6502/CLK
1
0
Fork 0
mirror of https://github.com/TomHarte/CLK.git synced 2024-11-20 15:31:10 +00:00
Code Issues Projects Releases Wiki Activity
5,361 Commits 8 Branches 121 Tags 496 MiB
0da1881a07
Commit Graph

466 Commits

Author SHA1 Message Date
Thomas Harte
0b771ce61a Removes all instances of the copyright symbol. 2018-05-13 15:19:52 -04:00
Thomas Harte
3fdefb94e4 Removes requirement for correct sector epilogues. 
It's now a length test that at present accepts 6-and-2 sectors only.
 2018-05-12 23:03:08 -04:00
Thomas Harte
49592ebaf3 Ensures initialisation of scanner and that sectors overlapping the end of track are captured. 2018-05-12 18:42:07 -04:00
Thomas Harte
f410dcb3f3 Ensures proper test: not having a number of sectors that is a multiple of the track count. 2018-05-12 18:05:33 -04:00
Thomas Harte
0cb5362c6f Disambiguates whether Step will occur in addition to below zero/beyond maximum. 2018-05-11 21:44:08 -04:00
Thomas Harte
9089bf6535 Adds step events. 2018-05-10 21:58:14 -04:00
Thomas Harte
ef19a03efc Drives can now deliver activity events. 2018-05-10 21:54:10 -04:00
Thomas Harte
8ee34fafa6 Switches default DSK volume to 254. That seems to resolve Pravetz booting issues. 2018-05-09 20:28:58 -04:00
Thomas Harte
f65c65569a Makes disk head position explicitly something with sub-integral precision. 
Also as a drive-by fix, corrects accidental assumption of 10 sectors for all MFMSectorDump descendants.
 2018-05-06 23:17:36 -04:00
Thomas Harte
36f8b165cf Makes the epilogue test a bit more thorough. 2018-05-05 20:52:42 -04:00
Thomas Harte
ba7fbc4032 Reroutes all Disk II types through the Disk II analyser and returns actual sector from the Apple GCR parser results. 2018-05-05 16:37:33 -04:00
Thomas Harte
c36d7b4972 Makes first attempt at 6 and 2 decoder. 2018-05-04 23:11:12 -04:00
Thomas Harte
0dece80b5d Improves documentation. 2018-05-04 18:02:55 -04:00
Thomas Harte
e3b4aebf1a Introduces the Disk II as a unique media target platform. 
As it makes a little more sense to analyse Apple GCR images to determine target platform than it does to have the potential platforms vote over them.

Also starts on the parser that'll be necessary for making a decision.
 2018-05-04 18:02:36 -04:00
Thomas Harte
2e20191c01 Relocates and cleans up what is currently written of Apple GCR handling as the encoder. 
A decoder will be forthcoming.
 2018-05-03 22:40:45 -04:00
Thomas Harte
fcc750784a Switches interleaving logic around: having inspected some NIBs sectors are numbered in order, but scatter read from the image. 2018-05-02 21:28:18 -04:00
Thomas Harte
3787d094ec Deals with potential precision pitfall. 2018-05-02 21:26:39 -04:00
Thomas Harte
4b4ea4a103 Corrects final two bytes of Apple GCR low nibble encoding. 2018-05-02 21:06:18 -04:00
Thomas Harte
5e34c1b6b8 Switches to producing a single segment for NIBs and DSKs. 
I've now seemingly verified that the values read back by the CPU are those I'm intending to produce, so I'm at a loss.
 2018-05-01 20:31:42 -04:00
Thomas Harte
f4097290c2 Made various corrections following a quick for-loop constness audit. 2018-04-30 22:23:57 -04:00
Thomas Harte
9da481b060 Slightly simplifies syntax. 2018-04-30 22:08:51 -04:00
Thomas Harte
79002d6962 Adds an additional assert. 2018-04-30 22:07:54 -04:00
Thomas Harte
10c0e687f5 Attempts to introduce sleeping for the Disk II. 2018-04-29 17:51:10 -04:00
Thomas Harte
a9d4fe0b41 Introduces filetype wiring for DO and PO files. 
Also corrects sector numbering logic to ensure there is a sector 15.
