* Split YM2154 into a separate device
* Created fake YM2154 ROMs as placeholders
* Created new BBD sound device in src/devices/sound/bbd.cpp
* Created new mixer device in src/devices/sound/mixer.cpp
* Connected YM2154 and BBD devices
* Approximated BBD driver behavior
* Exposed analog sliders as adjusters (accessible via ` menu)
* Added sliders to layout and animated them
* Sync with upstream. I/O callbacks are now consolidated into a single read callback and a single write callback, with an access type specifier.
* Initial working implementation of YM278B. Most features implemented, except vibrato.
* Implement vibrato and status register flags. Fix envelope rate computation.
* Rename ymfm_interface::external_type to access_class and clean up the fallout.
* Formally replace the old YMF278B engine with the one from ymfm
* Rotated YMF278B outputs into a more logical order.
* Re-evaluted envelope calculations and 2x works better than the weird 15/8 I came up with before. Also changed the way FM resampling is computed to be more precise (and simpler). Turned off extraneous debugging.
* Start of/reset to a null state with no loaded waveforms.
* Fix YM2608 I/O ports.
Software list items promoted to working
---------------------------------------
vsmile_cart: V.Smile Smart Keyboard (USA), V.Smile Clavier Tip Tap (France), V.Smile Schreibspaß (Germany) [Ryan Holtz, PulkoMandy]
ymfm: refactor the code into a separate 3rdparty library
* Moved ymfm core implementation to 3rdparty/ymfm
* Split out each family (OPM/OPN/OPL/etc) into its own source file
* Added preliminary OPQ and OPZ support, still WIP
* Put all 3rdparty code into its own namespace ymfm
* Fixed various bugs reported in #8042
* Created interface class for communication between the 3rdparty engine and the emulator
* Standardized MAME implementation of all Yamaha devices based on a template class
* Created standard base class ym_generic that can be used when multiple YM chips are swapped in
* Changed YM2203/2608/2610 to embed a YM2149 as a subdevice instead of deriving from ay8910_device
* Also provided compile-time option to use a simplified built-in SSG rather than using MAME's at all (currently off)
* Consolidated MAME header files from one-per-chip (ym2151.h, ym2203.h, etc) to one-per-family (ymopm.h, ymopn.h, etc)
* OPL prep. Define FAMILY constant in register classes, and use that instead of template specialization for family-specific behaviors. Expand channel masks to 32 bits. Add is_keyon() helper.
* Made FM channel and operator mapping more flexible. Operators are now owned by the engine and can be dynamically assigned to channels. Register classes now provide a mapping between a linear set of operators and channels. The register data array is now a regular array instead of a vector.
* Minor change for consistency.
* Introduce OPL registers and create a ymopl_engine. Add support for sustain-less notes and the OPL envelope clocks (which does not divide by 3).
* Moved keycode calculations into register class. Removed unnecessary recalc in phase generator. Lined up OPL frequency, feedback, algorithm, and total level.
* Implement key scale level and OPL-specific LFO and phase handling.
* Create new YM3526 device based on new OPL. Fix keyon and sustain behaviors.
* Fix weird OPL multiple values. Clean up and further document OPL LFO.
* No busy flag on OPL, so no need to do the work. Add a right shift of 1 to the output stage to line up volume with old implementation.
* More compact way of representing key scale level values.
* Move the KSL bitswap into the registers since it's apparently fixed in OPLL and perhaps others.
* Add support for ryhthm generation in OPL. Change compute_volume to take a phase value directly, and expose operator phase.
* Fix OPL timers.
* Start LFSR with a non-zero value to ensure it actually starts generating for OPL.
* Fix silly bug in sustain logic for OPL. Fixes a lot of previously missing sounds.
* Create OPL2 engine. Add waveform support for OPL2/3 waveforms.
* Wire up YM3812 to the new OPL2 engine.
* Reset OPL timers when the mask is written as well.
* Manage rhythm key ons separately. Fixes Wardner awfulness.
* Explicitly make channel logic handle 0, 2, or 4 operators rather than checking each one for null. Also simplify the combining logic for the 2 operator case.
* Reverse bit order of LFSRs to make things a little simpler. Fix OPL LFSR so that it has its full 23-bit period.
* Change outputs along the path to arrays rather than stereo items. This allows for four channel output. Also add a constant for the number of outputs to the register class.
* Move status register bit definitions to the registers class. Generically support the IRQ bit.
* Create shared helpers for FP encode/decode/roundtrip and use them throughout. Also update TMNT to use the FP decoder.
