes all fail tests, run on stein2
- Three browser counter-probes fail, the build went through each
  time: no hit on the line, no halo, a band that skips lines
- Screenshots of a picked link and a picked transition looked at
  before the last build, none after it

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-01 20:51:15 +02:00
adminandClaude Opus 5.5 1aa824ecb4 Keep objects through a Load scene with a switch (#165)
game-maker / gate (push) Successful in 2m37s
game-maker / suite (push) Successful in 13m2s
A Load scene destroyed every object of the old scene. A player, a HUD
or a music object had to be built again in each scene, with its state
pushed through shared variables. A top-level object with the new switch
Keep on scene load now travels into the next scene as the same entity,
with everything below it.

- The switch is a common property right under Enabled, off by default,
  on the asset and on a placed object. The inspector shows the row on
  top-level objects of a scene only, as answered in the issue.
- runtime/keptObjects.ts holds the rules as pure functions.
  keptEntities picks what travels, objectsToBring copies the stored
  objects as the run parents them, and bringObjects puts them into the
  next scene.
- World.instantiate takes a handed entity as it is instead of making
  one. The next syncToModel then writes the run's values into the
  scene.
- BehaviorSystem.release takes the running instances out before the
  old world is destroyed, and the next system adopts them. They hear
  no On destroy and no second On start, and their sounds play on.
- PhysicsSystem.resume gives a dynamic body the velocity and spin it
  had.
- Duplicates: the object brought along wins. The next scene does not
  build its own object with the same id, which is the way back into
  the scene the object came from.
- Decided here: the next scene does not build a top-level object of
  the same asset that is kept itself either. The issue only names the
  id. With the id alone a level that places its own player would hold
  two after a swap.
- Decided here: kept objects land behind the scene's own, the switch
  is no animation row, and collisions are found anew.

A finding, fixed along with it: Load scene did not build a scene as it
is stored, but as the run had left it. A moved object stayed moved and
a destroyed one stayed gone. The way back into a scene needs a clean
build.

- GameRun copies a scene's objects and groups before it first builds
  it and puts them back before it builds it again.
- GameRun.end puts every such scene back. The player calls it on stop,
  so spawned objects no longer stay for "Click to play again".
- The consequence: an edit made in a scene during Play is gone the
  next time that scene is loaded.

Tests:
- Gate: 199 files, 2944 tests (2929 + 15), exit 0
- e2e, full run: 1625 checks passed (1619 + 6), exit 0
- Smokes: 24 / 19 / 21, exit 0
- 15 unit tests in tests/core/keep-on-scene-load.test.ts
- Six e2e checks in module 35: the switch is set with a real mouse,
  the child has no row, the object is in the new scene, its counter
  still reads 1, its sound plays on, and the stored scenes are
  unchanged after Stop
- Eleven tests with fixed row lists only count the new schema row
- 16 unit counter-probes all fail tests, run on stein2. One did not
  fail at first (same id not left out), it has its own test now.
- The browser counter-probe fails 4 of the 6 new checks
- Not looked at in the browser: the switch as a port of the Properties
  node, and a released game with a kept object

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-01 17:30:16 +02:00
adminandClaude Opus 5.5 c4a9280140 Preview inspector edits once per frame, not per input (#274)
Scrubbing a number in the inspector and dragging a colour stop of a
gradient told the editor about every input. Pixi repeats each pointer
move shortly after, so both gestures paid for two previews per move.
The issue asked to measure first and to build only from 10 % per
move. The measurement showed about 50 %.

- previewingHandlers still writes the model on every input but
  requests the preview through a FramePreview of its own. The commit
  cancels a waiting preview first.
- The preview handler of the GradientBar in shapeSection.ts does the
  same.
- Wider than the issue describes: the row handlers also serve typing,
  the arrow keys and the colour picker. All of them preview once per
  frame now.

