295 Commits

Author	SHA1	Message	Date
Thomas Harte	e0bca1e37b	Reinstated the 16 and 32 kb Atari pagers, and ensured the 6532 always starts in a valid state.	2017-03-18 17:34:34 -04:00
Thomas Harte	55ce851bb2	Fixed types of the 8k cartridges, ensured the 6502 starts without an IRQ request history.	2017-03-18 17:04:01 -04:00
Thomas Harte	e8d34f2eb4	Having farmed out the bus, the Atari itself no longer is/owns a 6502.	2017-03-18 16:34:41 -04:00
Thomas Harte	bb3daaa99b	Sought to reintroduce the Atari 8k paging scheme, at the same time deciding to do away with the copy and paste of holding on to ROM data.	2017-03-18 15:04:01 -04:00
Thomas Harte	36b58d03b7	Formalised read bus value guarantee from the 6502, fixed missing clock signal wiring on the Atari cartridge class, reintroduced CommaVid support.	2017-03-18 14:46:46 -04:00
Thomas Harte	7958459db9	In theory unpaged cartridges should now work. Time to debug.	2017-03-18 14:01:04 -04:00
Thomas Harte	14a76af0d3	Started trying to float out bus control to cartridges.	2017-03-17 20:28:07 -04:00
Thomas Harte	afbd9fd41b	Fixed declared line length.	2017-03-14 21:33:38 -04:00
Thomas Harte	7302703039	Implemented the Activision stack paging scheme.	2017-03-14 20:24:05 -04:00
Thomas Harte	97a8a96593	Rejigged how the memory map is handled and implemented MNetwork support.	2017-03-14 20:07:54 -04:00
Thomas Harte	3b29276228	Implemented the MegaBoy paging scheme.	2017-03-14 17:40:01 -04:00
Thomas Harte	b3632a4e86	Institutes CBS RAM Plus emulation.	2017-03-14 17:25:10 -04:00
Thomas Harte	f0d944847b	Fixed setting of the second 1kb.	2017-03-11 18:16:29 -05:00
Thomas Harte	add14fb43a	Made an attempt to implement Tigervision paging.	2017-03-11 17:44:35 -05:00
Thomas Harte	33bda2d40c	Switched to image inspection for RAM guesses rather than disassembly. Which fixes the other Parker Bros titles.	2017-03-11 13:04:23 -05:00
Thomas Harte	2b5e3a600e	Made a first attempt at implementing the Parker Bros pager within the emulation.	2017-03-11 12:43:12 -05:00
Thomas Harte	9c72ce5bd2	Ensured a settling delay is permitted before an NTSC/PAL decision is made. To avoid false switches just due to startup.	2017-03-06 19:37:35 -05:00
Thomas Harte	79784a8e57	Followed tip: missiles are locked to the position after four output pixels, not four input pixels.	2017-03-04 22:23:50 -05:00
Thomas Harte	4751615623	Fixed NTSC colour cycle count, and hence the 2600's reported line lengths and phase offset.	2017-03-04 17:31:39 -05:00
Thomas Harte	d3257c345a	Tested against public ROMs and corrected. Also moved the deferred adjustment into a more canonical place.	2017-03-04 17:00:28 -05:00
Thomas Harte	837cccdf83	Switched to deferred updates for the 6532.	2017-03-04 14:58:28 -05:00
Thomas Harte	93d1573481	Added a fix for certain homebrews.	2017-03-01 07:59:25 -05:00
Thomas Harte	893a5dd007	Added an artificial low pass filter, attempting to capture post-digital effects.	2017-02-28 21:29:55 -05:00
Thomas Harte	06dd98b23c	Pulled the reset time for horizontal blank extend up to position 224.	2017-02-28 20:28:54 -05:00
Thomas Harte	184c8ae707	Extended to emulate the CommaVid.	2017-02-27 20:50:59 -05:00
Thomas Harte	9d7985c1e1	Added Super Chip emulation.	2017-02-26 17:47:29 -05:00
Thomas Harte	e6a84fd26b	Attempted a hardware-correct implementation of missile-to-player latching. This completes the last of the **knowing** inaccuracies. The rest are as-of-yet unwitting.	2017-02-26 15:12:31 -05:00