 2018-04-29 16:34:10 -04:00
Thomas Harte
5cd15147eb Introduces interleaving of sector numbers. 2018-04-29 16:18:14 -04:00
Thomas Harte
c62db6665a Corrects storage of lower two bit pairs. 
It turns out the non-integral result of 256/3 is handled differently than my guess.
 2018-04-29 11:20:23 -04:00
Thomas Harte
fabcb261dc Corrects data prologue usage and off-by-one error in checksum placement. 2018-04-28 23:17:06 -04:00
Thomas Harte
45cf28e0eb Corrects sync lengths. 2018-04-28 15:48:49 -04:00
Thomas Harte
5b35c88be2 Corrections: data segments now correctly announce their number of bits, and tracks aren't oversized. 2018-04-28 15:47:50 -04:00
Thomas Harte
7f03f5d02f Makes a first attempt at six-and-two encoding for DSKs. 2018-04-28 15:18:48 -04:00
Thomas Harte
b98d5b790a Finally unifies disk image file exceptions, and adds a placeholder for Apple DSK. 2018-04-27 23:18:45 -04:00
Thomas Harte
5c74044e62 Unifies constants. 2018-04-27 21:38:08 -04:00
Thomas Harte
992a99d792 Improves validation of suspected sync regions. 2018-04-27 19:53:35 -04:00
Thomas Harte
41075356e2 Makes a first attempt at NIB support. 2018-04-26 22:49:07 -04:00
Thomas Harte
d59db504a3 Adjusted stepper logic; some disks load now. 2018-04-25 21:59:18 -04:00
Thomas Harte
4c6dc597f4 Converts Time::get into a template, introduces a via-a-double fallback for the timed event loop. 2018-04-25 19:54:39 -04:00
Thomas Harte
376b26c1e4 Simplifies the rotational multiplier upon construction, to mitigate against scale issues later. 2018-04-24 20:16:14 -07:00
Thomas Harte
7061537ff5 Makes joined-up attempt to run data through the Disk II. 2018-04-24 19:44:45 -07:00
Thomas Harte
4cbe5068a9 Works further towards NIB, but still isn't close. 2018-04-23 20:01:12 -07:00
Thomas Harte
38b2302b59 Corrects minor documentation errors. 2018-04-23 19:59:19 -07:00
Thomas Harte
bce0702745 Makes some faulty steps further towards providing Apple GCR assistance. 2018-04-23 19:59:03 -07:00
Thomas Harte
d447e81abd Adds provisional support for WOZ files. 2018-04-23 19:57:45 -07:00
Thomas Harte
6592745e53 Adds the bare minimum to respond to attempts to open NIB files with an Apple II. 2018-04-21 21:21:57 -07:00
Thomas Harte
72bc5f8d7b Adds a class to contain the Disk II and begins Apple GCR conversion routines. 2018-04-21 14:33:42 -07:00
Thomas Harte
e1c4035812 Switches away from C strings and allows Vic-20 region inference from filenames. 2018-04-06 17:42:24 -04:00
Thomas Harte
3bb496f9ae Enforces a maximum sector size to avoid impossible sizes. 
Such as 128 * 2^255.
 2018-03-22 22:00:26 -04:00
Thomas Harte
aafdff49be Implements the ugly stuff of converting a DMK back to flux. 2018-01-09 22:13:04 -05:00
Thomas Harte
4ef583813a Minor tidying of PCMSegment and Oric MFM DSK. 2018-01-09 22:12:34 -05:00
Thomas Harte
9f97fb738e Merge branch 'master' into DMK 2018-01-09 19:42:27 -05:00
Thomas Harte
4e124047c6 Introduces enough DMK support to progress to failure to parse a track. 2018-01-08 21:57:11 -05:00
Thomas Harte
6eb56a1564 Corrects various comment typos. 2018-01-08 20:55:40 -05:00
Thomas Harte
35fc0a5c16
Corrects assumption of double sidedness. 2018-01-08 09:35:29 -05:00
Thomas Harte
0ccc104027 Corrects start sector and track interleaving for MSX DSK. 
MSX DSKs start with sector 1; Acorn disks still begin with sector 0. Also it turns out that MSX DSKs are indeed interleaved.
