Nouvelle update u2 + u3:<\/p>\n
chdman: Properly handle and warn on bin\/cue output of images with Source Changes Rewrite INS8250 to use diserial, updated Microtouch to modern and made Checksum the delay slot in single-instruction blocks in the MIPS3 and mc146818 - Set period even if timer is disabled, required for windows diserial - added optinal rx and tx timers and callbacks [Carl]<\/p>\n ins8250 - serial rx line high on start and fix for msr register [Carl]<\/p>\n m68k: Add API to install a non-debugger instruction hook. Usage is Williams blits with bit 2 set take approximately 2x as long because i386: One more to fix cmpsb (dir \/s in MS-DOS works now) and remove Improvements to Golden Poker driver [Roberto Fresca] Added a trivial Y reject to triangle draw commands in the N64 RDP. N64: Various PIF fixes [Ryan Holtz]<\/p>\n info.c: several fixes and additions to listxml output [Fabio Priuli] i386: Added 32-bit versions of LMSW, and made LMSW only change the Fixed Windows 95 blue screen. [Carl]<\/p>\n Preliminary driver for Merkur's 'Mahn Oh-Mann'. German board game Removed need for *_dev.lst files [Miodrag Milanovic] N64: Force mtc0 opcodes that initiate to-RSP DMAs to end recompile N64: Don't flush the RSP DRC cache on DMA. Fixes Mario 64 and others. N64: Miscellaneous machine fixes [Ryan Holtz] i386: Prevent page faults when returning to same privilege and Documentation of the video data on Dragon Gun. [David Haywood]<\/p>\n N64: Fix audio sync issues in Super Mario 64 and probably others, and N64: Change PI DMA speed from 3 cycles\/word to 5.08 cycles, as i386: LOCK prefix should depend on IOPL on 386. Windows NT beta gticlub: Rewrote the 3D renderer. Implemented gouraud shading, Create a new device_mixer_interface, derived from Better management of FLAC block sizes, aiming to keep them around 2k Re-did the non-CD FLAC codec to be \"smart\" in that it can choose Added missing proms for Fax and Fax 2. [W. Belk, Smitdogg, The New games added or promoted from NOT_WORKING status New clones added New games marked as GAME_NOT_WORKINGSource Changes
\n--------------
\nFixed listxml and validation. [Fabio Priuli]<\/p>\n
\nsubcode data [R. Belmont]
\n[\/CODE]<\/p>\nMAMETesters Bugs Fixed
\n----------------------
\n- 04704: [Gameplay] 25pacman: Unable to continue while playing Galaga
\n (Tafoid)
\n- 04694: [Documentation] m5fmonty: The correct description is \"Full
\n Monty (The) (Empire) (MPU5)\".
\n- 04692: [Documentation] j6kapang: The correct description is \"Kapang!
\n (Crystal) (IMPACT)\".
\n- 04691: [Documentation] j6cpal and clone: The correct description is
\n \"Caesars Palace (Whitbread \/ Crystal) (IMPACT, set 1)\". The
\n\t set 'j6cpala' is his clone.
\n- 04693: [Documentation] m5cockdd: The correct description is \"Cock A
\n Doodle Dough! (Empire) (MPU5)\".
\n- 04696: [Crash\/Freeze] Crash when loading samples in certain games
\n (Aaron Giles)
\n- 04697: [Compiling] Unable to compile LDRESAMPLE.exe tool (Aaron
\n Giles)
\n- 04698: [Compiling] Unable to compile LDVERIFY.exe tool (Aaron Giles)<\/p>\n
\n--------------
\nchdcd: Fix crash when parsing bin\/cue [R. Belmont]<\/p>\n
\nit serial device, updated connected drivers [Carl]<\/p>\n
\nRSP DRCs. [Ryan Holtz]<\/p>\n
\nnt [Carl]<\/p>\n
\nstrongly NOT recommended except for dire cases. [R. Belmont]<\/p>\n
\nthey are bus-shared with RAM. Should impact some timing behaviors such
\nas later levels in Robotron, where approximately 10% of the blits are
\ndone with bit 2 set. Further refinements may be necessary but this
\naddresses the most significant issue. [Sean Riddle, Aaron Giles]<\/p>\n
\nunused variables. [Carl]<\/p>\n
\n * Inputs from the scratch.
\n * Switched manufacturer 'Playman' to PM \/ Beck Elektronik, since it's
\n PM and Beck Elektronik\/Computer\/etc...
