Nouvelle update.<\/p>\n
The most major change is the final retirement of the global Machine The second big change is that the CPUs are now full-fledged devices, Starting now, the focus for MAME will be to iron out remaining issues MAMETesters Bugs Fixed Source Changes Funworld.c improvements [Roberto Fresca] Made CPUs into proper devices. CPUs are now added in the machine Changed cpu_type from an enumeration into a pointer to the CPU's Added function cpu_get_type() to return the CPU type. [Aaron Giles]<\/p>\n Changed several cpu_* functions over to the device_* equivalents, The device system now maintains a parallel list of devices based on Cleaned up code that looped over CPUs via the machine->cpu array to Removed start\/stop\/reset\/nvram functions from the device_config in Cleaned up the generic interrupt_enable code to work with CPU devices Mapped the devtag_* functions to device_* functions via macros instead Removed cpuintrf_init() which is no longer necessary. [Aaron Giles]<\/p>\n Changed a number of APIs in the system to accept running_machine, Added the concept of device \"owner\", for devices owned by other Modified laserdisc players to walk back to their global device state Finalizer update [poUzVdb90] Made the concept of a \"clock\" native to devices. The clock is now Defined a generic device execute function callback, though it Changed various NVRAM devices to fetch their default memory region Moved device startup to *before* the DRIVER_INIT is called. This Rewrote the device iteration logic to make use of the typenext Modified the parameters of the stream_update_func callback. As part of Cleaned up i86\/i286\/i386\/nec\/v30mz headers: [Aaron Giles] Cleaned up many other CPU headers to move memory read\/write macros Removed index and clock parameters from CPU_INIT function. Some PIC8259 updates: [Wilbert Pol] Fixed proms in boobhack. [robiza]<\/p>\n Decoded some graphics and setup some RAM in the Virtual Combat driver. Cleaned up device and sound interfaces to match the CPU interfaces Added the concept of a cpu_state_table, which is supplied by the CPU Removed the old debugger register list, which was never used. Replaced Renamed REG_PC -> REG_GENPC, REG_SP -> REG_GENSP, and More cleanup. Added address-space-specific constants for the various Changed ADSP21xx cores to accept a configuration struct instead of Memory view fixes: [Aaron Giles] Added PAL to arkgcbl. [robiza]<\/p>\n Added Country Girl to the Jangou driver,working with sound Added running_machine callbacks for the Amiga code. [Nathan Woods]<\/p>\n Updated sscope2 to use newly-dumped roms. [robiza]<\/p>\n Added running_machine parameters to the f3853 callbacks. [Wilbert Pol]<\/p>\n Added running_machine parameter to the callback in the sid interface. Removed MDRV_INTERLEAVE(x), which specified the minimum scheduling Changed MDRV_WATCHDOG_INIT_TIME(x) to automatically prepend Added new MDRV_QUANTUM_PERFECT_CPU(x) to specify that the minimum Fixed some color issues in the goldstar.c driver [David Haywood]<\/p>\n Merged goldstar and lucky8 drivers. [David Haywood]<\/p>\n Improvements to ddragon3\/shadfrce\/wwfsstar\/wwfwfest: [Mamesick] Fixed Agress' title screen. [David VanVoorhis]<\/p>\n Fixed update_handler for the dsp56k and polygonet. [Andrew Gardner]<\/p>\n Improvements to the dmndrby.c driver. [Mike Green]<\/p>\n Continued rewrite of the Motorola DSP56k CPU core. [Andrew Gardner] Improvements to the skylncr.c driver: [Roberto Fresca, David Haywood] Rewrote 8080\/8085 interrupt handling so that it makes sense. Changed Converted warpwarp driver to raw video parameters. [Aaron Giles]<\/p>\n Fixe 'run to cursor' and 'toggle breakpoint' operations in the Improvements to goldstar.c driver. [Roberto Fresca] New games added or promoted from NOT_WORKING status New clones added New games marked as GAME_NOT_WORKING
\nNOTE: As of this release, we have two more major changes. These should
\nbe the last \"huge\" changes of this dev cycle. <\/p>\n
\npointer, which has been part of MAME for many years. Frankly, I never
\nthought we'd hit this point so quickly, so my hats off to Atari Ace
\nwho helped push through the final round of changes needed to make this
\nhappen finally. As a result of this change, it should be possible in
\nthe coming year to finally write a driver that runs two full machines
\nsimultaneously.<\/p>\n
\nand participate like all other devices in the system configuration.