 2018-01-07 21:59:18 -05:00
Thomas Harte
2f59226300 Fixes: DiskROM drive motor control, track_for_sectors' sides. 2018-01-07 20:02:40 -05:00
Thomas Harte
999a0c22d4 Adds superficial support for MSX .DSK. 
In the sense that the file format itself is properly parsed, but the MSX doesn't actually yet have disk hardware.
 2018-01-07 16:35:57 -05:00
Thomas Harte
0eaac99d74 Avoids implicit signed/unsigned comparison in the G64 reader. 2017-11-12 17:48:11 -05:00
Thomas Harte
792061a82b Corrects warnings in the CSW, CPC DSK, ZX8081 data encoding, and PRG and binary cartridges. 2017-11-12 17:46:06 -05:00
Thomas Harte
5aef81cf24 Commutes cross-platform #pragma marks to //MARK:s. 2017-11-12 15:59:11 -05:00
Thomas Harte
bce58683fa Corrects insert explicitly to supply a shared_ptr rather than a raw one. 2017-11-11 18:22:41 -05:00
Thomas Harte
2e15fab651 Doubles down on <cX> over <X.h> for C includes, and usage of the namespace for those types and functions. 2017-11-11 15:28:40 -05:00
Thomas Harte
cb0f58ab7a Corrects order-of-initialisation errors in the CPC (again), TextureBuilder, TextureTarget, Z80, MFM parser and binary tape player. 2017-11-10 22:57:03 -05:00
Thomas Harte
d9e56711ce Corrects order-of-initialisation errors in the Amstrad CPC, Vic-20, Oric, Commodore File, MFM disk controller, UEF and Commodore tape parser. 2017-11-10 22:47:10 -05:00
Thomas Harte
fabaf4e607 Adds missing include files, corrects bad include paths and eliminates the Clang-specific __undefined. 2017-11-10 21:56:53 -05:00
Thomas Harte
c0055a5a5f Further builds up SConstruct, correcting many missed imports and a couple of improper uses of C99 in C++ code. 2017-11-09 22:04:49 -05:00
Thomas Harte
f95515ae81 Eliminates a large number of instance of end-of-line tabs. 2017-11-07 22:51:06 -05:00
Thomas Harte
23d5849cda Attempts to map recognised [M]FM errors back to FDC status codes. 2017-11-03 21:29:42 -04:00
Thomas Harte
5a3ca0e447 Adds output for modified CPC DSKs. 2017-11-03 21:10:22 -04:00
Thomas Harte
e384c50580 Switches FileHolder to have a usage much closer to FILE *. 
Thereby opens a route for file format implementations such as that appearing for CPC DSK that create an in-memory copy and perform a full rewrite.
 2017-11-02 22:32:00 -04:00
Thomas Harte
f807a6b608 Generalises the concept of multiple samplings of an FM/MFM sector, simplifying CPC DSK support and paving the way for generic weak/fuzzy bit support. 2017-10-31 21:32:28 -04:00
Thomas Harte
833f8c02a4 Switches the CPC DSK implementation to building an in-memory version of the structure up front. 
Preparatory to making these things writeable.
 2017-10-31 19:41:16 -04:00
Thomas Harte
ad9df4bb90 Commutes uint8_t *, uint16_t *, uint32_t *, size_t, off_t and long to functional-style casts. 2017-10-21 22:30:15 -04:00
Thomas Harte
e983854e71 Converts all uint8_t and uint16_t casts to the functional style. 2017-10-21 21:50:53 -04:00
Thomas Harte
ec999446e8 Commutes int and unsigned casts to the functional style. 2017-10-21 21:00:40 -04:00
Thomas Harte
5e3e91373a Switches all unsigned int and double casts to functional style. 2017-10-21 19:49:04 -04:00
Thomas Harte
7f2febeec9 Ensures complete DPLL initial state assignment. 2017-10-17 22:13:37 -04:00
Thomas Harte
2d7a4fe5f0 Switches the MFM shifter to unsigned accumulation. 
Since left shifting signed numbers is undefined behaviour.
 2017-10-17 22:12:04 -04:00
Thomas Harte
b4583e976e Corrects buffer overwrites resulting from failure to treat a number of records of 0x80 as a special case. 2017-10-17 20:52:16 -04:00
Thomas Harte
44cdc124af Switches to providing a full record of changes to disk images, rather than feeding them a track at a time. 
