33 Commits

Author	SHA1	Message	Date
Thomas Harte	313b36944f	Ensured that the final bit of a tape isn't dropped even if the tape ends exactly after it, by not posting a false pulse, being less restrictive about what can cap a bit, and in any case using a long gap as the end-of-file bookend.	2017-07-22 15:41:33 -04:00
Thomas Harte	456fdda6c2	Ensured that the mark_end step can't overwrite another recognised symbol.	2017-07-22 15:40:22 -04:00
Thomas Harte	06ea81fdb2	Made sure that unrecognised waves don't block the symbol queue, and allowed any type of nonsense to be skipped before finding a byte.	2017-07-21 20:23:26 -04:00
Thomas Harte	d69b1e2d11	Switched parsing logic to looking only for upward zero crossings, that being my new understanding of the ROM.	2017-07-21 19:39:38 -04:00
Thomas Harte	130d598ec9	Corrected some minor out-of-style breaks, and ensured that the name of ZX81 files is returned.	2017-07-17 19:52:54 -04:00
Thomas Harte	2179edc7fe	Adjusted to allow the very first thing found to be data, and ensured that unrecognised symbols break files just as gaps do.	2017-07-17 07:43:47 -04:00
Thomas Harte	c2bc34fd87	Eliminated the PLLParser class. I think the proper abstraction if and when another machine requires PLL-esque parsing is probably to template out the Acorn wiring of a PLL to a Parser, and/or generalise the Acorn shifter. It'll be easier to decide when the time comes.	2017-07-16 19:25:26 -04:00
Thomas Harte	1d3ae31755	Abstracted the concept of an Acorn shifter away from being a PLLParser. The Acorn tape parser now skips using that class and uses the shifter. The actual Electron also uses the shifter. So the two are completely aligned. Net result: the Electron should successfully load exactly when static analysis was successful.	2017-07-16 19:24:01 -04:00
Thomas Harte	e5188a60dc	Settled on the new average-of-length approach to a PLL window sizing, eliminating the old errors-of-phase approach. Since it anchors automatically to the original target clocks per bit, killed the explicit mention of a tolerance.	2017-07-16 19:03:50 -04:00
Thomas Harte	e71d13c090	With the new PLL implementation, switching to a deeper window size returns the Acorn tape parser to: working.	2017-07-16 17:12:12 -04:00
Thomas Harte	c8575fe6e0	Mild clean ups, and a tweak to permitted top and bottom phase.	2017-07-16 13:39:08 -04:00
Thomas Harte	253f9603ed	Split the normal tape parser class into two in order to add a new option: a PLL-driven tape parser. Decided to see what happens if I attempt to use that to parse CSW Acorn data.	2017-07-15 19:07:35 -04:00
Thomas Harte	648618d280	Tweaked bit timing decision.	2017-07-13 21:26:05 -04:00
Thomas Harte	2f42874fd3	Another fix to deal with real-time fighting: allow 8 and 18 pulses to be recognised as 1s and 0s. That's because the hand-off from ROM routines to parsing may occur very shortly before the first pulse of a valid sequence, making it look like there's a ghost. A cleaner solution needs to be found, probably revolving around allowing parsers to be attached to tapes and therefore to run constantly.	2017-07-06 22:33:03 -04:00
Thomas Harte	919fc48cc5	Fixed dumb out-of-bounds access error.	2017-06-22 22:28:50 -04:00
Thomas Harte	2c6414ce11	Adjusted to allow inspect_waves to swallow a gap before a bit if necessary, increasing the opportunities for its call.	2017-06-11 18:31:09 -04:00
Thomas Harte	b07af2660d	Adjusted to make sure that the very end of a tape is properly measured.	2017-06-08 21:33:35 -04:00
Thomas Harte	bc0d70b2f7	Added: a shout-out when the tape has been exhausted.	2017-06-08 21:32:27 -04:00
Thomas Harte	7e10c7f9d8	Relocated the ZX80/81 concept of a 'file' out from Tape into Data, given that it's an exact duplicate of memory.	2017-06-08 19:09:51 -04:00
Thomas Harte	c47128f433	Widened tolerances and ensured zero bits aren't prematurely discarded.	2017-06-07 17:50:03 -04:00
Thomas Harte	dbd2944c13	Took an initial run at the ZX80/81 parser.	2017-06-07 17:27:05 -04:00
Thomas Harte	4603fa6f24	Extended explicitly to support a token of lookahead, which is pretty much what was on offer anyway. Also corrected instance variable names, as per better adoption of C++ norms.	2017-06-07 17:21:57 -04:00
Thomas Harte	60300851ea	Started sketching out a tape parser for ZX80 and '81 files. I think this'll help me to verify whether the .O input is working.	2017-06-07 10:12:13 -04:00
Thomas Harte	58312ea2b7	Updated to new standardisation on curly bracket placement.	2017-06-07 10:05:43 -04:00
Thomas Harte	2807e3134f	Implemented speedy header finding. So that's half of it.	2017-05-07 20:32:48 -04:00
Thomas Harte	3f7f2c6117	'Tape' has joined the new underscore orthodoxy.	2016-12-03 12:05:19 -05:00
Thomas Harte	fce48b9b8c	I am instructed that the Oric actually catches only positive transitions, and compares the distance between those to a threshold. So here's an altered version of the tape parser that does that.	2016-11-09 06:37:10 -05:00
Thomas Harte	c257e7f58d	Implemented better sync-to-zero and discovered a header counting bug plus, probably, a misleading representation of gaps in the Oric TAP decoder.	2016-11-07 20:17:06 -05:00
Thomas Harte	80702616ea	Had a quick go at using the Oric parser for static analysis. Found out that synchronisation is lost. Need to investigate.	2016-11-06 22:56:38 -05:00
Thomas Harte	7205c3f82b	Made an attempt to build in Oric slow/fast detection.	2016-11-06 21:31:10 -05:00
Thomas Harte	d1ef2f7c63	Made an attempt to consolidate what I learnt of Oric encoding while building this hastily and untidily directly into the Oric implementation. ... (while adding support for the slow tape encoding mode)	2016-11-06 19:22:09 -05:00
Thomas Harte	353c1c8ea3	Shifted ownership of PETSCII -> string conversion down to the storage layer, where it's useful for tape parsing.	2016-11-06 18:43:51 -05:00
Thomas Harte	1b15bc3a6c	Started relocating the tape parsers down from static analyser to storage, to signify that they may be used by the emulation (if fast loading is supported on that machine).	2016-11-06 16:13:13 -05:00