\nThis is mainly an internal structural change, but in the end removes
\nsome redundant code and behavior. In a future dev cycle, a similar
\nthing will happen with the sound cores.<\/p>\n
\nand release a 0.129 sometime around the 1st of the year. Please keep
\nreporting bugs to MAMETesters as you find things.<\/p>\n
\n----------------------
\n- 02745: [Crash\/Freeze] all sets using autosave: Games using autosave
\n segfault on startup (Aaron Giles)
\n- 02750: [Sound] All Sets in equites.c: no background music
\n (Aaron Giles)
\n- 02757: [Documentation] sengoku3: manufacturer should be \"Noise
\n Factory \/ SNK\" (Phil Bennett)
\n- 02753: [Crash\/Freeze] gradius4: hardware error -1 (Aaron Giles)
\n- 02755: [Crash\/Freeze] All Sets in metro.c: Crash with access
\n violation. (Aaron Giles)
\n- 02756: [Crash\/Freeze] hyprduel, hyprdelj: Colors on the title screen
\n wrong, games crash mame with ACCESS VIOLATION (Mamesick)<\/p>\n
\n--------------
\nRemoved old cherrym driver (meant to be removed last time, oops).
\n[Aaron Giles]<\/p>\n
\n - Added new set: Jolly Card (italian, encrypted bootleg).
\n No coins... Only remote credits. After nvram init, set the Payout
\n DIP to 'manual' to allow the remote credits mode to work.
\n - Created inputs from the scratch for jolycdib.
\n - Updated technical notes.<\/p>\n
\nconfiguration just as any other device, and the standard CPU
\nconfiguration is performed via the inline configuration macros.
\n[Aaron Giles]<\/p>\n
\nget_info function, very similar to device behavior. Updated all
\ndrivers to #include the necessary header files for the CPU's they
\nreference. [Aaron Giles, Atari Ace]<\/p>\n
\nspecifically for get\/set info and for reset. [Aaron Giles]<\/p>\n
\ntype and class, for faster iteration through all devices of a given
\ntype or class. [Aaron Giles]<\/p>\n
\nnow loop using the type-based device list. [Aaron Giles]<\/p>\n
\nfavor of grabbing them as needed. [Aaron Giles]<\/p>\n
\ninstead of numbers. [Aaron Giles]<\/p>\n
\nof parallel implementations. [Aaron Giles]<\/p>\n
\naddress_space, or device_config types, in order to remove the use of
\nthe global Machine variable. Removed the global Machine variable now
\nthat it is no longer directly references. [Atari Ace, Aaron Giles]<\/p>\n
\ndevices. [Aaron Giles]<\/p>\n
\nvia the owner, rather than brute-force searching for the first
\ninstance and hoping that is the right one. [Aaron Giles]<\/p>\n
\n - converted to tilemap
\n - added flipscreen
\n - changed sound chip name from SN76496 to SN76489A<\/p>\n
\nspecified when the device is added, and the clock is available in
\nthe device_config directly via device->clock. Updated all devices
\nthat have a clock to specify it when adding the device, rather than
\nas part of their configuration. As part of this work, also created
\ndevice-specific _ADD and _REMOVE macros to simplify configuration.
\n[Aaron Giles]<\/p>\n
\nis not used yet. The long term plan is that any device with an
\nexecute callback will be scheduled along with the CPUs. Now that
\nCPUs are devices, their scheduling will be moved over to this
\nlogic eventually. [Aaron Giles]<\/p>\n
\nfrom the device->region rather than specifying it in the
\nconfiguration. [Aaron Giles]<\/p>\n
\nis to allow the DRIVER_INIT to configure devices that have been
\nproperly allocated. So far I don't see any negative effects, but
\nbe on the lookout if something weird shows up. [Aaron Giles]<\/p>\n
\nfield and the newly-introduced classnext field for iterating more
\nefficiently through devices of a given type or class. [Aaron Giles]<\/p>\n
\nthis work, macro-ized all the callbacks to STREAM_UPDATE, and added
\nthe sound device as a parameter. [Atari Ace]<\/p>\n
\n * renamed i86.h -> i86priv.h
\n * renamed i86intrf.h -> i86.h
\n * consolidated i88intrf.h, i186intf.h, i188intf.h, etc into i86.h<\/p>\n
\nout of them, since they are inappropriate and should be prive to the
\nCPU core. [Aaron Giles]<\/p>\n
\n[Aaron Giles]<\/p>\n
\n - Fixed operation of edge-triggered interrupts
\n - INT output line is now cleared when an IRQ is acknowledged
\n - Implemented basic OCW3 features<\/p>\n
\n[Andrew Gardner]<\/p>\n
\nwhen handling strings. Namely, the generic get_info functions
\nallocate a temporary string and the device in question copies its
\nstring to the target, instead of assigning a const char *. Updated
\nall device and sound cores to operate this way. [Aaron Giles]<\/p>\n
\ncores and which describes all the register state accessible to the
\ndebugger and other subsystems. The format of the table is such that
\nmost data can be simply fetched from memory without the further
\ninvolvement of the CPU core, including the display of common formats.