* No need to clamp when using the roundtrip.
* Clear the EOS flag when execute is turned off on ADPCM-B. Fix combine_status in YM2608 to ignore previously set flags.
* Add missing note_select in base class. Don't add 1 to the OPL release rate.
* Move Y8950 over to new OPL engine.
* Remove old y8950, along with fmopl and ymdeltat
* Add updates prior to status reads for ADPCM systems.
* Add status_mask and irq_reset logic into the core. Clean up documentation on family-specific registers. Includes some temporary gross debugging stuff.
* Made debugging less gross by giving operators and channels a reference back to their owner.
* Fix status port address in OPL chips. Reduce ADPCM volume to match previous implementation.
* Fix Y8950 ADPCM start. Return masked status properly.
* Initial cut at OPLL mapping.
* Add YM2413 support based on ymfm; renamed vrc7snd to ds1001; added YM2423 amd YMF281 variants as well. Instrument data is now loaded via external ROMs. Added 'depress' envelope support to the core engine. Fixed a number of issues in the ymopll_engine. Documented hard-coded values. Moved register clear into register-specific reset.
* Add missing identifier.
* Y8950 is OPL not OPL2.
* Some documentation cleanup. Consistency fixes in the register classes.
* Consolidate large comments. Add support for delayed modulators for OPL. Broke out 2-op and 4-op cases to help simplify logic. Fixed overflow handling in fp encoding.
* Fix silly bug.
* Changed operator assignment mechanism to be more readable. Added prepare method to be called at the start of sound update. Added ALL_CHANNELS constant to register files. Updated all consumers to call prepare and use constants where applicable.
* Move YMF262 and YMF278B to use new FM engine for OPL3/4. Fix several issues in OPL3 logic, which now seems to work ok.
* Minor fixes. More documentation.
* Fix MSVC build.
* Add caching of data to prepare methods to improve performance. Moved non-register decoding logic out of the .h file and into .cpp file. Move phase_step calculation into register class.
* More notes. Removed keycode from cache. Split 2/4-operator outputs into separate functions. Changed OPN/OPL to use templates for variants. Added channel/operator_offset helpers.
* Fairly substantial overhaul of register interface. Register interface is now stateless and contains family-specific state. Channel and operator accessors are prefixed by ch_/op_ now and require an offset to the specific channel or operator. Moved LFO/noise generation into register class, along with keyon logging.
* Add noise back to OPL/OPLL
* Added early-out for low envelope. Moved waveform logic out into family-specific code. General clean up of ordering. Reduced family base class to minimal needed.
* More aggressively track active channels to help performance.
* Use only summing outputs for consideration of active channels. Centralize the logic of determined 4-op vs 2-op.
* More conservative channel deactivation.
* Add helper to compute the sample rate and use it in all implementations. Remove unneeded chnum/opnum members.
* Fix error in YM2612 that caused crashes.
* Switching parameters and locals to 32-bit values gives a noticeable performance boost. Checkpoint 1.
* More moving to 32-bit values. Checkpoint 2.
* Last of the 32-bit promotions.
* Ensure SSG inverted flag is only tested on systems with SSG support in the innermost loops.
* Make most constants 32-bit as well. Expose some constants via the engine class.
* Expand waveforms ahead of time. Optimize the attenuation lookup a bit.
* If volume is low, don't erase output, just leave it alone. Fixes missing sound in raimais.
* Replace a couple of magic numbers with constants.
* Normalization of FM engine usage across consumers.
* Removed explicit external prepare() call; this logic is now automatically done after writes and periodically. Changed OPL/OPLL to use native formats for block/fnum. Fixed waveform generation. Fixed PM and AM depth on OPL/OPLL.
* Cache multiple value. Clean up output handling a bit.
* Move multiple caching ahead of phase step caching.
* Fully split OPLL from OPL. Remove many hacks now that OPLL registers can have state. Created shared helpers between OPL/OPLL. Removed more aggressive channel disables since it was not rhythm friendly and didn't really buy much.
* Remove old comment
* Remove bad write in OPL mode case. Fixes fsoccer intro. Only call set_reset_status() once per mode call.
* Remove FM output boost in YM2608/2610. Not sure why I did that. Better matches previous volume now.
* Make AM/PM logging less confusing.
* Let's actually set DYNAMIC_OPS properly, eh?
* Improved logging.
* Comment cleanups. Add constant for dynamic phase. Pre-shift sustain level. Srcclean.
* Fix memory regions on YM2608 games.