Measured on stein2, fixture project with 120 extra objects, 120 moves
8 ms apart, three runs each:
- Scrub: 236 project events before, 51 to 61 after. 45 to 50 ms per
  move before, 22 to 27 after.
- Colour stop: 238 project events before, 50 to 51 after. 44 to 49 ms
  per move before, 22 after.
- One preview still costs 13 to 17 ms. That price per frame is what
  remains.

Finding, not fixed here: in the pixel art tab the palette colour and
the opacity of a layer still preview on every input
(PixelArtSide.ts), and so do the head fields of the curve tab.
Neither was measured.

Tests:
- Gate: 198 files, 2907 tests (2903 + 4), exit 0
- e2e, full run: 1616 checks passed (1615 + 1), exit 0
- Smokes: 24 / 19 / 21, exit 0
- Four new unit tests in tests/editor/frame-preview.test.ts, two for
  a row and two for a colour stop
- One new e2e check in module 02: a held scrub gave 10 project events
  in 12 frames, and the value is there before the release
- Four unit counter-probes fail, 1 to 2 tests each. They ran on
  stein2, not in the container.
- The browser counter-probe fails the new check with 50 events in 16
  frames

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-01 11:25:37 +02:00
adminandClaude Opus 5.5 0f6e628616 Check the declared type when a graph writes a typed variable (#271)
A list variable with an item type only kept its shape in the editor.
A graph could write any list into it at runtime, and the List node had
no type at all. Now the declaration decides what a write may store.

- Writing a list into a typed list variable leaves out every item that
  does not have exactly the fields of the type. The rest is written and
  the console names each item that was left out, like Add item does.
- The same rule holds for typed list fields inside a struct variable
  and for lists inside typed items, at any depth.
- A typed struct variable refuses a struct with a field the type does
  not have, and a value that is no struct. A struct with missing fields
  still merges in as before, so a partial Pack keeps working.
- The check runs in the Variables node (own variables and Variables
  assets) and in the Properties node when it sets a behavior variable
  of another object.
- Ports learn their declaration through GraphLookup: what a variable
  declares about its type, and which behavior owns a graph. The rule
  itself lives in the new src/core/graph/typedValues.ts, fitsType moved
  there from SetItem.
- The List node gets a Type setting. With a type its items are shaped
  and marked, follow the struct asset, count when the type is deleted
  and offer Split into fields in For each. The mark never reaches the
  running game.
- A type on a List node holding something other than structs is
  refused with a console line.
- Setting the List node back to None removes the marks, unlike a list
  variable. A node list has no per-item type row, so marked items could
  never be freed there. Undo brings everything back.
- tests/core/kinds.test.ts looked for the value as the first parameter
  of a literal. Type now comes first on the List node, so the test
  finds the value by its key.

Tests:
- Gate: 198 files, 2869 tests (2856 + 13), CHECK EXIT 0
- e2e 55: 19 checks (+55j write at runtime, +2 for 55k List node)
- e2e 54: 16 checks, smokes 24 / 19 / 21
- Five mutations, each fails between 1 and 4 tests
- e2e total 1606 is computed (1603 + 3), no full run

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-09-30 22:08:22 +02:00
adminandClaude Opus 5.5 ded71a92c9 Pick the target of an object variable in the Scene View (#276)
game-maker / gate (push) Successful in 2m32s
game-maker / suite (push) Failing after 10m45s
Choosing the target of an object variable meant finding its name in
a dropdown. With many objects, or several with the same sprite, the
name alone does not say which one it is. The eyedropper lets you
click the object where you see it instead.