Thomas Harte	440467ea3e	Started communicating which copy is being requested.	2017-02-26 13:39:25 -05:00
Thomas Harte	98376de9ad	Started returning 'no effect' for pot ports, rather than doing nothing. Still very much TODO though.	2017-02-25 22:58:58 -05:00
Thomas Harte	e61e355251	Moved to the maximum possibly required queue length of 4. Though the emulated 2600 should never need more than 2 slots as per the current calling pattern, it's not a contractual guarantee.	2017-02-25 17:25:10 -05:00
Thomas Harte	c898c8a99e	Ensured the missiles and ball don't attempt to enqueue. Because I don't think they're supposed to.	2017-02-25 17:13:22 -05:00
Thomas Harte	8c9062857c	Added a single-slot queue for player objects to defer drawing, thereby deferring pixel lookup. Which I think is correct. Though more slots might be needed.	2017-02-25 17:10:24 -05:00
Thomas Harte	77ed4ddc05	Slightly simplified ready line release logic.	2017-02-23 21:08:32 -05:00
Thomas Harte	82f392fada	This should be the other way around. I want whichever is later.	2017-02-22 21:54:49 -05:00
Thomas Harte	8291a63d5f	Fixed loss of audio when switching to PAL.	2017-02-22 21:15:37 -05:00
Thomas Harte	6120dae61a	While I'm using the hacky approach to player/missile synchronisation, I need to seed adder.	2017-02-22 07:39:11 -05:00
Thomas Harte	4f5f191cd6	Fixed: will no longer attempt to output pixels from before the pixel part of a line on which sync was disabled abruptly.	2017-02-22 07:33:36 -05:00
Thomas Harte	21abf4e9fc	Enshrined a terminology switch, albeit without any flow change behind it.	2017-02-22 07:29:48 -05:00
Thomas Harte	144d6b70d9	Minor cleaning.	2017-02-22 07:14:30 -05:00
Thomas Harte	b769f22ca0	Switched back to the collision_buffer_ being part of the TIA object, added one more assert.	2017-02-21 22:26:20 -05:00
Thomas Harte	7019d396d0	Threw in some asserts, discovering a bug in missile positioning.	2017-02-21 22:04:27 -05:00
Thomas Harte	f4447fd9cd	Attempted to fix failure of sprites properly to wrap when performing motion.	2017-02-21 21:53:09 -05:00
Thomas Harte	36396b3d62	Made a slightly better, albeit still inaccurate, version of missile-player lock.Enough for Combat to do reasonable things.	2017-02-21 20:45:20 -05:00
Thomas Harte	d1dbf8c21f	Missile to player lock is supposed to be a toggle; also factored out the commonalities of missile and ball drawing.	2017-02-21 07:58:37 -05:00
Thomas Harte	1bde0fed6f	Simplified relationship between Objects and the usage-specific components through inheritance.	2017-02-21 07:37:20 -05:00
Thomas Harte	7ab2358bba	Made an attempt to reintroduce missiles.	2017-02-20 22:22:39 -05:00
Thomas Harte	99547181f1	Attempted to template this thing. Without yet a plan in place for pixel lookup timing.	2017-02-20 21:42:59 -05:00
Thomas Harte	2bf784535c	Simplified calllng.	2017-02-20 18:04:40 -05:00
Thomas Harte	57f434c199	Reorganised state, with an eye towards unifying object motion and triggers.	2017-02-20 17:58:28 -05:00
Thomas Harte	99a35266e1	Attempted to bring frequency-switching logic into the cross-platform realm. Which for now creates an issue with the OpenGL context.	2017-02-19 21:20:37 -05:00
Thomas Harte	51bcaea60c	Disabled incorrect 'optimisations'.	2017-02-19 12:00:04 -05:00
Thomas Harte	e00339ef0a	Attempted to reintroduce the ball.	2017-02-19 08:02:54 -05:00
Thomas Harte	53cd125712	Added stub calls to draw the missiles and ball.	2017-02-19 07:28:24 -05:00
Thomas Harte	04693b067c	Fixed failure of the optimised route to pump the pixel clock; removed optimisation entirely for now.	2017-02-18 21:36:48 -05:00
