Charm chips are slowly climbing up the screen again when using the gr…

[Snagglefern]

Charm chips are slowly climbing up the screen again when using the grow upwards option. After the chips completely disappear and then come back on on the next mez, they move about the height of one chip. In the screenshot they started right above the extended target window.

---

EQBuddy 1.80.1 · Windows 26200

[DranakCorps-bot]

Found it, and you were right to say "again" — this is the same drift as #122 arriving by a completely different route, which is why the first fix didn't cover it.

What's happening. A grow-up stack has to remember its bottom edge, because its top depends on how many chips it's holding. When the last chip expires the window closes entirely, and the bottom edge gets written to settings from the Closed handler — at which point the window is already torn down and its measured height is zero. So the "bottom" that got saved was really the top edge. On the next mez the stack reopens, subtracts a chip's height from that saved value to work out where its top should be, and lands exactly one chip higher than it was.

That's why it's precisely one chip per disappear/reappear cycle rather than a random creep, and why it only bites in grow-up mode. #122 was the same symptom from restoring the saved top across reopens; this is the same idea one layer down, in the value being saved.

The fix: the anchor now remembers the bottom edge continuously while the window is alive, and that remembered value is what gets persisted — never a measurement taken from a window that no longer exists. It also refuses to record a zero height at any point, so a closing or not-yet-measured window can't collapse the anchor onto the top edge again.

Applied to the spawn-timer stack too, which had the identical code and would have drifted the same way.

Thanks for the screenshot and especially for "about the height of one chip" — an exact quantity is what turns this kind of report from "the window moves sometimes" into a specific arithmetic error with one place to look. Shipping in the next release.

[DranakCorps-bot]

Shipped in v1.82.0 — the stack should hold its parked spot now, however many times it empties and refills. Thanks again for measuring the drift; "about the height of one chip" is what made it findable.