- Every object dropdown gets an eyedropper button next to it. This
  covers variables, struct fields and list entries, in the Inspector
  and in the Variables tab.
- The button opens the Scene View or the container tab the object
  lives in. A strip above the view names the field and the current
  candidate and offers Take and Cancel. The cursor is a crosshair.
- A click picks a candidate and writes nothing yet. The candidate
  gets its own cyan frame, the selection stays on the object with
  the variable.
- Clicking follows the selection rule. `pickOn` in
  `src/viewport/pick.ts` uses `pressOn` and only counts objects the
  dropdown offers. A click on empty space keeps the candidate.
- Enter or Take writes through the same handlers as the dropdown.
  That is one undo step, or none if nothing changes.
- Esc, Cancel, a right click in the view or a second click on the
  button cancel without writing. So do a tab switch, another
  selection, undo, redo, play and loading a project.
- After Take or Cancel the editor goes back to the tab the pick came
  from.
- While the eyedropper runs, the view does not drag, marquee, paint,
  show handles or open its context menu. Panning and zooming stay.

Found on the way:

- Started from the Variables tab, the click missed the object. The
  Scene View is hidden while that tab is open, and the layout skips
  hidden children. The strip appeared then, the view moved down by
  32 px but kept its old height. `Editor.display()` now lays the row
  out again after showing it.

Tests (on melektaus-atx):

- unit pick and object-pick: 27 passed, 13 new
- gate: 196 files, 2844 tests, CHECK EXIT 0. Two more come from
  svg-icons, which checks every icon twice.
- e2e 57: ALL PASSED, 12 checks, 5 new (57e to 57g)
- Counter-checks work. Without the layout fix 57g fails. Picking
  objects the dropdown does not offer fails 2 tests. Writing on
  cancel fails 2 more.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-09-30 06:31:41 +02:00
adminandClaude Opus 5.5 14972d057f Refactoring round 8: hide data files, safe saves, Esc as Cancel (#85)
game-maker / gate (push) Successful in 3m53s
game-maker / suite (push) Successful in 11m55s
Recurring cleanup round over the code since round 7 (bc688f3). The
issue now also asks to fix bugs and small inconsistencies found on the
way, so unlike the rounds before, this one changes behaviour in a few
places. Duplicated code in api went from 23 clones with 228 lines to
none (jscpd, at least 50 tokens). In src it stays at 22 clones with
271 lines, the same before and after.

Bugs:
- nginx served project and snapshot files as they were, so anyone who
  knew a project id could read it without signing in. Every file the
  API writes now goes through AtomicFile (temp file, mode, rename).
  Projects, snapshots, releases, accounts and sessions are 0640, and
  nginx answers 403 for them.
- Any signed-in account could delete a shared image, font or sound,
  which projects of other accounts may wear. Deleting one is now for
  administrators only. The editor never deleted a file, so the
  client's delete is gone.
- Two saves of one project could run at once. The lock file was
  removed after each save, so a waiting save and the next one locked
  different files. The lock file now stays, and deleting a project
  takes the lock as well.
- Bytes that are no picture got a 500 instead of a 422, because the
  error handler also threw for warnings silenced with @.
- A session file was rewritten in place. A request reading it at that
  moment found it empty and was signed out. It is replaced atomically.
- Delete did nothing for a shape picked in the vector art list. Delete
  now goes through removeSelection, one case per kind of selection,
  and a kind without a case fails the type check.

Inconsistencies:
- Escape on a question does what its Cancel does. Cancel on "Delete
  project?" and "Restore snapshot?" leads back to the picker or the
  list like the other questions did, and so does deleting a project.
- Editor texts use American spelling throughout, like the port names:
  48 British spellings in 45 lines (colour, centre, initialised).
- objectName and groupName replace eleven copies of "own name, else
  the asset's" and three of "Group". One copy fell back to the object
  id instead of "?".

Duplicates, dead code, comments:
- routeFiles serves images, fonts and sounds instead of three copies.
  HashedFileStore holds what the three stores shared: keeping a file
  under its hash with its name, listing and deleting. The stores keep
  only which bytes they take and what a record says.
- AtomicFile replaces nine copies of temp file plus rename.
- removeShapes is the one way the canvas, the inspector button and
  Delete take shapes out.
- GraphInterpreter.fire was called by tests only. The 54 calls use a
  helper in tests/helpers.ts now.
- FileStore.delete, HttpFileStore.delete and the PluginApi type were
  unused. A doc comment in actions.ts sat above the wrong function.