Thomas Harte	cd7876a746	Reintroduced the extra clocking delay.	2017-02-18 20:18:50 -05:00
Thomas Harte	ed5ff49ef5	Fixed vertical delay, retreated from my previous thought about adding the one extra cycle of sprite delay, at least temporarily.	2017-02-16 20:52:01 -05:00
Thomas Harte	8d502a0b03	Decided not to run before I can walk and switched to storing the motion time and next step explicitly per object.	2017-02-16 20:28:37 -05:00
Thomas Harte	5ea232310f	Added a check against negative runs.	2017-02-16 18:55:58 -05:00
Thomas Harte	09309aa74f	Attempted to prevent extraneous moves.	2017-02-16 18:52:39 -05:00
Thomas Harte	b5357860b9	Made an attempt to split things apart so as to be able to introduce the proper sprite latency.	2017-02-14 20:56:16 -05:00
Thomas Harte	cd90118a0f	Added two, extraordinarily simple tests.	2017-02-12 20:32:53 -05:00
Thomas Harte	25776de59d	I think unit testing this thing is the only way forwards. Started adding appropriate hooks.	2017-02-12 19:55:02 -05:00
Thomas Harte	600bdc9af7	In C++, I think the implicit cast to bool negates the need for any manual collapsing?	2017-02-12 18:18:35 -05:00
Thomas Harte	0c9be2b09e	Shunted the collisions buffer onto a separate area of the heap for the time being, as a debugging aid. Also added a few more initial values.	2017-02-12 18:16:50 -05:00
Thomas Harte	df8a5cbe6d	Made attempts (i) to respect the delay flag; and (ii) to account for border-region sprite clocking.	2017-02-12 17:35:09 -05:00
Thomas Harte	9ce68c38ae	Made an effort to implement proper pixel output for sprites.	2017-02-12 14:01:50 -05:00
Thomas Harte	40954d6a2a	Attempted to factor out parts I expect to reuse for missiles and the ball.	2017-02-12 11:31:17 -05:00
Thomas Harte	ac444a3f34	Corrected both position increments and target time calculation.	2017-02-11 21:24:14 -05:00
Thomas Harte	b8abeced6d	Made an attempt to introduce the proper eventful loop for player output. With debugging yet to occur.	2017-02-11 21:01:58 -05:00
Thomas Harte	aeff59addc	Implemented motion 'correctly', for programs written to do all work outside of the pixel area.	2017-02-11 20:25:49 -05:00
Thomas Harte	6381e4e1b0	All that's happened to position is that numbers have been added to it. So it can't be negative, given that it wasn't before. So a regular modulo will do.	2017-02-11 13:34:36 -05:00
Thomas Harte	905ed1f87b	Switched to the more natural type, which is also signed, making my logic less prone to error.	2017-02-11 13:16:53 -05:00
Thomas Harte	8de6caf6ff	Started trying to get into a proper structure here. Chickened out.	2017-02-11 12:59:13 -05:00
Thomas Harte	40d3f5f7f6	Attempted properly to respect start.	2017-02-11 08:26:09 -05:00
Thomas Harte	64d5712d1d	Added an incorrectly-coded version of horizontal move, at least so that I can verify that information is going into the correct slots.	2017-02-10 07:23:43 -05:00
Thomas Harte	3b20d862f0	Made an initial attempt to mark sprite positions. But without hmove implemented, they're all over the place.	2017-02-09 20:53:42 -05:00
Thomas Harte	2e9ef2b0ef	Took a shot at reinstating the horizontal blank extend flag.	2017-02-09 18:37:19 -05:00
Thomas Harte	70745286a5	Ensured this array is properly aligned for the uin32_t accesses I intend to make for background drawing.	2017-02-08 20:25:23 -05:00
Thomas Harte	dcb7584060	Added the four-cycle playfield output latency and ensured you can't get smaller-than-usual pixels by rapid register value changing.	2017-02-08 07:30:32 -05:00