\n * Added technical and game notes.<\/p>\n
\n[Ryan Holtz]<\/p>\n
\n * appended to xml info about devices which can be optionally mounted
\n to slots
\n * added a description attribute to slotoption to show which devices
\n can be mounted in each slot
\n * added to xml info of cpus\/inputs\/image\/slots which are relative to
\n devices (e.g. CPUs or Dipswitches in the device itself)
\n * tags of elements directly attached to the root_device do not
\n display anymore the root colon (slot tags were already behaving in
\n this way); elements attached to optional devices use tags starting
\n with colon (to be interpreted as a relative tags which gets
\n appended to the device tag, when such a device is selected in a
\n slot)
\n * modified when necessary output routines to pass a root_device and a
\n root_tag as parameters, in order to simplify output of relative
\n devices and avoid duplicated code
\n * updated DTD according to the new elements and attributes These
\n changes allow frontends to know in advance how drivers can be
\n configured through the various slot options (by matching option
\n descriptions with device descriptions), without the need to run
\n -lx with the various slot combinations. This is especially of use
\n if a device adds e.g. dipswitches or image devices (like
\n additional floppy drives or additional slot expansions), but also
\n if one wants to know e.g. how many sound channels a soundcard
\n would add to a pc driver, or which kind of screen parameters would
\n be added by a specific choice of video card.<\/p>\n
\nlower 4 bits of CR0. [Barry Rodewald] <\/p>\n
\nsimilar to Ludo, derivated from the indian game Parchisi. Coin-op
\nmachine for 1-4 players. No screen, just artwork and lamps.
\n[Roberto Fresca]<\/p>\n
\n * enforced short names for slot card devices
\n * updated validation, romverify and listxml output accordingly<\/p>\n
\nsequences to force a code check. Fixes certain microcodes from
\nexecuting garbage code after DMA. [Ryan Holtz]<\/p>\n
\n[Ryan Holtz] <\/p>\n
\n * Increase the amount of time consumed by PI DMA, makes certain RSP
\n microcodes that test DMA speed land in the correct window.
\n * Force the main CPU to yield when the RSP reads its atomic
\n semaphore.
\n * Force the RSP to yield when PI DMA completes.
\n * Flush the RSP DRC cache when touching IMEM with SP DMA.<\/p>\n
\nunmapped pages are above the stack. Windows 95 boots now. [Carl]<\/p>\n
\nalso bring PI DMA speed into the realm of reality. [Ryan Holtz]<\/p>\n
\naverage PI speed is between 2.53 and 2.55 cycles\/byte. [Ryan Holtz]<\/p>\n
\nboots. [Carl]<\/p>\n
\nlighting and fog. [Ville Linde]<\/p>\n
\ndevice_sound_interface, which mixes all its inputs to a single output
\nstream. Redefined the SPEAKER device to use this and remove the
\nduplicate underlying logic. The main purpose of this new interface is
\nto allow device-ification of an entire sound board, which can itself
\nbecome a mixer of all of its sound components to a single output
\nstream. This stream can then be routed by the device's owner to the
\nappropriate speakers. A real-life example will show up soon. [Aaron
\nGiles]<\/p>\n
\nsamples. Increased CD sectors per hunk to 8 for better compression.
\n[David Haywood]<\/p>\n
\neither big-endian or little-endian on the fly and uses the best one.
\nAdded FLAC as a default codec for hard disks, which helps a lot for
\ncertain disks. [Aaron Giles]<\/p>\n
\nDumping Union]<\/p>\n
\n---------------------------------------------------
\nCasino Poker (Ver PM86LO-35-5, German) [Roberto Fresca, Team Europe]<\/p>\n
\n----------------
\nAltered Beast (FD1094 317-0069)
\n [Stefan Lindberg, Chris Hardy, The Dumping Union]
\nPuzzLoop 2 (Japan 010226) [idc\/Team Avalaunch]
\nStreet Fighter II': Champion Edition (Japan 920322)
\n [idc\/Team Avalaunch]
\nMortal Kombat (Yawdim bootleg - Set 2)
\n [ScheissNussen, Smitdogg, The Dumping Union]<\/p>\n
\n------------------------------------
\nMann, oh-Mann [Roberto Fresca, Team Europe]
\nNFL Classic Football (NCF3 Ver.A.) [Smitdogg, The Dumping Union]
\nArea 51: Site 4 (September 11, 1998) [ninjakid]
\nPot O' Gold [Bicyclerepairman]
\n[\/CODE]
\n<\/p>\n