Fix memory leak when element becomes inactive after src swap

[guilhermesimoes]

When an element swaps srcs, there's a brief moment in time where it has 2 different textures: this.__texture (new) and this.__displayedTexture (old). When the element becomes inactive, the hook _unsetActiveFlag gets called which does this:

Lightning/src/tree/Element.mjs

Lines 406 to 408 in 6ca16cd

	if (this.__texture) {
	this.__texture.decActiveCount();
	}

But it forgets to do the same for the old (but still displayed) texture! Since the old texture's usage count is not properly decremented, the old texture and its source appear to still be in use, and they get stuck forever in memory. Calling stage.gc() still won't release them, because the source's isUsed will always return true and allowCleanup will always return false.

[Copilot]

🟡 Changes recommended

The new ordering can corrupt texture active counts during re-entrant txUnloaded callbacks.

1 open finding

• Avoid reentrant texture state corruption during txUnloaded callback · New

What changed in this PR

Fixes texture-source retention when an element becomes inactive after swapping textures.

Changes:

• Reorders active-state handling during texture deactivation.
• Adds regression coverage for garbage-collecting the previous source.

File	Description
src/​tree/​Element.mjs	Adjusts texture deactivation ordering.
tests/​textures/​test.textures.js	Tests cleanup after a texture swap.

🧠 Review effort: Balanced

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Copilot]

🟡 Changes recommended

The unresolved synchronous txUnloaded re-entrancy can leave replacement textures counted as active after the element becomes inactive.

1 open finding

• Avoid reentrant texture state corruption during txUnloaded callback

🧠 Review effort: Balanced

[jfboeve]

@guilhermesimoes the decActiveCount is called here? so im not sure this is actually a fix..

[guilhermesimoes]

But the test that I added fails without the current change in Element 🤔 but you're right, let me take a look

[guilhermesimoes]

Ah, we're dealing with different textures! This leak only occurs when the element is swapping srcs / textures. I've updated the PR which hopefully should make the issue clearer.

[Copilot]

🟢 Approval recommended

The focused fix preserves inactive event behavior and includes suitable regression coverage.

0 open findings

1 resolved since last review

• Avoid reentrant texture state corruption during txUnloaded callback

🧠 Review effort: Balanced

[Copilot]

🟢 Approval recommended

The active-count fix preserves lifecycle ordering and the regression test covers the reported leak.

0 open findings

🧠 Review effort: Balanced

[Copilot]

🟢 Approval recommended

The fix preserves existing lifecycle ordering and directly addresses the leaked active count with regression coverage.

0 open findings

🧠 Review effort: Balanced

[Copilot]

🟢 Approval recommended

The active-count fix correctly addresses the leak and includes focused regression coverage.

0 open findings

🧠 Review effort: Balanced