* Clean up ymadpcm to line up with recent ymfm changes.
* y8950: Reshuffle read/write handlers. Rename them to less confusing names.
* ym2413: Reshuffle read/write handlers. Rename them to less confusing names.
* ym3526/ym3812: Reshuffle read/write handlers. Rename them to less confusing names.
* ymf262: Match read/write details to datasheet and previous tests.
* Use a constexpr function instead of macro for packing operator numbers. Pre-compute OPM LFO waveforms.
* Generate OPL4 engine to support the proper clock divider and new flags. Update YMF278B to use FM timers and status rather than replicating the logic.
* Fix 4-operator enable on new OPL4 instance.
* Fix FM downsampling and adjust balance in YM278B.
What's missing:
- parameters (like the disk name when it exists)
- possibly a cleanup of ram_open and friends (but not sure of the appropriate direction in which to go)
New BSD-licensed implementation of Yamaha OPN and OPM FM audio chips, along with new device drivers for YM2203, YM2608, YM2610, YM2610B, YM2612, YM3438, and YM2151 based upon these.
New working machines
--------------------
Basketball (Mattel) [hap, Sean Riddle]
QuizKid Speller [hap, Sean Riddle]
New working clones
------------------
QuizKid Racer (MM5799 version) [hap, Sean Riddle]
If you want to build 64-bit and 32-bit in the same tree without them
stomping on each other, use SEPARATE_BIN=1 (you already need to do this
for TOOLS=1 anyway).
* Separated Acorn IOC and MEMC into devices.
* Emulated 8051-based serial keyboard.
* acorn_machine/memc.cpp: Ensure only one logical page is mapped to a single physical page.
* Fixed RISC OS POST IOC register test.
* aa310.cpp: Added debug code to display RISC OS POST failures.
-machine/archimedes_keyb.cpp: Dumped Acorn Archimedes keyboard microcontroller. [Phil Pemberton]
- sa1110: Added rudimentary support for the ICP and DMA blocks. [Ryan Holtz]
- sa1111: Various changes. [Ryan Holtz]
* Hooked up I2S audio DMA support.
* Hooked up IRQ support.
* Hooked up basic GPIO support.
- uda1344: Added initial pass at audio support, based on DMADAC. [Ryan Holtz]
- sed1356: Added support for BitBLT solid-fill command. [Ryan Holtz]
- jornada: Added more keys, touch pen support, and fixed EEPROM access. [Ryan Holtz]
- render.h: Split out layout class declarations into rendlay.h, with some adjustments for the resulting incomplete types (std::reference_wrapper unfortunately does not allow these by C++17 rules)
- rendlay.h: Move old header contents to layout/generic.h
- vecstream.h: Revert changes made in aa29519528. The std::string_view conversion has been made a non-member function (util::buf_to_string_view) and moved to coretmpl.h.
- strformat.h: Remove the using declaration importing util::string_format into the global namespace. It has been moved to emucore.h and a few tool sources; other references have been qualified.
- osdcore.h: Split out file, directory and path classes and methods to a new header (osdfile.h), Doxygenizing the documentation comments.
- Disaggregate many #includes that were including other standard or custom headers. emu.h now includes basically the same things that it did, but other headers have been streamlined; for instance, emucore.h no longer stealth-includes osdcore.h several ways.
Added methods for enabling and disabling breakpoints and watchpoints,
and made debugger views update when breakpoints/watchpoints are
manipulated from Lua. Made breakpoints and watchpoints objects rather
than tables. (It’s not possible to enable/disable a breakpoint or
watchpoint from the object itself, you have to go through its owners'
debug interface.)
Exposed more device_t members for dealing with child/sibling tags and
devices. Also provided a way to get regions/shares/banks from a device
using relative tags rather than going through the memory manager with
absolute tags.
-netlist: Give devices the C++17 namespace treatment.
-Tidied up compiler warning options for 3rdparty.
-emu/render.cpp: Exposed a few information view item properties.
I've guessed whether break or [[fallthrough]] is appropriate. In cases
where it looked particularly suspicious, I added a FIXME comment. All
of these changes should be reviewed by someone familiar with the code.
- Implemented memory mapper for CP/M.
- Serial printer port.
- External RAM is now scrambled.
- SuperCart bankswitching board implemented for Aquaricart.
- Cartridge images of size 4K, 8K, 16K now supported, was previously 16K only.