Checked and left as is:
- solidCells is used by tests only. It is the per-cell view before the
  merge into rectangles, the one place the largest-scale rule shows.
- An undo checker ran all 124 commands of the unit tests and compared
  the project before and after undo. The only differences are empty
  ports, groups and notes lists where there were none, which the file
  format drops anyway.

Docs:
- README: the check sum.
- Wiki, own commit: API (admin-only deletes, 422, file modes, lock
  file), Home (the "Not yet" list named three things that exist),
  Concepts, Keyboard, Vector art, Architecture and Development, and
  the 38 characters Gitea flagged as ambiguous on eight pages.
- .gitignore: the lock files, and pasted-*.png in any folder, which
  keeps the screenshot archive out.

Tests:
- ./check.sh: 195 files, 2817 tests, exit 0 (+2 for the name helpers).
- PHP: accounts 52 in the web container and 53 in the CI image with
  the race test (+2), snapshots 15 (+1), releases 20.
- Full e2e run: ALL PASSED, 1586 checks (+6 in 12, 47, 49 and 50).
- Smokes jumprun, tetris, actor: 24, 19 and 21 checks. The example
  project files were restored byte for byte afterwards.
- The file stores of HEAD and of this commit give byte-identical
  output for the live folders and for store, list and delete in a
  throwaway folder.
- Counter-checks, all fail as expected: a 500 for bad bytes fails 12,
  no admin check fails 49, removing the lock file again loses saves
  in 3 of 3 runs (151 to 170 of 321), Delete ignoring shapes fails 47,
  Escape that only closes fails 50, and dropping the shape case fails
  the type check (TS2366).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-09-29 18:12:55 +02:00
adminandClaude Opus 5.5 bc688f3efc Refactoring round 7: split BehaviorSystem, fewer copies (#85)
game-maker / gate (push) Successful in 3m50s
game-maker / suite (push) Successful in 12m3s
Recurring cleanup round over the code since round 6 (056a622), 105
commits. Behaviour is unchanged. Duplicated code in src went from 29
clones with 229 lines (0.32 %) to 14 clones with 109 lines (0.15 %),
measured with jscpd and the same settings as round 6.

- BehaviorSystem.ts shrinks from 941 to 589 lines. Collecting the
  pulses (keys, timers, collisions, animations) moved to
  behaviorPulses.ts. The NodeContext the nodes receive moved to
  behaviorContext.ts. Imports from outside stay the same, TimeScaled
  is still exported from BehaviorSystem.ts.
- PointKindButton and drawPointGrip replace the "Point:" button with
  its menu and the point and handle drawing that the curve and the
  vector art tab each had on their own.
- lerpVec, cubicAt and splitCubic in src/core/util/vec.ts replace four
  copies of the cubic curve and two de Casteljau splits.
- mapNested in Properties.ts is the one walk over lists and structs
  that liveObjects and objectRefs did separately. withFieldRenamed
  and withoutField replace three copies of renaming a struct field
  and two of removing one.
- jsonBody and staleRevision build the JSON body and turn a 409 into
  StaleRevisionError for all stores. The auth store turns a 401 into
  its error in one place.
- Smaller helpers: structAssetFrom for the two "Make struct type"
  commands, shownObject in the inspector, selectedItem for the size
  gizmo and the polygon corners, solidTiles in the tilemap collision,
  requireVariablesAsset and requireBehaviorVariable.
- Left as is on purpose: the plain forwarding from the collision and
  physics systems to Solids, the pointer wiring of the canvases and
  the identical last lines of the two make commands.
- Removed unused code: tilemapTileIds and tilemapsUsing were never
  called since #128 and #131. EditorState.runtimeProject became unused
  with GameRun in #266. The assetId getter of the make commands was
  read only by a test.
- 60 values used only in their own file lost their export. The shader
  and particle node catalogues stay exported as in round 6.
- Five comments that told history now describe the current state. The
  doc comment of BehaviorSystem sits on the class again.
- README lists the struct, clip area and curve assets and mentions the
  game link from #266.