Gets explicit about `override`s while doing so, to ensure full adaptation.
 2017-10-07 19:37:36 -04:00
Thomas Harte
b37787a414 Ensures lifetime-linked track flushing without relying on virtual calls within a destructor. 2017-10-07 19:14:18 -04:00
Thomas Harte
53b99ea248 Uses Disk::flush_tracks to elide replacement of dirty tracks. 2017-10-06 22:07:42 -04:00
Thomas Harte
97a2be71e3 Introduces flush_tracks to Drive, while switching its interface to using Track::Address and adjusting associated integer types. 2017-10-06 21:45:12 -04:00
Thomas Harte
edb9fd301c Begins this project's conversion to functional-style casts. 2017-10-03 22:04:15 -04:00
Thomas Harte
0fb363ea0e Adds writing support for HFEs. 2017-10-03 21:24:20 -04:00
Thomas Harte
1cc85615d5 Factors HFE track seeking out from the track fetching method. 2017-10-03 20:33:55 -04:00
Thomas Harte
7b01c1bee6 Revokes direct visibility of is_read_only_ to subclasses of FileHolder. 2017-10-03 19:36:06 -04:00
Thomas Harte
35705c5345 Factors out bit reversing from the HFE class. 2017-10-03 19:12:45 -04:00
Thomas Harte
ef605eda51 Factors out commonalities in SSD/DSD and ADF implementations. 2017-09-30 20:30:15 -04:00
Thomas Harte
326857a84d Corrects FM/MFM selection when looking for sectors. 2017-09-29 22:48:00 -04:00
Thomas Harte
5dd3945695 Factors out the more egregious similarities between ADF and SSD. 2017-09-29 22:07:23 -04:00
Thomas Harte
19eb975c73 Adds an intermediate step in CP/M directory parsing. 
To reduce amount of time spent allocating and reallocating buffers.
 2017-09-29 21:38:16 -04:00
Thomas Harte
698ffca51b Recasts the [M]FM parser in terms of the new factoring. 
Temporarily breaks SSD writing support.
 2017-09-29 20:08:36 -04:00
Thomas Harte
fe3cc5c57c Removes dead pragma. 2017-09-28 20:47:25 -04:00
Thomas Harte
f488854720 Switches Oric MFM DSK serialisation to feeding a track serialisation to a shifter. 
Thereby eliminates the parser's need to offer get_track.
 2017-09-27 22:14:50 -04:00
Thomas Harte
51c0c45e04 Turns MFM bit length into a globally-available constant. 2017-09-27 21:30:09 -04:00
Thomas Harte
c3e1489a8e Introduces Track::Address, a parallel to Sector::Address to enable more uniform storage. 2017-09-27 21:29:06 -04:00
Thomas Harte
e3420f62c6 Switches the Acorn ADF implementation to using the new track_serialisation/sectors_from_segment route for decomposition of a track into sectors. 2017-09-26 22:05:33 -04:00
Thomas Harte
9f4a407f94 Switches the track serialiser to a more standard header + implementation separation. 
Also introduces a full priming of the PLL before deserialisation begins.
 2017-09-26 22:01:32 -04:00
Thomas Harte
5dda897334 Changes function name to sector_size — into line with idioms. 2017-09-26 22:00:19 -04:00
Thomas Harte
3982e375e3 Introduces a route from a PCMSegment to a list of [M]FM sectors. 2017-09-25 19:57:11 -04:00
Thomas Harte
a8524daecb Marks the move constructor as noexcept, to improve usage with vector. 2017-09-25 19:53:22 -04:00
Thomas Harte
d1ce764201 Provides SectorsFromSegment, a bitstream to sector converter. 2017-09-24 22:41:16 -04:00
Thomas Harte
8875982e1f Ensures Sectors are move constructible (and still default constructible), and adds proper const qualifiers to Sector::Address. 2017-09-24 22:40:38 -04:00
Thomas Harte
3319a4f589 Isolates those Sector fields that describe its address and makes them usable as a set key. 2017-09-24 21:57:21 -04:00
Thomas Harte
c7f27b2db4 Renames MFM.[c/h]pp as per its new remit: encoding only. 2017-09-24 21:40:43 -04:00
Thomas Harte
631f630549 Severs the MFM parser from the overweight single MFM.hpp. 2017-09-24 20:31:19 -04:00
Thomas Harte
2a08bd9ecc Factors shifting plus stateful [M]FM token recognition out of the MFMDiskController. 