* Implement TT5665 emulation
Similar as OKIM6295, but with Support more ROM capacity and phrase spaces, 2 sound outputs
used in bowltry, igs_m036 (for some hardwares with TT5665, ex: cjddzsp)
bowltry.cpp: Fix game name (BOWLING TRY! in both poster and PCB)
* bowltry.cpp: Typo
* Fix endline
* tt5665.cpp: Reduce unnecessary argument
This fixes the "typing on all keyboards at once" issue. You can now
enable and disable keyboard/keypad inputs per device in the Keyboard
Mode menu. Default is to enable the first device with keyboard inputs,
and all device with keypad inputs but no keyboard inputs. The settings
are saved in the CFG file for the machine.
Typing in natural keyboard mode only ever types on one keyboard at a
time, but now you can control which keyboard it types on, as it will be
the first enabled keyboard.
You can easily try this out with something like:
mame64d zorba -rs232 terminal cpm
-ui/inputmap.cpp: Show device descriptions as well as tag paths.
-mac128.cpp: Fixed mouse axis wrap compensation, cleaned up mouse code,
eliminated static variables for mouse input state.
You could see the issue with wrap detection easily enough just by
running mac128k/mac512k/macplus and tapping the arrow keys to move the
mouse one pixel at a time. As you moved past the point where the axis
count wrapped, it would move one pixel in the opposite direction.
There were two function static variables related to mouse input state,
probably still lurking from when the code was initially made to use a
driver state class. This obviously messes with save states and prevents
multiple instances.
- bus/a2bus/mouse.cpp: Fixed mouse axis wrap compensation.
This device had the same bug with wrap compensation as mac128k.cpp.
-----------------------------------
vsmile_cart: V.Smile Tanz Mit Center (Germany), V.Smile Défi Gym (France), Gimnasio Interactivo V.Smile (Spain)
-vsmile: Added support for the Jammin' Gym Class dance mat. [bmx, Ryan Holtz]
* Split hc55516 core into separate cores/subclasses for hc55516 and hc55532 (new 'digital' implementation based on decap/die tracing) as well as mc3417 and mc3418 (old existing 'analog' implementation left alone) and hooked the hc55516 and mc3417 implementations to the appropriate hardware drivers. This should vastly improve CVSD sound quality in drivers that use the hc55516. [Lord Nightmare, Sean Riddle]
* Made Exidy's mouse trap use a timer to clock the mc3417 and update the state readable by the z80 rather than relying on the mc3417 to accept a clock parameter and update the z80 state via a callback. The timer implementation ensures proper synchronization and prevents missed clock transitions from the z80 side causing noise in the CVSD audio. [Lord Nightmare]
* Added a biquad-based 2nd order filter emulation for an audio stream (to go along with flt_rc 1st order filter emulation) and hooked it to Exidy's mouse trap (mc3417) and williams drivers (hc55516) following the original schematics. For hc55516 this is necessary as the original chip produces a very pronounced audible 16khz 'carrier' in its output audio even on the real chip, and these filters suppress this. This also happens to a lesser extent with the 8khz quieting waveform noise on both mc3417 and hc55516. [Lord Nightmare]
* Made Williams System 11 and s11_bg relative sound mixing more accurate to the original audio flow on the PCB itself. [Lord Nightmare]
* Switched Williams Joust 2 to use the s11_bg common sound device rather than its own implementation. [Lord Nightmare]
* fmtowns: add a SCSI card slot for the original models
- Add a slot device for the Model 1/2 dedicated SCSI slot
- Add a device for the FMT-121 SCSI Card
- Modify the I/O maps and machine configurations so the slot and the
integrated controller don't overlap
Moved MS DIB parser out of ICO file reader and made it available for
artwork and layout images.
Added more efficient I/O and better error checking for JPEG file loading
(MAME will no longer exit immediately on a bad JPEG file).
Made caller responsible for opening files for loading images, to avoid
decompressing images used in ZIP/7z artwork multiple times.
Added support for JPEG and Windows DIB to picture_image_device.
Added support for SVG image files in external artwork.
Added support for using I/O port value for animation state and masking
animation state values.
Made bounds elements more flexible in layouts.
Reworked headers to reduce dependencies.
Updated layout file format documentation.
machines promoted to WORKING
----
TV MegaMax active power game system 30-in-1 (MegaMax GPD001SDG) [David Haywood]
* added files for the VT APU type, currently no extra functionality, but will be built upon
* psixptty: Fix for FreeBSD
We need to check for __FreeBSD__ not __FreeBSD_kernel__
* FreeBSD: Use gl includes from pkg-config
FreeBSD doesn't install the includes from 3rdparty software in a default
location so use pkg-config to get the correct path.