Findings:
- Deleting a tile that is painted on a tilemap leaves its cells in
  place, and they draw nothing. The delete dialog does not mention the
  tilemaps. That was probably the job of the unused tilemapsUsing.
  Left as is, because fixing it changes behaviour.
- The solid filter for tile surfaces was untested. A bent filter
  stayed green. A new test in tests/core/surface.test.ts now fails
  against it.

Tests:
- ./check.sh: 195 files, 2815 tests, exit 0 (one test removed with
  tilemapsUsing, one added for tile surfaces).
- Full e2e run: ALL PASSED, 1580 checks.
- Smokes jumprun, tetris, actor: 24, 19 and 21 checks, all passed.
  The example project files were restored byte for byte afterwards.
- Counter-checks, 12, all fail as expected: 10 in unit tests (among
  them mapNested, splitCubic, cubicAt, staleRevision, the stay pulse,
  the late timer and the time scale), 2 in the browser (Point button
  always enabled fails 47, selectedItem without a hit fails 29).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-09-29 16:16:37 +02:00
adminandClaude Opus 5.5 89aa92d2e3 Put a shader on a text and draw pixel fonts sharp (#169)
game-maker / gate (push) Successful in 4m5s
game-maker / suite (push) Failing after 11m40s
A text could not wear a shader, and a pixel font was always smoothed.
Pixi's Text takes no custom shader. It also shares one texture between
texts of the same content and style, and the filter is not part of that
key, so two equal texts with different filters would overwrite each
other.

- TextPicture rasters a text into a texture of its own and draws it
  like a sprite. Shader, time animation and edge mode work without a
  special path. A new text, font or size rasters again and keeps the
  shader.
- The old texture is dropped only after the shader holds the new one.
  Before, Pixi warned about a destroyed texture that was still bound.
- A text has a Shader row with the shader's property rows below, as a
  sprite has. Overrides, saving, animation and the Properties node
  (type Text) know them. A shader parameter named like an own row of
  the text loses to that row, as on the particle source.
- The font asset has a Filter row, Linear or Nearest. The row existed
  three times already (texture, vector art, pixel art) and is now one
  helper filterRow.
- Nearest rasters the text 8 times finer and cuts every 8x8 block to
  one texel: at least half covered is full, else empty. Chrome on
  Windows smooths canvas text sideways through DirectWrite whatever
  filter we set later. The cut removes those half pixels on every
  system, and the texture stays one texel per world unit.

Found on the way:

- On a font that is no pixel font thin strokes break apart under
  Nearest. For a pixel font at its own size this does not matter.

Not done:

- The preview in the Shader tab does not take a selected text as its
  picture, it shows the test picture then.

Tests:

- ./check.sh: 189 files, 2759 tests, CHECK EXIT 0.
- text-shader.test.ts with 10 tests: rows and name precedence, a
  clamped override, save and load, animation, the font filter, and
  sharpCoverage on a smoothed edge and at the texture border.
- e2e 30 has 7 new checks: red glyphs through the shader picked with
  the mouse, the shader survives a new text, red in the game, soft
  edges under Linear, clean 4x4 blocks under Nearest, and half covered
  texels (Linear 238, Nearest 0).
- Full local run before the Nearest cut: all passed, 1543 checks.
  GM_STEPS=1,30 on the final bundle: all passed, 32 checks. The sum
  1544 in the README is counted from these two runs, not measured.
- Counter-checks fail as expected: in the browser no shader on a text
  fails 30h-30j, an always linear sampler fails the Nearest check, the
  old 1:1 raster leaves 107 half covered texels and fails 30k-ii.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-09-29 06:01:28 +02:00
adminandClaude Opus 5.5 d7d0a6e7d4 Let Tab reach the game in the Game tab (#261)
The game did get Tab, but the editor moved the keyboard focus to a
button at the same time. The next Space then pressed that button
instead of reaching the game. The key field of On key down could not
capture Tab at all, so an inventory on Tab could not be built.

