- DISCRETE_TASK_START now requires a parameter TASK_GROUP (>=0, <=9)
- Tasks are scheduled in the order of their task group
- Nodes are automatically buffered between task groups
- Discrete core determines nodes which need buffering to minimize overhead (information in DISCRETE_LOG)
- A discrete block list now must put each stepped node into a task if it uses tasks
- Drivers not using tasks will get one task allocated automatically
- Updated drivers accordingly
- Some more constification
> From: Atari Ace [mailto:atari_ace@verizon.net]
> Sent: Wednesday, September 30, 2009 7:56 AM
> To: submit@mamedev.org
> Cc: atariace@hotmail.com
> Subject: [patch] More static qualifiers
>
> Hi mamedev,
>
> This patch makes more of MAME static, primarily targeting functions
> exported by header files that are in fact unused outside their own
> file, and the chip emulators in machine/snes.c. It also degenericizes
> some exported names in archimds, bublbobl, and lucky74.
>
> ~aa
- preparation work so that a task node output buffer may be read by
more than one following task.
- target: implementation of task groups: tasks in a task group run parallel, task groups serial. The current main task may than just be task (in the last task group)
Sent: Thursday, October 01, 2009 5:25 AM
To: Aaron Giles
Subject: MD update
I'm trying to get the HazeMD code running again (by request, and so
that
I can look at fixing some of the bugs that have surfaced, including
ones
which affect MAME)
This shuffles a bunch of stuff around in order to help with that
process, I've also killed off some old megaplay code which wasn't
really
needed anymore, and used the newer code instead.
> From: Atari Ace [mailto:atari_ace@verizon.net]
> Sent: Sunday, September 27, 2009 7:58 AM
> To: submit@mamedev.org
> Cc: atariace@hotmail.com
> Subject: [patch] More _NAME macros
>
> Hi mamedev,
>
> MAME's idiom for function/data macros is to first implement
> <name>_NAME, then implement the other macros in terms of the _NAME
> macro. Then in principle only a single line needs editing to change
> the naming convention.
>
> This patchset implements this idiom more completely. The first patch
> adds some missing _NAME macros and fixes cases in source files that
> should be using the macros. The second patch then changes header
> files where the macros should have been used, but weren't. This
> required changing the idiom for removing a machine driver function
> pointer from MDRV_<FUNCTION>(NULL) to MDRV_<FUNCTION>(0), to avoid
> problems with NULL being macro expanded. This actually unifies the
> handling of all such cases, as we already had ipt_0 and driver_init_0.
> It also required reworking the devtempl.h implementation in a way that
> triggered a warning on MSVC about using empty macros, so vconv.c
> needed to be updated. The third patch then renames all the _NAME and
> _0 macros to verify that all the cases have been covered, so it isn't
> intended to be applied.
>
> ~aa
> Sent: Sunday, September 27, 2009 5:45 AM
> To: submit@mamedev.org
> Subject: Increase the number of PROFILER_USERx slots when profiling
>
> Hi,
>
> sometimes, when I profile a game in order to see what is the most time
> consuming part of its emulation, the 4 PROFILER_USERx are just not
> enough.
> The attached patch increases this number to 8 in order to give more
> flexibility.
>
> It also removes the now unused and obsolete PROFILER_END constant in
> order
> to be sure that 'profiler_mark_start' will never be called with it
>
> Hope this help.
> Best regards,
>
> CJ
Sent: Friday, September 25, 2009 12:15 PM
To: Aaron Giles
Subject: 360 pad crash
This 'fixes' the bug where MAME crashes and corrupts your config files
if you connect a pad at runtime. Previously there was an assert, now it
simply doesn't execute that block of code if devcode is null.
I don't know if this is the best fix (it would probably be quite nice if
it could detect pads being added removed at runtime and adapt to that)
but at least it stops it crashing and corrupting files.