Thomas Harte	a477499724	Got a bit more explicit with range returned by get_cycles_until_horizontal_blank and hence attempted a more thorough (/correct) version of WSYNC.	2017-02-07 22:14:45 -05:00
Thomas Harte	944d835eea	Switched explicitly to an accumulation model for filling the collision buffer.	2017-02-06 21:59:28 -05:00
Thomas Harte	8f5039130c	Changed index naming order to ensure no out-of-bounds accesses.	2017-02-06 21:48:41 -05:00
Thomas Harte	ba165bb70a	Made an attempt properly to populate collision registers from the collision buffer.	2017-02-06 21:15:55 -05:00
Thomas Harte	474e2e8d2c	Fixed once again to respect mid-line palette changes.	2017-02-06 20:09:12 -05:00
Thomas Harte	8b8eb787df	Fixed complete invisibility.	2017-02-06 18:42:58 -05:00
Thomas Harte	66bcdd36f3	Made an attempt to introduce an intermediate buffer that ends up with a bit mask of all graphical components present on it, and to use that to infer collision flags and colours, based on playfield priority and colour palette. Immediately yielding: a blank screen. Good work!	2017-02-06 18:29:00 -05:00
Thomas Harte	6bcf95042c	Started trying to be a bit more explicit about usage, and to divide up drawing responsibility.	2017-02-05 17:51:56 -05:00
Thomas Harte	f2437cb257	Added some additional documentation, started making steps towards returning sprites, fixed a counter bug that would exhibit as incorrect sync.	2017-01-31 20:30:32 -05:00
Thomas Harte	abe04334c2	Attempted to retain more player information, and removed the output cursor from class storage as I think it's acceptable as a temporary.	2017-01-30 22:42:27 -05:00
Thomas Harte	8545707b54	Reinstituted the playfield. Probably needs more buffering though. Time to look into delays.	2017-01-30 21:38:58 -05:00
Thomas Harte	2b08758b2b	Started capturing playfield/ball and background colours.	2017-01-30 08:08:03 -05:00
Thomas Harte	764b528891	Made a first attempt at switching to a model that respects blank and sync.	2017-01-30 07:19:19 -05:00
Thomas Harte	92754ace7a	Some mild fixes get me up to having a rolling screen of vertical lines. Which is what I was hoping for right now!	2017-01-29 22:16:23 -05:00
Thomas Harte	d51f185dc7	Made an attempt to reintroduce the basic horizontal loop.	2017-01-29 15:43:57 -05:00
Thomas Harte	2390358c24	Prevented unbounded CPU usage, albeit without yet deciding who has authority for the clock rate.	2017-01-29 14:19:26 -05:00
Thomas Harte	2432a3b4d7	Fixed condition — >= is smarter.	2017-01-29 14:00:01 -05:00
Thomas Harte	9c3597c7e3	Attempted to reintroduce enough logic to handle [most of] line timing, such that WSYNC works. Initial objective is to get back to having a working background.	2017-01-29 13:47:36 -05:00
Thomas Harte	fba6baaa9c	Stubbed and disabled to get back to building.	2017-01-28 21:56:01 -05:00
Thomas Harte	a246530953	Supposing the TIA were implemented, this is (more or less) what the Atari 2600 would now look like.	2017-01-28 21:46:40 -05:00
Thomas Harte	0ffded72a6	Created a placeholder class for a factored-out TIA. There's a bit more it'll need to do, like vending (or receiving) a CRT but this is the full hardware stuff, I think.	2017-01-28 16:19:08 -05:00
Thomas Harte	6d087ca054	Restored 2600 audio.	2017-01-25 21:29:19 -05:00
Thomas Harte	be1cb2a551	Fixed NTSC phase.	2017-01-11 21:31:24 -05:00
Thomas Harte	a549fd1ecc	Introduced the ability simply to piggy-back off the CRT's natural phase for the colour burst, thereby eliminating a couple of redundant independent attempts in the Oric and Electron.	2016-12-10 13:42:34 -05:00