- FocusManager asks tabBelongsElsewhere before it moves the focus.
  The editor answers yes while the Game tab is in front, so Tab stays
  with the game as long as no field has the focus.
- A widget can claim Tab through takesTab. The key capture button
  does so while it is capturing, so On key down accepts Tab.
- The harness knows the key Tab.

Found on the way:

- The first version of check 20g left the Behavior tab in front, and
  module 21 then failed in the chained run. 20g now goes back.

Tests:

- ./check.sh: 2746 tests, CHECK EXIT 0.
- e2e 20 has two new checks: Tab reaches the game and leaves the
  keyboard with it, the key field captures Tab. Both failed against
  the old bundle.
- GM_STEPS=19,20: all passed, 44 checks. The full local run with all
  three topics of this push passed with 1543 checks.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-09-29 06:01:26 +02:00
adminandClaude Opus 5.5 f84defb9b6 Pick an object of the same scene for an Object variable (#258)
game-maker / gate (push) Successful in 3m32s
game-maker / suite (push) Failing after 10m2s
An Object variable could only be set by a graph while the game runs.
A placed object always showed "(null)", although the other objects of
its scene are there to choose from and the runtime can already find
any object by its id. Now the placed object stores a reference to one
of them, and the running game turns it into the real object.

- Stored as {"$object": "<id>"}. The empty placeholder stays as it
  was, so old projects load unchanged and empty still means "(null)".
- The row of an Object variable on a placed object is a dropdown with
  None and every object of the scene, indented like the Hierarchy.
  The same goes for Object fields in structs and Object items in
  lists, also in the Variables and List tabs.
- The behavior asset keeps "(null)". An asset has no scene, so "Apply
  to asset" is gone from Object rows.
- The runtime resolves the reference when the Variables or Properties
  node reads it. A destroyed target reads as null, a missing one shows
  "(missing)" in the row.
- Inside a container only objects of the same container are offered.
  The reference keeps the template id, so every placement finds its
  own copy.
- Deleting the target clears the reference in the same undo step.
  Paste points references inside the pasted group at the copies and
  leaves outside references alone. Create container and Unpack carry
  references along with the new ids.
- A value a graph writes while running stays in the run. The scene
  keeps the picked object, also in "Run in editor".

Found on the way:

- Setting an override coerced every object slot to empty, in plain
  values, struct fields and list items. The shared objectPlace() now
  keeps the reference.

Not done:

- No eyedropper to pick in the scene, the dropdown is enough for now.
- Deleting a template object in the container editor does not clean up
  references in other scenes. They read as null there.

Tests:

- ./check.sh: 187 files, 2724 tests, CHECK EXIT 0
- e2e 1-9, 51, 57: ALL PASSED, 389 checks
- e2e 1, 57 on the final bundle: ALL PASSED, 17 checks
- 13 unit tests in object-references.test.ts, 5 checks in e2e 57
- Counter-checks: 8 unit breaks fail their tests, the browser break
  without a pick list fails 57b-57d

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-09-28 14:35:17 +02:00
adminandClaude Opus 5.5 e8b3ac0a3b Keep behavior variables private to their own graph (#257)
game-maker / gate (push) Successful in 3m40s
game-maker / suite (push) Successful in 10m56s
A behavior often needs working state such as a timer or a counter.
Until now every variable showed up on each placed object and in the
Properties menu, so the inspector filled up with values nobody should
touch. A private variable works like any other inside its graph but
is not offered anywhere outside.