* FreeBSD: genieos is named simple bsd
* Most OpenGL link issues fixed. Two remain from drawogl.cpp.
This isn't working anyhow.
* "-video accel" now uses WebGL. This allows GL based scaling and
provides a performance improvement.
* Fixed pong and other games by adding another function to the
exception whitelist
* Target now is mame.html. This allows emrun to be used for testing.
mame.js is created as well and thus the current behaviour
unchanged.
This adds support for the mini digital cassette recorder that can be
found inside a P2000t. This implementation is based on documentation
that can be found in https://github.com/p2000t/documentation.
In memory of NPM Jansen, who taught me all the magic of bits and bytes.
* animalc: Pile kludge upon kludge for poorly understood video timing register
* gocowboy, itazuram: Increase frequency of one timer interrupt (and hopper timing in gocowboy)
* kc82, kp69: Modernize state_add syntax
* Removed device and macro header files.
* All of those can be generated automatically so going forward there is
no need for these any longer.
* Introduced the modules concept. Modules are netlists for which
automatic lib entries are generated.
* Going forward you just store them in macro/modules and they will be
automatically registered as device elements.
* You need to do a "make generated" is src/lib/netlist/build
* Some_device.cpp still needs to be added to netlist.lua
* Added documentation on how to add devices to netlist.
* Please refer to adding_devices.md for more information.
* Ported Cirrus Logic CS8900A Crystal LAN MAC emulation from VICE and hooked it up to Apple II card device.
* Adds Ethernet networking support for Apple IIgs.
* removed include directory src/lib/netlist from various genie files to
avoid potential issues.
* Code using netlist should use #include "netlist/*".
* Updated includes.
* Fixed standalone makefile depend target to properly deal with relative
paths.
IMGUI_DISABLE_OBSOLETE_FUNCTIONS was defined in osd/modules.lua but not
in 3rdparty.lua. As a result, two different variants of struct ImGuiIO
were being defined, causing a C++ One Definition Rule violation
* Move DIPs for 82S16, 82S115, and 2102A devices into nlm_proms
* Moved 7448 DIP to a macro. Replaced 7442 with truthtable and macro.
* Moved 74LS629 DIP into macro.
* Expand truthtable to handle 10 outputs.
- Rewrite memory system, now allows supporting expansion devices and better prepared for contention emulation
- Add expansion interfaces for rear expansion, drive ports and mouse
- Now supports the following rear expansion devices:
* Blue Alpha Sound Sampler
* Dallas Clock
* 1 Mb Interface
* SAMBUS 4-slot Expansion Interface (with clock)
* SID Interface (6581 and 8580 variants)
* S.P.I. SAM Parallel Interface
* Voicebox
- Added support for the Atom HDD interface, used in place of a floppy drive
- Simplified and cleaned up driver
- Temporarily deactivated joystick code, interferes with the keyboard
-tsb12lv01a: Added a skeleton device for the TI TSB12LV01A IEEE 1394 link-layer controller. [Ryan Holtz]
-ibm21s850: Added a skeleton device for the iBM 21S850 IEEE 1394 PHY controller. [Ryan Holtz]
* palloc.h/pmatrix2d.h: Fix static_assert warnings at the origin.
* Rework hints to broaden their use and fix NC hint.
* 74377: use NC hint
* plists.h: Fix debugging in MSVC
* Include cleanup: Move everything not needed by netlists from
nl_setup.h into core/setup.h
* Fix some clang tidy warnings
* srcclean
Available keyboards are us (M0110, U.S.), gb (M0110B, British), fr
(M0110F, French), pad (M0120F, numeric keypad with passthrough port)
and plus (M0110A, U.S. with integrated numeric keypad). The mac128k,
mac512k and mac512ke drivers default to the numeric keypad with the
U.S. keyboard connected to the passthrough port; the macplus driver
defaults to the U.S. keyboard with integrated numeric keypad.
Note that the numeric keypad may seem strange. Four of the operators
work as cursor arrows if you don't hold shift. There is a comma on one
of the keys, but by the time System 6 was released, Apple had decided
an equals sign was more useful, so that's what it will produces on
newer system versions. The U.S. keyboard with integrated numeric
keypad emulates these aspects of the stand-alone keypad - pressing the
operator keys on the keypad sends fake shit key down/up events, and
using the arrow keys while holding shift will produces operator
characters rather than selecting text.