> Sent: Wednesday, September 30, 2009 4:05 AM
> To: submit@mamedev.org
> Subject: Cheat
>
> I have added a simple system for auto-detect the ram region for make
> more fast the cheatinit. But I have a problem on some address
> translation, for example in seattle.c the ram region have a physical
> address of 0x00000000 - 0x007fffff and the logical address is
> 0x80000000 - 0x807fffff, I not have found a way for convert the
> physical
> address to the logical address. For now the only way for initialize the
> cheat on seattle.c is force the address to the right range ("ci ub,
> 0x80000000, 0x7fffff") but is not a good solution.
> I hope can help me to fix this problem.
>
> Regards
>
> Sandro Ronco
Fixed the logical/physical issue by having the cheat system always
work at the physical layer and output cheats that explicitly point
to the physical space.
by prepending with an 'l' or 'p'. Logical remains the default. Example:
ppb@1000 = physical program space byte at address $1000. ldw@2000 =
logical data space word at address $2000.
New games added or promoted from NOT_WORKING status
---------------------------------------------------
Othello (version 3.0) [Tomasz Slanina, Stefan Lindberg]
------------------------------------
48-in-1 MAME bootleg (ver 3.09) [Guru]
48-in-1 MAME bootleg (ver 3.02) [Guru]
I lost the decrypt for these in an inbox cleaning accident, but even with
that out of the way the protection is much heavier than 39in1's so they're
not really worth the effort.
Credit: MetalliC
MAMETesters bug ID 03403 "Can't init Roll Fruit game."
Hooked up inputs for Roll Fruit
Added hopper emulation, payout now works
Added information on how to initialize Roll Fruit
Added in missing rom to a few MultiFish sets
----------------
Street Fighter II - The World Warrior (Quicken Pt-I, bootleg) [D. Beneke, Guru, Smitdogg, The Dumping Union]
Street Fighter II' - Champion Edition (Accelerator!, bootleg) [D. Beneke, Guru, Smitdogg, The Dumping Union]
- the return type is UINT8, consequently the designer needs to ensure that the node delivers adequately scaled data.
- input nodes allow the input data to be scaled. The return value is the node value *after* scaling.
Sent: Wednesday, September 23, 2009 2:54 PM
To: submit@mamedev.org
Subject: Tiny speed up in 'src\mame\video\neogeo.c'
Hi,
here is a patch against 'src\mame\video\neogeo.c'
This patch give a tiny speed up in neogeo rendering and, IMO, gives the
source code more readable.
This is achieved by unrolling a loop in 'draw_fixed_layer'
Doing so, the logic inside the loop is simplified, less arithmetic is
performed and branches are avoided.
Hope this helps,
Best regards,
Christophe Jaillet
> Sent: Wednesday, September 23, 2009 4:41 AM
> To: submit@mamedev.org
> Subject: optional flag in -listxml
>
> This patch:
> - adds an "optional" flag for rom/disk entries to -listxml output
> - removed the obsolete "dispose" flag from the DTD in the -listxml
> output
Sent: Monday, September 21, 2009 2:54 PM
To: Aaron Giles
Subject: Photo2 coinage
I've added dipswitches for the coins, you can now select between 1 and 6
minutes per coin.
I've also added information about the original games that these have
been bootlegged from.
> Sent: Monday, September 21, 2009 9:02 AM
> To: Aaron Giles
> Subject: NeoGeo update from JohnBoy
>
> roms are public.