- Each behavior variable gets a lock switch next to Keep. It starts
  off, so existing projects stay as they are (`privateVariables`).
- A placed object shows no private variable, neither in the
  inspector nor while the game runs. If all are private, it says so.
- Switching a variable to private drops its overrides on every
  placed object in one undo step. Switching back does not restore
  them.
- A running game does not write a private value back as an
  override, and a project change during the run does not reset it.
- `+ Property` no longer offers private variables. A port that
  names one by hand still reads and writes it without a warning.
- The default value of a private variable is read-only in the asset
  inspector. Its shape (name, fields, type, delete) stays editable.
- Keep is now a pin icon instead of a word. With the lock added, the
  header of a list variable had no room left and the lock covered
  `+ Item`.

Not done: the animation track picker still offers private variables.

Tests: ./check.sh -> 186 files, 2711 tests, CHECK EXIT 0.
e2e 1,19,20,21,56 -> ALL PASSED, 119 checks. New module 56 has
7 checks with a real mouse. Counter-checks fail as expected: unit
without the write-back filter, the sync guard, dropping overrides or
loading the flag. Browser with all variables on the object, an
unfiltered menu, or no value lock.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-09-28 13:54:57 +02:00
adminandClaude Opus 5.5 b0d0868f55 Give a list one struct type so every item keeps its shape (#254)
game-maker / gate (push) Successful in 3m28s
game-maker / suite (push) Successful in 10m38s
A list of structs used to let every item carry its own fields. Adding
an item meant typing the same fields again, and nothing kept the items
in step with the struct asset from #253. A list can now name a struct
type for its items, and the asset keeps them in shape.

- A list variable and a list field get the row Items with Mixed, every
  struct asset and Make struct type. Make struct type collects every
  field any item has, so no item loses a field.
- A typed item carries its type under the key `$struct`
  (ITEM_TYPE_KEY). listItems does not count it as a field.
- A list that holds anything but structs refuses a type, also when
  the offending item sits in an override. The console names the list
  and the count, and the list stays as it was.
- In a typed list `+ Item` adds an item with the defaults, without the
  kind menu. Field names are locked, values stay editable.
- The rows of a typed item take their kind from the struct asset
  (declaredKind in InlineList.ts). An asset field shows the asset
  picker instead of a text field, as reported in the issue.
- A struct item in a mixed list gets its own Type row. Switching a
  list back to Mixed keeps the type on each item.
- syncStructTypes brings marked items into shape and carries renames
  into them, also inside overrides and nested lists. One undo step.
- The delete dialog of a struct asset also counts typed lists.
- The game never sees the key. World and sharedState strip it with
  withoutItemTypes and bagWithoutItemTypes, so a struct stays a
  plain object at runtime.
- Add item and Set item get the setting Type. With a type they only
  accept a struct with exactly its field names (fitsType in
  SetItem.ts). Anything else is left out and logged.
- The item indent from #252 is back to its old width, the narrow one
  squeezed the value fields.

Not done: writing a variable does not check the type, and the List
node offers no type.

Found on the way: a test helper in list-of-structs.test.ts called
nodes without params. It now passes the same arguments as the
interpreter.

Tests:
- ./check.sh: 183 files, 2679 tests, CHECK EXIT 0
- e2e GM_STEPS=1,21,22,32,54,55: ALL PASSED, 150 checks
- e2e GM_STEPS=1-6,21,22,32,54,55 before the follow-up: 404 checks
- New: tests/core/typed-lists.test.ts, e2e module 55 with 13 checks
- Counter-checks bite in unit and browser, originals restored by cmp

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-09-28 07:23:00 +02:00
adminandClaude Opus 5.5 61bbb96a96 Define a struct once as an asset and reuse its shape (#253)
game-maker / gate (push) Successful in 3m30s
game-maker / suite (push) Successful in 10m33s
A struct used to declare its fields wherever it stood. The same shape
in two variables meant typing it twice, and Pack/Expand copied it once
without staying in sync. A Struct asset now holds the shape in one
place, and every struct that picks it as its type follows it.