Thomas Harte	ebb62a2d78	Switched the 2600 to postfix and non-camel-case instance variable names.	2016-12-03 14:07:38 -05:00
Thomas Harte	b81c058c0a	Factored out the Atari 2600's 6532 connection, as a low-hanging fruit.	2016-12-03 13:41:55 -05:00
Thomas Harte	3361d6b93a	Factored out the Atari 2600 speaker and adjusted it to postfix underscores.	2016-12-03 13:39:46 -05:00
Thomas Harte	922dd6a586	Hit up the other two kinds of audio generator.	2016-10-07 17:10:00 -04:00
Thomas Harte	ee8510984f	Added just enough wiring to restore the 2600 to functionality.	2016-09-15 19:34:45 -04:00
Thomas Harte	d9016909ed	Added some wiring for PAL/NTSC mode switching on the Vic, making an attempt to simplify the whole loop of having different clock rates.	2016-08-14 13:33:20 -04:00
Thomas Harte	8db44eed3e	Added a default implementation of Speaker::skip_samples.	2016-07-04 20:48:27 -04:00
Thomas Harte	bfa237397d	Settled on assigning to nullptr as the appropriate style for emptying shared_ptrs; fixed the Vic properly to release its output resources.	2016-07-04 20:34:11 -04:00
Thomas Harte	88964ceac0	Eliminated plain pointer passing for object types.	2016-07-04 19:33:55 -04:00
Thomas Harte	0182b0483a	Added a 'power on' flag that is set automatically at construction but can be declined. Saves all that stuff of every machine having to set and then unset the RST line, and fixes an Electron bug related to that.	2016-06-29 19:13:24 -04:00
Thomas Harte	454e98302d	Slight rearrangement plus a setting of the CSMachine delegate corrects audio after a clock rate change.	2016-06-20 22:18:41 -04:00
Thomas Harte	31b26b8118	Made a first attempt to implement the full top-to-bottom change of clock rates, giving the Atari the ability to change rate when it switches to PAL mode, as it should always have been.	2016-06-20 21:47:27 -04:00
Thomas Harte	88e2b382e5	Made an attempt at a full and thorough 6532 implementation (and got a bit more explicit about flag size in the 6502).	2016-06-20 18:57:35 -04:00
Thomas Harte	065ec618c4	Factored out and templated the 6532, finally taking the opportunity to add a means to control the on-console switches of an Atari 2600.	2016-06-19 19:36:34 -04:00
Thomas Harte	65413f078c	Factored out the 6532, eventually to make it testable.	2016-06-19 18:57:40 -04:00
Thomas Harte	00a9f1bf24	Fixed best-effort updater actually to run, finally moved clock rate announcement into the C++ side of things, started working on a mechanism to allow clock rate to be updated.	2016-06-16 20:39:46 -04:00
Thomas Harte	12d1cac751	Merge branch 'master' into Vic-20	2016-06-10 18:23:55 -04:00
Thomas Harte	805e2d15c6	Fixed timing when clock is divided.	2016-06-10 18:20:16 -04:00
Thomas Harte	581eace478	Increased logging slightly, ensured all of colour RAM can be read, slightly improved the 2600 pixel decoder.	2016-06-07 22:01:14 -04:00
Thomas Harte	604862b20b	Fixed playfield/missile and playfield/ball collisions.	2016-06-02 22:36:52 -04:00
Thomas Harte	e3b95b8d2b	Back to asynchronous updates and drawing, to try to improve guarantees on audio latency; experimenting with whether other parts of the approach are fundamentally flawed; added a broad-phase for scheduled updates on the 2600.	2016-06-02 21:22:55 -04:00
Thomas Harte	11073daee1	Issues appear to be around timing generally. Working on it.	2016-06-02 20:15:48 -04:00
