"Extractor can be built here" indication in the build menu overlaps the build icons #144

New issue

Closed

opened 2026-08-28 13:45:41 +00:00 by metaheavens · 1 comment

metaheavens commented

2026-08-28 13:45:41 +00:00

Owner

No description provided.

metaheavens commented

2026-08-29 10:52:03 +00:00

Author

Owner

Done in 2206b55.

The cause is a reference-frame mismatch I introduced in #137. The status line is anchored to the top of the build panel; the grid is anchored to its centre. So the gap between them is half the panel height - and the band the grid lived in was a pair of written-down constants, measured at 1280 tall. At 1080 the grid's first row started at -252 while the status line's bottom sat at -168, i.e. four pixels above it, and drew straight through.

Nothing about that geometry is written against a screen size any more. The band is measured from the header that is actually there and the panel's own height; the row pitch closes up to fit it, floored at a row's own height (which is what makes overlap impossible rather than unlikely); and the column count rises until the rows fit, so a short window gets a wide grid instead of one that runs off the bottom.

Ten structures now lay out as 5 columns at 900px tall, 4 at 1080, 3 at 1280, 2 at 1600 - clearing the header at every one. The test walks those four heights and checks both "nothing above the status line" and "nothing overlapping anything else"; the old constants fail it at 900 and 1080.

Done in `2206b55`. The cause is a reference-frame mismatch I introduced in #137. The status line is anchored to the **top** of the build panel; the grid is anchored to its **centre**. So the gap between them is half the panel height - and the band the grid lived in was a pair of written-down constants, measured at 1280 tall. At 1080 the grid's first row started at -252 while the status line's bottom sat at -168, i.e. four pixels *above* it, and drew straight through. Nothing about that geometry is written against a screen size any more. The band is measured from the header that is actually there and the panel's own height; the row pitch closes up to fit it, floored at a row's own height (which is what makes overlap impossible rather than unlikely); and the column count rises until the rows fit, so a short window gets a wide grid instead of one that runs off the bottom. Ten structures now lay out as **5 columns at 900px tall, 4 at 1080, 3 at 1280, 2 at 1600** - clearing the header at every one. The test walks those four heights and checks both "nothing above the status line" and "nothing overlapping anything else"; the old constants fail it at 900 and 1080.