Given the proliferation of MFM-related classes, establishes a subdirectory for them.
 2017-09-24 20:07:56 -04:00
Thomas Harte
f789ee4ff0 Introduces a track to segment decoder. 
This will be needed to make formats like G64 and HFE writeable, but probably also will be usable to speed up static analysis.
 2017-09-23 22:39:19 -04:00
Thomas Harte
d72dad2d1a Severs the DiskImage implementation from its public header file. 2017-09-22 22:46:31 -04:00
Thomas Harte
698e4fe550 Tidies the Disk file hierarchy. 2017-09-22 22:39:23 -04:00
Thomas Harte
b5406b90cd Introduces a new class hierarchy for disk images. 
Increasing independence of format-specific stuff and generic caching without mangling them into a common namespace, and allowing in some cases for a decrease in read/write blocking.
 2017-09-22 20:28:11 -04:00
Thomas Harte
44a56724cb Speeds up byte decoding within sectors for the ahead-of-time MFM parser. 2017-09-16 20:28:24 -04:00
Thomas Harte
a85909198f Adds defences against double calls to end writing. 2017-09-16 17:07:36 -04:00
Thomas Harte
da082673d7 Drives now have a finite number of heads. 
The Amstrad volunteers itself to be single sided. Everything else stays as it was.
 2017-09-15 21:18:36 -04:00
Thomas Harte
662d031e3c Adds exposition on the meaning of a disk controller being in write mode. 2017-09-15 19:14:36 -04:00
Thomas Harte
bf20c717fb The Drive now no longer produces input when in writing mode — other than announcing the index hole. 2017-09-14 22:32:13 -04:00
Thomas Harte
4d4a0cf1d2 Puts the disk controller back into the loop with knowledge about reading mode, and uses that knowledge to cut off the PLL. 2017-09-14 22:30:40 -04:00
Thomas Harte
82b13e98f2 Implements the real hardware ready test for Drives — motor on plus two index holes. 2017-09-11 22:27:50 -04:00
Thomas Harte
96bf133924 Withdraws requirement for DiskController users to specify a PLL multiplier or to provide rotation speed. 
In the latter case because it's no longer of any interest to the controller, and in the former because I'd rather it be picked automatically.
 2017-09-10 22:56:05 -04:00
Thomas Harte
6d6cac429d Fixes extra time accumulation during track running. 
Introduces a bunch of further asserts, which aided me in determining the fix, i.e. that Drives being responsible for their own setup_track could double-pump the event loop.
 2017-09-10 22:44:14 -04:00
Thomas Harte
8882aa496f Corrected wiring to get advance signals through to Drive event delegates. 2017-09-10 20:51:05 -04:00
Thomas Harte
0622187ddf Strips Controller of all capabilities now housed on the Drive. 2017-09-10 19:23:23 -04:00
Thomas Harte
523e1288fa Updates the MFM parser to use SingleTrackDisk rather than the equivalent withdrawn Drive functionality. 2017-09-10 17:34:52 -04:00
Thomas Harte
1a96cce26f Implements SingleTrackDisk, a Disk that contains only a single, specified, track. 2017-09-10 17:34:14 -04:00
Thomas Harte
a4e275e1fc Provides an implementation of Drive's new interface. 
Mostly lifted from DiskController. `set_disk_with_track` has been withdrawn in favour of providing a suitable wrapper `Disk` subclass, as being an unnecessary complexity and intermingling of concerns.
