Asset dependencies & break detection¶
The dependency graph records how your assets reference one another — a sprite is a frame of an animation, a sprite is the source image of a tileset, a tileset is used by a tilemap — and the break indicator passively warns you when changing one asset would break another that points at it. Both are reached from the &Library menu, alongside the asset library panels they build on.
What a dependency is
A dependency is a directed reference from one catalog asset to another: the referencing asset (the source) points at the referenced asset (the target). The reference also remembers the target's content at the moment it was made, so the library can later tell whether the target has changed. Dependencies form a directed acyclic graph — the library refuses to record a reference that would create a cycle (an asset cannot, directly or transitively, end up depending on itself).
Opening the dependency graph¶
The &Library menu gains a Dependency Graph dock toggle. The Dependency Graph view visualises the references for either the whole catalog or the single asset currently selected in the Asset Library panel, in two directions:
| Relation | Reads as | Shows |
|---|---|---|
| Depends on | what this asset references | the assets the selected asset points at |
| Referenced by | what references this asset | the assets that point at the selected asset |
Both relations are shown as stable, alphabetically ordered lists. By default the view shows the direct neighbours; the same relations can be read transitively (the full chain — for example every asset a tilemap ultimately draws on, through its tilesets and their source sprites).
Why the graph never hangs
Because the stored graph is always acyclic, the view only ever lists the direct neighbours the model returns — it performs no recursive walk of its own. Even a deliberately tangled set of references cannot make the view spin: a cycle is caught when the reference is recorded and reported plainly, and a very deep chain is bounded rather than followed forever.
The passive break indicator¶
When you change an asset, anything that references it may now be pointing at content that no longer matches. The library surfaces this passively — it flags the problem, it does not block your edit:
- Breaks are flagged, never blocked. You can always make the change; the indicator simply tells you that a reference is now broken so you can decide what to do. Nothing is prevented, and no modal dialog interrupts you.
- It refreshes on catalog change. The indicator re-evaluates whenever the catalog or the graph changes, so it always reflects the current state — you never have to ask it to recheck.
- It is shown in place. The Dependency Graph view shows a break summary and per- reference status, and the Asset Library list carries a Status column so a broken asset is visible right where you browse.
A reference is reported broken for one of two reasons:
| Reason | Meaning |
|---|---|
| Missing | the referenced asset is no longer in the catalog (it was removed). |
| Hash mismatch | the referenced asset is still present, but its content has changed since the reference was made. |
An asset whose references are all still valid shows no indicator; only a genuinely broken reference is flagged, so the warning stays meaningful (no false positives).
Accessibility, themes & language¶
The Dependency Graph view's controls each have an accessible name and are reachable from the keyboard, every label is fully translatable, and the view and the Status column render correctly in both the light and dark themes.
Related¶
Dependency tracking and break detection build on the asset library catalog. The companion capability — an append-only version history per asset, plus cross-project reuse, export/import and optional cloud backing — is documented in Asset versioning & cross-project reuse. A reused asset is a reference like any other, so the break indicator described here flags a reuse whose target later goes missing or changes.