- New asset type "Struct" (not placeable, group Structs, own icon).
  The inspector shows its fields with name, kind and default value,
  and "Open" shows them in the variables tab.
- Struct variables and struct fields get a Type row: own fields, any
  Struct asset, or "Make struct type". With a type the fields come
  from the asset and cannot be added, renamed or removed there. The
  values stay per variable, overrides on placed objects still work.
- Pack and Expand get the setting Type. With a type their ports are
  the fields of the asset and read-only in the inspector.
- A change to the asset reaches every typed place in one undo step
  (syncStructTypes). A new field arrives with its default. A removed
  field goes along with its port, edge and typed value. A renamed
  field keeps its value in variables, overrides and on ports with
  edges, also through nested types.
- Undo for these edits records plain assignments (Recorder), because
  the reach of a change is not known in advance.
- Cycles such as Loot in Loot, also through another type, are refused
  and not offered in the picker.
- Deleting the asset names how many places use it. Afterwards each
  place keeps its fields as its own.
- Loading reshapes a stale copy in a project file to its type.
- "Make struct type" creates an asset from an existing struct and
  links the place to it.

Differs from the plan:

- A changed default value applies to new fields and newly typed
  places only. Existing values stay.
- Removing a single field shows no warning. Undo brings it back.

Typed lists are left for #254. Lists stay dynamic for now.

Found by e2e: after deleting or renaming the type the inspector did
not rebuild and kept showing the old name. Fixed, the check was red
before.

Tests:
- ./check.sh: 182 files, 2661 tests, CHECK EXIT 0
- GM_STEPS=1,4,21,22,54: ALL PASSED, 176 checks
- 14 unit tests in tests/core/struct-types.test.ts
- New e2e module 54 with 14 checks, real mouse. Module 04 counts
  23 asset types.
- Counter-probes: without the rename pass 4 unit tests fail, without
  reshaping values 8, without the cycle rule 1. In the browser,
  without spreading changes and with + Field despite a type, 3
  checks fail.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-09-28 06:04:40 +02:00
adminandClaude Opus 5.5 4857bb75e6 Show lists in the inspector and variables in their own tab (#252)
game-maker / gate (push) Successful in 3m24s
game-maker / suite (push) Successful in 10m22s
The List tab could show one list at a time and was the only place to
see its items. A list now shows its items right where it lives, and a
new Variables tab shows all variables of an asset or placed object.

- InlineList draws the items of a list under its row in the inspector,
  in struct fields and in the Variables tab. A list inside a list or a
  struct field is drawn the same way, one step indented.
- The index column is the indent. Items sit under their list without an
  own margin, and the index column is 16 px wide instead of 28 px.
- Above 20 items (INLINE_LIMIT) the rest collapses into "... N more -
  open in tab". During a run the rows show the values from the game.
- ListHead puts "N items" and "+ Item" in one row. "N items" opens the
  list in a tab. A separate Open button squeezed the counter to zero
  width in the narrow inspector column.
- VariablesEditor shows every variable of a behavior asset, a variables
  asset or a placed behavior. On an asset it edits the declarations. On
  a placed object an edit only overrides that object, with the usual
  reset and push buttons.
- Opening a list inside a tab adds a new tab and keeps the open one.
  valuesKey gives each list its own document, so a list that is already
  open is shown instead of opened twice.
- Titles read "Pickup · Variables" or "Listy · nums · [3]". The tab
  closes when its asset or object is deleted.
- Lists on node inputs stay inline only, a node owns no variables.
- The List tab, ListEditor and the __gm.list hook are gone.

Tests:
- ./check.sh: 181 files, 2645 tests, CHECK EXIT 0
- GM_STEPS=1,21,22,32: ALL PASSED, 123 checks
- e2e 21 is rewritten for inline lists and the new tab, 22 and 32 are
  adapted. Sum in README: 1486 (+6 in 21, -1 in 22), counted by hand.
- Counter-probes fail as expected: sub-lists not told apart (1 unit
  test), sub-list replacing the open tab, all items drawn inline (60
  rows instead of 20), no live values (0 rows at 24 items in a run).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-09-27 21:56:21 +02:00