The ISO layout keyboards (M0110B and M0110F) produce different scan
codes to the ANSI keyboards (M0110 and M0110A) but they don't report a
different identification byte. To use an ISO keyboard, you must open
the Keyboard control panel and change the layout to International (and
change it back to Domestic if you switch back to an ANSI keyboard).
This doesn't actually work at the moment due to issues with 6522 VIA
emulation, but it will work with macplus sys603 if applied on top of
revision 963a2c166d.
-----------------------------------------------------------------------
Kidniki now achieves up to 910% when run with static solvers and with
nltool. That is significant better than the 860% we have seen
previously.
This increase is driven by using a global memory pool in the solver
code.
In addition the following refactoring and code maintenance work is
included. Please excuse the large commit, some of this took interfered
with other work and the detail development steps were ugly.
- gsl support: This commit adds pgsl.h which implements a very limited
number of the functionality of the gsl header described in the c++ core
guidelines.
- clang-tidy fixes
- A significant refactoring of palloc.h. Aligned hints were removed,
they added complexity without a significant performance gain. Vector
operations should better be done on special spans/views.
The code has been tested on linux with g++-7, g++-9, clang-11.
On Windows mingw-10 and VS2019, OSX clang-11.
- Added emulation of the SPG290 CDServo
- Added joypad inputs
- Added RFID card support
- Split SPG290 PPU, Timers and I2C into separate devices
- Added a softlist for the RFID cards
* Support Colecovision Megacart.
Assume that a rom file that is more than 32K in size is a megacart
and that it should be bankswitched using Megacart protocol.
* Put megacart functionality in its own cartridge type.
----------------------------------
Novation BassStation Rack Analogue Synthesizer Module [DBWBP]
Novation Drum Station [DBWBP]
Novation Super Bass Station [DBWBP]
Add disassembler and skeleton CPU device for Panasonic MN1880 architecture [AJR]
Factory elements can now pass additional parameters to device
constructors. This makes the design of interface objects like analog
callbacks easier.
The change also allowed to remove some "deep" calls into the core from
the MAME interface in netlist.h
plists.h was splitted into plists.h, pmulti_threading.h and
ptimed_queue.h. In addition removed plists.h from a number of files it
wasn't used in.
Certain minor adjustment needed to be made for cuda toolkit 10.1 and
10.2.
- Memory references in expressions no longer default to the console's visible CPU if no device name was specified, except when entered through the console itself. Expressions in view windows now use the context of the currently selected device instead.
- The pcatmem debug command and similar qt mouseover function now produce an error message if the initial address translation fails.
Related internal changes (nw)
- The debugger_cpu class no longer interprets memory accesses. The existing routines have been moved into symbol_table (which used to invoke them as callbacks), and reimplemented in most other places. Thecode duplication is a bit messy, but could be potentially improved in the future with new utility classes.
- The cheat engine no longer needs to hook into the debugger_cpu class or instantiate a dummy instance of it.
- The inclusion of debug/express.h within emu.h has been undone. Some debugging structures now need unique_ptr to wrap the resulting incomplete classes; hopefully the performance impact of this is negligible. Another direct consequence is that the breakpoint, watchpoint and registerpoint classes are no longer inside device_debug and have their own source file.
- The breakpoint list is now a std::multimap, using the addresses as keys to hopefully expedite lookup.
- The visible CPU pointer has been removed from the debugger_cpu class, being now considered a property of the console instead.
- Many minor bits of code have been simplified.
The last(?) two changes are:
- Add a template parameter to everything (theoretically the address
space width, in practice a level derived from it to keep as much
compatibility between widths as possible) so that the shift size
becomes a constant.
- Change the syntax of declaring and initializing the caches and
specifics so that they're embedded in the owner device. Solves
lifetime issues and also removes one indirection (looking up the base
dispatch pointer through the cache/specific pointer).
For reasons unknown to me compile optimizations do not behave for
template code. If the implementation is in separate compile units, the
code compiles and performs.
This needs more attention since for certain compilers there is a
considerable performance degregation. It looks like this is only
triggered if too many variants are declared in one cpp file and the
compiler stops inlining.
Thanks to Aaron Giles who made me think about a different approach.
This is a rewrite from scratch for rom devices. It uses a generic
template to implement rom devices which is used together with a
description struct to define a rom device. This leads to highly
efficient code since all information is available at compile time.
This is also a step forward to support tristate outputs. All rom devices
covered by this approach have tristate or open collector outputs and
thus all code changes to support tristate outputs can now be made
consistently in one file.