 2017-09-10 17:33:01 -04:00
Thomas Harte
ff6e65cca9 Introduces necessary storage and interface for writing. 2017-09-10 16:23:31 -04:00
Thomas Harte
90c7056d12 Started devolving timed event loop logic down to the drives, moving them closer to modelling real life. 2017-09-10 14:43:20 -04:00
Thomas Harte
c6e6c3fcfb Resolves all current analyser warnings. 2017-08-24 22:18:44 -04:00
Thomas Harte
49285e9caa Attempted to implement Sleeper in Drive and therefore in DiskController. Also corrected a couple of nonconformant file names. 2017-08-20 11:54:54 -04:00
Thomas Harte
5344e3098b Minor: made has_disk something that is decided on insertion/deletion. 2017-08-20 10:55:08 -04:00
Thomas Harte
ef07c33741 Merge branch 'Plus10' into MFMParserDensity 2017-08-18 15:48:20 -04:00
Thomas Harte
e559a65ede Ideally I would be able to kill this multiplier, as it could easily be derived at runtime. But, for now, just turned it up so that the analysis-oriented parser is better at parsing different bit rates. 2017-08-18 15:47:46 -04:00
Thomas Harte
af61a7fa28 Two quick fixes: correctly set segment size, and flip bytes to match HFE's bit ordering to PCMTrack's. 2017-08-17 22:28:00 -04:00
Thomas Harte
c8c1792c3f Made a first attempt at HFE support. 2017-08-17 22:20:02 -04:00
Thomas Harte
e6683e7f2d Added the base skeletal stuff of HFE support. 2017-08-17 21:48:48 -04:00
Thomas Harte
0c1714b695 Relaxed a little to allow +10% in track length. 2017-08-17 21:36:14 -04:00
Thomas Harte
2c2dd8073c Modified to return nullptr if asked for an extended disk image track that doesn't exist. 2017-08-17 15:32:24 -04:00
Thomas Harte
4f8b89772e Improved logic for detecting when all sense has been derived from a track to spot any repeated track, not necessarily the first one. That avoids sectors that run over the index hold and obscure the first throwing things. 2017-08-17 15:31:53 -04:00
Thomas Harte
733ee5a5c3 Ensured no attempt to put a null track into the cache 2017-08-17 15:30:02 -04:00
Thomas Harte
fedf5a44a6 Imposes a maximum track length. 2017-08-17 15:20:49 -04:00
Thomas Harte
da3e8655e9 Withdrew some caveman debugging nonsense. 2017-08-17 13:25:19 -04:00
Thomas Harte
42ad670ec8 Fixed: catalogue bitmap is in blocks, not sectors. 2017-08-17 12:47:47 -04:00
Thomas Harte
468770b382 Removed debugging nonsense. 2017-08-15 22:06:58 -04:00
Thomas Harte
6cfc3daacb Introduced a test within the disk controller so as not to request illegal tracks from disks, instead automatically substituting an 'unformatted' track. Which is just empty. 2017-08-15 21:52:12 -04:00
Thomas Harte
7b92b235e1 Further upped asserts, thereby discovering the mistake I'd recently introduced: seeking properly within the event source as per its potential left-clipping, but then not allowing for that in the calculated current time. 2017-08-15 16:25:46 -04:00
Thomas Harte
c196f0018f Upped the assert quotient. 2017-08-15 16:15:09 -04:00
Thomas Harte
73080d6c36 Added an easy way for disk controllers to clamp termination of written data exactly to the index hole. 
This commit also temporarily provides a whole load of extra logging and minor logic improvements from the 8272. I'm mid-flow on finding a particularly vicious error in its handling of writing; wait for the pull request. But, at least: now waits for the first part of a post-ID gap before writing data, and attempts partially to handle appearance of the index hole during writing a track. More work to do on that though.
 2017-08-15 16:05:10 -04:00
Thomas Harte
9541a2a5f0 Corrections: seek_to now takes the segment_start_time into account, correcting a windowing error where segments overlay other segments. Also added some asserts while bug hunting, and corrected the steps taken when inserting a longer-than-a-track segment so that each is correctly windowed. 2017-08-15 15:54:09 -04:00
Thomas Harte
944222eba4 Added: write_id_data_joiner can now be instructed not to write the first portion of gap. Which makes more sense as an option, to avoiding splicing errors. 2017-08-15 15:29:23 -04:00
Thomas Harte
a366298022 Factored out the standard [M]FM gap and mark groups, to increase 8272 readability and because it's pretty-much certain I'll need them again if ever I try to tackle e.g. the 8271. 2017-08-14 16:03:35 -04:00
Thomas Harte
4df9307d25 Factored out the dull and repetitious stuff of writing n bytes of the same value. 2017-08-14 15:50:36 -04:00