Thomas Harte	7aac306a12	Quick fix to the polynomials.	2016-06-02 19:56:02 -04:00
Thomas Harte	253e5a4203	With no regard to performance whatsoever, endeavoured to bring sound to the 2600. It's very scratchy, so something is wrong.	2016-06-02 19:50:16 -04:00
Thomas Harte	1e0fcbbee8	Made a very basic stab at a couple of the tone generators, added straight-through path for the speaker when input rate exactly equals output rate.	2016-06-01 19:53:16 -04:00
Thomas Harte	40c4544fb7	Got the basic sound pipeline pumping, just enough for static-level audio seemingly to work. Berzerk VE says "intruder alert, intruder alert", anyway.	2016-06-01 19:27:04 -04:00
Thomas Harte	8dc66167be	Made an attempt to consolidate the Objective-C++ side of things based on the incoming CRTMachine::Machine.	2016-05-31 22:16:20 -04:00
Thomas Harte	5c4f35e13f	Finally started on generalising the C++ stuff so as to be able to be able to get a working audio binding on the OS-specific side without further repetition by factoring an appropriate protocol out from the Electron and sketching out the correct speaker class for the Atari. Added a method to ask it what a good output frequency would be.	2016-05-31 21:23:44 -04:00
Thomas Harte	3fc80ea01f	Rethought, hopefully perfecting, sprite timing.	2016-05-30 19:56:36 -04:00
Thomas Harte	8b342f77a9	Made an attempt further to rationalise timing.	2016-05-27 21:51:27 -04:00
Thomas Harte	91d3453cc1	Switched to looking backwards rather than forwards.	2016-05-27 14:33:08 -04:00
Thomas Harte	106ddae907	Sprites are still a pixel off but better; made some attempt to move things outside of the loop.	2016-05-25 21:43:19 -04:00
Thomas Harte	40c2c0bd2d	Minor simplifications and improvements. Still trying to figure out what's causing the non-linear edge addressing.	2016-05-25 21:12:25 -04:00
Thomas Harte	b01b474e36	I'm back to thinking that resets are deferred but pixel counts are live.	2016-05-25 07:32:25 -04:00
Thomas Harte	758806f924	Introduced a separate queue for pixels, which may or may not be correct.	2016-05-24 21:39:57 -04:00
Thomas Harte	0b081831a9	Set up a pipeline for all pixels that attempts to allow for appropriate delays.	2016-05-24 07:58:26 -04:00
Thomas Harte	5c1e594937	Realised there's a clocking delay on starting horizontal move. Which fixes Coke Zero.	2016-05-22 21:45:40 -04:00
Thomas Harte	c3e719c4ab	Added missile-to-player offsetting. Completing the list of graphics-related TODOs.	2016-05-22 17:01:56 -04:00
Thomas Harte	4c6d4d899d	Ensured that resetting the horizontal timer affects phase.	2016-05-22 16:29:53 -04:00
Thomas Harte	df93d7849d	Re-enabled the ball.	2016-05-22 14:38:14 -04:00
Thomas Harte	0242924fb4	Various bits of caveman debugging appearing and disappearing, switched to latching ball behaviour.	2016-05-22 14:26:02 -04:00
Thomas Harte	5e220562e4	Actually, the four-clock delay, with palette taking effect immediately, appears to be correct. So back to the drawing board on that.	2016-05-21 21:44:54 -04:00
Thomas Harte	4eded9b9d0	Re-enabled collisions.	2016-05-21 13:05:36 -04:00
Thomas Harte	b1c84f5402	Counters run every scan line regardless of blank and sync. Also played further with timing.	2016-05-21 12:54:39 -04:00
Thomas Harte	470c90428a	Switched to two-register implementation of VDELP, etc.	2016-05-21 10:26:27 -04:00
Thomas Harte	36d19cb6cf	Made an attempt at properly pumping timers during hmove.	2016-05-21 10:18:15 -04:00
Thomas Harte	a952813036	Fixed missile sizes, played about with ball placement.	2016-05-19 18:30:17 -04:00