>
> - Renamed sets based on them being used for the initial AES production
> runs
> Renamed tophuntra to tophuntrh, added correct s1
> Renamed kof97a to kof97h
> Renamed bstars to bstarsh
>
> . Added correct c1 to neocup98
> . Added correct m1 and v1 to androdun (MVS version layout)
> . Tagged lastbld2 as newer revision
> . Identified pbobblena as bootleg, renamed to pbobblenb, renamed roms
> to
> 'correct' chip label
> . Renamed p1 in mslug4, see notes
>
> New sets added: [Johnboy]
> . Added Baseball Stars Professional (set 1)
> . Added Metal Slug 4 (set 2)
> From: Atari Ace [mailto:atari_ace@verizon.net]
> Sent: Sunday, September 20, 2009 9:54 AM
> To: submit@mamedev.org
> Cc: atariace@hotmail.com
> Subject: [patch] Eliminate more #ifdef LSB_FIRST
>
> Hi mamedev,
>
> This patch recodes more cases where LSB_FIRST is used to conditionally
> compile separate code for LSB and MSB targets. The atari.h chunk was
> dead code (both in MAME and MESS) so I simply removed it.
>
> ~aa
----------------
Exzisus (Japan, dedicated) [Stefan Lindberg, Phil Bennett]
(This is now the parent set. The existing set becomes exzisusa)
Added correct ROM locations for exzisus and flagged exzisusa b23-03.14l as BAD_DUMP. [Phil Bennett]
- implemented reset for TMS5220C if /RS and /WS are pulled down together
- Inserted LNs email and corrections as comment into source
- Created a TODO section.
> From: Sandro Ronco [mailto:sandroronco@gmail.com]
> Sent: Wednesday, September 16, 2009 12:48 AM
> To: submit@mamedev.org
> Subject: MAME cheat update
>
> I have added the byte swap to the cheat search.
Sent: Thursday, September 10, 2009 7:10 AM
To: submit@mamedev.org
Cc: upstephh_wip@yahoo.com
Subject: Quick fix for the F6 toggle
Hi
My last diff got altered before getting included, I appreciate this would likely get done (and many
thanks for doing it) but it did lose the F6 functionality when a cheat was selected. Eg. if you have
the cursor bar over Infinite Time then F6 will not work. My original solution using numbered calls
was not ideal but it did work, I'd spent ages trying to find the appropriate routine to call without
joy. Annoyingly I just tried again with 0.134 and it took less than 5 minutes to fix it
properly....DOH!!!
Anyway I've attached a simple one-line diff for 0.134 to hook up F6 wherever the cursor bar is on
the cheat menu screens.
It just changes this in uimenu.c.:-
- if (exclusive_input_pressed(menu, IPT_UI_TOGGLE_CHEAT, 0))
+ if (ui_input_pressed_repeat(menu->machine, IPT_UI_TOGGLE_CHEAT, 0))
Martin 'Pugsy' Pugh
- this emulates the RS/WS and READY handshaking according
to the tms5220 manual. Drivers reading the ready line
should most probably use it.
- Note: For cases where the READY handshake is done by the cpu
sufficient cycles have to be burnt to simulate that the
cpu waited. This yet has to be implemented.
- use tms rom clock as sense input for the s2650. darkwar now speaks again on startup.
- only one latch used for sound & speech.
- hooked up 393Hz signal to "dac3"
- pcb pictures indicate discrete logic on the sound board. Marked all games as GAME_IMPERFECT_SOUND since we seem to miss some sounds.
- 4bit dac remains mysterious
(Off the record)
I verified this thing against IDA Pro's 56156 disassembler and the docs.
Every time I found a bug in IDA's disassembler, I cross-referenced the
manual. There remain 3 opcodes which are questionably disassembled,
since even the manual is ambiguous on the details, but beyond that,
this thing should be 100% correct.
Whew. This might have actually fleshed out a bug in the
disassembly/execution. Time will tell...
CVS:
- hook up speech cpu
- get rid of "temporary" code which hardwire ctl & pdc writes
tms5110:
- Implement ctl port input/output for status reads.
- Input/output is switched by clocking pdc after a TALK STATUS command.
- update drivers to support new tms5110_ctl_r
Not worth mentioning, I think. NSS bios does not need update, so it's a minor change for MAME (but necessary for the possible legal issues of including those bits in the source)