\nExtensibility points are available for custom display and for
\nimporting\/exporting the data to intermediate variables for more
\ncomplicated scenarios. Updated the ADSP21xx, TMS340x0, i86, Z80, Z180,
\nand 8085 cores to use this. [Aaron Giles]<\/p>\n
\nit with using ordering from the cpu_state_table. [Aaron Giles]<\/p>\n
\nREG_PREVIOUSPC -> REG_GENPCBASE. Updated a few spots that were using
\nthese directly. Moved these definitions into the end of the register
\narea rather than leaving them outside which put them in a weird range.
\n[Aaron Giles]<\/p>\n
\nbus width and shift CPU interface constants. Changed all the cores
\nto use them. [Aaron Giles]<\/p>\n
\nusing set_info to specify serial port callbacks. Simplified the
\nADSP21xx get\/set info significantly. Removed support for only
\nincluding certain variants of the chips; they are now either all
\nsupported or all unsupported. [Aaron Giles]<\/p>\n
\n * Fixed bug that computed the wrong total size if the address
\n space was a full 32 bits
\n * Improved display to show unmapped memory with *s
\n * Disabled enumeration of disk and 0-length regions<\/p>\n
\n[David Haywood, Angelo Salese]
\n - Fixed transparency pens in the Jangou driver (so removed the
\n GAME_IMPERFECT_GRAPHICS flag); [Angelo Salese]
\n - Added a kludge for Jangou Lady RNG,will be worked on in the future;
\n [Angelo Salese]<\/p>\n
\n[Wilbert Pol]<\/p>\n
\nquantum in terms of \"frames\" (a dubious concept now with multiple
\nscreens and changing refresh rates). Replaced it with a new
\nMDRV_QUANTUM_TIME(x) which specifies the minimum scheduling quantum
\nas a time value. Time can be specified as HZ(x), NSEC(x), USEC(x),
\netc. Updated all drivers to use this, assuming 60 was the frame
\nrate (this is not perfect but should work for almost all cases).
\n[Aaron Giles]<\/p>\n
\nUINT64_ATTOTIME_IN_ to the parameter, ensuring there is no improper
\nuse of this macro and bringing it in line with the MDRV_QUANTUM_TIME()
\nmacro. Updated all callers. [Aaron Giles]<\/p>\n
\nquantum should be enough to ensure that the specified CPU tag only
\never executes a single instruction at a time. This can be used to
\nexplicitly require \"perfect\" synchronization for drivers that have
\nmultiple CPUs with shared memory. Turned this on for the arknoid2
\ndriver for now as a test (the interleave on that driver was already
\nvery close to perfect anyway). [Aaron Giles]<\/p>\n
\n - converted driver to be fully scanlines based using timers
\n - updated driver to new video screen raw parameters
\n - hooked up interrupts (ddragon3, wwfwfest)<\/p>\n
\n * Fixed flag calculation and sign extension for numerous ops.
\n * Added rnd, mpysuuu, and dmac ops.
\n * Fixed do loop behavior to skip empty loops.
\n * Added stack underflow exception handling.<\/p>\n
\n - Complete reel support from the scratch.
\n - Full inputs, outputs & DIP switches from the scratch.
\n - Added new games.
\n - Updated technical and game notes.<\/p>\n
\ncallbacks to be specified in a config structure. Converted core to
\ncpu_state_table. Changed to a single HAS_808X define for both cores.
\nFixed several drivers that used 8080\/8085 interrupts in odd ways.
\n[Aaron Giles]<\/p>\n
\ndebugger when dealing with CPUs that have shifted address spaces.
\n[Andrew Gardner]<\/p>\n
\n - Hooked the 3x PPI 8255 devices to chryigld and ncb3.
\n - Reworked the chryigld and ncb3 inputs based on PPI 8255 connections.
\n - Added the unmapped PSG sound to chryigld and ncb3.
\n - Some works on their machine driver.
\n - Some cleanups.<\/p>\n
\n---------------------------------------------------
\nCountry Girl [David Haywood, Angelo Salese]
\n4nin-uchi Mahjong Jantotsu [David Haywood, Angelo Salese]
\nThree Ds - Three Dealers Casino House [David Haywood, Angelo Salese]
\nSky Lancer (Bordun, ver.U450C) [Roberto Fresca, David Haywood]
\nButterfly Video Game (ver.U350C) [Roberto Fresca, David Haywood]
\nMad Zoo (ver.U450C) [Roberto Fresca, David Haywood]
\nCasino Winner [Chris Hardy, Angelo Salese]
\nCherry I Gold (bootleg of Golden Star) [Roberto Fresca]<\/p>\n
\n----------------
\nJolly Card (italian, encrypted bootleg) [Roberto Fresca]
\nDouble Dragon 3 - The Rosetta Stone (prototype) [Guru]<\/p>\n
\n------------------------------------
\nGrand Prix 98 [David Haywood]
\nLeader [Roberto Fresca, David Haywood]
\nParent Jack [David Haywood]
\nJanshi [David Haywood]
\nCherry Master I (v1.01) [Roberto Fresca, David Haywood]
\n[\/CODE]
\n<\/p>\n