1267 Commits

Author	SHA1	Message	Date
Thomas Harte	a7eb0b14f4	Always divide lines; use interlaced merely as final weighting.	2026-02-23 16:48:52 -05:00
Thomas Harte	0643a81e4f	Always alternate fields.	2026-02-23 16:47:12 -05:00
Thomas Harte	26afdf972c	Explain fields.	2026-02-23 14:20:26 -05:00
Thomas Harte	486a67586a	Include interlacing information in Field.	2026-02-22 23:16:33 -05:00
Thomas Harte	93fb57a478	Set alphas less often.	2026-02-22 09:43:13 -05:00
Thomas Harte	29fa6b21ed	Fix to-weave transition, alphas.	2026-02-21 23:19:18 -05:00
Thomas Harte	cbf5b79a5b	Attempt different frame weighting for interlaced output.	2026-02-21 22:42:51 -05:00
Thomas Harte	2cf0c4f8aa	Switch horses only when certain.	2026-02-21 22:42:22 -05:00
Thomas Harte	ea3e9a96f1	Detect and expose is-interlaced flag. ... With plenty of exposition to avoid confusing the reader into thinking that the emulator is predicated on this in an unrealistic fashion.	2026-02-21 21:17:35 -05:00
Thomas Harte	69b9aa9b57	Remove done TODO.	2026-02-19 12:31:45 -05:00
Thomas Harte	4f114d8370	Add header for std::lock_guard.	2026-02-18 16:02:44 -05:00
Thomas Harte	fa9bae55d8	Add lock for using scan_target_.	2026-02-18 14:12:05 -05:00
Thomas Harte	9b65349908	Ensure safe usage of force_horizontal_scans.load.	2026-02-18 13:01:56 -05:00
Thomas Harte	406fc85186	First blush: force always-horizontal, always.	2026-02-16 23:35:30 -05:00
Thomas Harte	717235cede	Add a delegate route for requesting exactly-horizontal scans.	2026-02-16 23:07:14 -05:00
Thomas Harte	abe844505f	Eliminate LineMetadata, redundant memory barriers.	2026-02-15 15:53:26 -05:00
Thomas Harte	0562a5aa00	Consume Frames correctly.	2026-02-15 15:49:56 -05:00
Thomas Harte	742bf46297	Attempt to consume Frames. Seems faulty.	2026-02-15 14:03:48 -05:00
Thomas Harte	dc9e9f194c	Capture completeness.	2026-02-15 13:35:32 -05:00
Thomas Harte	5abff02d56	Factor out the stuff of being a circular counter.	2026-02-15 13:10:24 -05:00
Thomas Harte	a8761bdd43	Complete relocation.	2026-02-15 13:02:10 -05:00
Thomas Harte	45602cf5f1	Relocalise PointerSet definition.	2026-02-15 11:04:36 -05:00
Thomas Harte	694ffea92d	Add justification.	2026-02-15 07:54:15 -05:00
Thomas Harte	4b9e37f163	Attempt to update frame areas.	2026-02-15 07:51:38 -05:00
Thomas Harte	f24a8ea93f	Continue baby steps towards keeping per-frame records.	2026-02-14 18:17:08 -05:00
Thomas Harte	031f00cd1e	Eliminate 'frame's from 'PointerSet'.	2026-02-14 11:56:08 -05:00
Thomas Harte	2311707764	Restore header for lock_guard, etc.	2026-02-14 11:39:54 -05:00
Thomas Harte	9a55cc1f55	Eat own dogfood on spin locks.	2026-02-14 11:21:23 -05:00
Thomas Harte	ed061fa9b3	Factor out spin lock, get a bit stricter on PointerSets.	2026-02-14 11:14:20 -05:00
Thomas Harte	6d52fdea4d	Factor out and unify line/scan frame subdivision.	2026-02-13 21:57:59 -05:00
Thomas Harte	b64da038d6	Eliminate unused Line properties.	2026-02-13 21:17:17 -05:00
Thomas Harte	17ee68f227	Stop assuming consistency of glEnables.	2026-02-10 02:15:17 +00:00
Thomas Harte	06a16a62b9	Avoid accidental quadrature clamping.	2026-02-09 20:30:22 -05:00
Thomas Harte	0d83b4a808	Overtly rebind all textures. ... This is defensive, based on prior experience of Qt, which seems to reserve the right to mutate your OpenGL state.	2026-02-09 11:03:24 -05:00
Thomas Harte	2e663e60a1	Restore real output path.	2026-02-09 14:49:28 +00:00
Thomas Harte	1e38a0000f	Be more overt in creation tests.	2026-02-09 14:47:15 +00:00
Thomas Harte	2258871159	Introduce random fill in debug builds.	2026-02-08 22:30:47 -05:00
Thomas Harte	cdddfca788	Ensure successful compilation on stricter compilers.	2026-02-08 16:23:56 +00:00
Thomas Harte	c55ded88a5	Remove done TODO.	2026-02-07 22:39:15 -05:00
Thomas Harte	17ab2e2aaf	Clean up resources.	2026-02-07 22:38:39 -05:00
Thomas Harte	2d8c4a3a2e	Copy old output buffer when resizing; remove draw.	2026-02-07 22:32:51 -05:00
Thomas Harte	92302fd0b5	Fix Luminanice8Phase8 output.	2026-02-07 21:20:56 -05:00
Thomas Harte	887ef36fec	Strip all precision modifiers. ... I'll either reestablish them when I have a broader range of testing hardware, or just leave the highp defaults as is.	2026-02-07 20:43:25 -05:00
Thomas Harte	5633f09e43	Reduce number of varyings.	2026-02-07 18:22:54 -05:00
Thomas Harte	2de6b22fec	Apply alpha during scan-based RGB output.	2026-02-07 16:05:52 -05:00
Thomas Harte	8150d283a1	Correct scan output. ... Over-saturation seems to be a current issue.	2026-02-07 15:58:24 -05:00
Thomas Harte	52075f89b2	Avoid recalculation.	2026-02-07 15:32:54 -05:00
Thomas Harte	2d8f3d8e15	Factor out repetitive circular buffer logic.	2026-02-07 15:29:10 -05:00
Thomas Harte	02283f8bfc	Further eliminate redundant code.	2026-02-07 14:51:01 -05:00
Thomas Harte	3e85d280ca	Avoid code duplication.	2026-02-07 14:46:33 -05:00