* Removed .mask(), as it’s not reliable in the general case.
* Added asserts to things that assume power-of-two sizes.
* Got rid of virtual qualifier on pointer-to-member operator.
* Made helpers a bit more assertive about logging warnings.
-emu/rendlay.cpp: Use delegates to avoid hot conditional branches.
-docs: Finished off description of object finders and output finders.
* Made it an error to specify address map and explicit ROM region.
* Made it an error if explicitly specified ROM region is not found.
* Made the ROM region tag apply relative to the current device.
-Cleaned up formatting for a couple more documentation pages.
API impact:
- install_ram/rom/writeonly now requires a non-null pointer. If you want
automatically managed ram, add it to a memory map, not in machine_start
- install_*_bank now requires a memory_bank *, not a string
- one can create memory banks outside of memory maps with memory_bank_creator
- one can create memory shares outside of memory maps with memory_share_creator
Memory maps impact:
- ram ranges with overlapping addresses are not shared anymore. Use .share()
- ram ranges touching each other are not merged anymore. Stay in your range
Extra note:
- there is no need to create a bank just to dynamically map some memory/rom.
Just use install_rom/ram/writeonly
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.
Components may have multiple bounds and/or color child elements with
state attributes, allowing for piecewise linear position/size/colour
animation.
Components may have a statemask attribute, allowing for things like
using external images to draw a multi-segment LED/VFD display without
requiring dozens of outputs for the individual lines or thousands of
images for all possible states. (Texture caching still never releases
anything, so MAME can still exceed the maximum number of textures, but
that’s a separate issue.)
Image components with alpha now blend over previously drawn components.
Layouts have been changed to use yes/no for inputraw to match what's
used for flipx/flipy. External layouts with 1/0 will still work, but
complay.py will complain.
Changed name of element to "collection" and initial visibility attribute
to "visible", and added them to documentation. Also added them to
complay.py.
Fixed issue with collection inside group, and improved initial view
selection behaviour.
Updated some internal layouts to demonstrate new features, including
et3400, irrmaze, ltcasino, mekd3/mekd4, seawolf and vgmplay.
Removed all uses of cpanel, marquee and overlay from internal layouts
and removed them from complay.py to actively discourage use. Also
cleaned up view names in layouts that were using them in place of
spaces, and removed some superfluous name attributes on elements that
won't do anything useful with an output value anyway.
Made vgmplay cycle visualiser modes when visualiser screen is clicked.
Fixed a copy/paste error in bus/rs232/hlemouse.cpp while I'm at it.
This isn’t supposed to be too prescriptive. The C++ stuff just codifies
some things we’ve managed to mostly agree on for public interfaces. The
stuff for titles/descriptions is also just codifying existing rules so
there’s something to point people towards. This will need to be refined
as we go forward.
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 Laserdisc overlays.
This is a change in behaviour, and it means that games like Golly!
Ghost! will need an explicit blend mode specified in the XML. I'm not
entirely happy with the situation, but a better, more general solution
than this would require some serious refactoring to MAME's renderer.
* There is no longer a concept of "layers" - there are only screens and elements.
* Elements are now instantiated with <element ref="...">
* Screens and elements can have explicit blending mode specified with blend="..."
* Default blending mode for screens is "add" and default for other elements is "alpha"
* Other supported modes are "none" and "multiply"
* This removes the options to enable/disable layers individually - use views instead
* Legacy layouts can still be loaded, and support won't be removed for at least a year
The current artwork model is over-stretched. It's based on a Space
Invaders cabinet model, and isn't applicable to a lot of the systems
MAME emulates now. The fact that MAME has to switch to an "alternate"
mode to deal with games like Golly! Ghost! without requiring pre-matted
bitmaps shows that the Space Invaders model wasn't even adequate for
general arcade use. It shows in that for a lot of the systems that
heavily depend on artwork, people just seem to randomly choose layers
for elements until they get something that works. Also, the fact that
MAME will switch to an alternate (Golly! Ghost!) mode depending on the
combination of elements is a trap for people learning to make artwork.
There are cases that the current approach of implying the blending mode
from the layer doesn't work with. Examples include LEDs behind
diffusers (requires additive blending for layout elements), and mutliple
stacked LCD panels (requires RGB multiplication for screens).
For configurability, it's now a lot easier to make multiple views using
groups. For example, if you want to make it possible to hide the
control panel section of your layout, you can put the control panel
elements in a group and create views with and without it.
I will gradually migrate the internal artwork to use the new approach.
I have an XSLT stylesheet that helps with this, but I'm not comfortable
adding it because it isn't a complete solution and it still requires
manul steps.
I wanted to get the re-worked pointer handling done sooner so I could
push them both at the same time, but unfortunately various things have
prevented me from progressing as quickly as I wanted to. Sorry guys,
that stuff's going to have to wait.
This effectively reverts b380514764 and
c24473ddff, restoring the state at
598cd52272.
Before pushing, please check that what you're about to push is sane.
Check your local commit log and ensure there isn't anything out-of-place
before pushing to mainline. When things like this happen, it wastes
everyone's time. I really don't need this in a week when real work™ is
busting my balls and I'm behind where I want to be with preparing for
MAME release.
* Those chapter numbers are auto generated by sphinx
* Fix the chapters not been bold like the rest of the documentation
* Fix chapter numbers not in sync with the rest of the documentation
* Change the last two chapters to keep up with the document consistency
* Fix bold for nscsi_full_device
* Allow <orientation> and <color> to work on group references
* Fix some corner cases where group bounds could be miscalculated
* Fix a corner case where MAME could incorrectly refuse to instantiate groups
* Add more checks to complay.py
* Document more of the layout format
* Clean up some corner cases in layouts with repeating blocks
* Make complay.py validate many more elements and attributes
* Make complay.py easier to use for just validating a layout
* Remove redundant view from Sega VMU layout
* Make buttons visually respond to input in whousetc.lay
* Add view with LED displays as well as terminal for aim65_40 and use repeats
* Clean up some outdated "game" terminology in clifront.cpp
* Initiaise a couple of members in tap/tun network module
* Start documenting layout format
Disassemblers are now independant classes. Not only the code is
cleaner, but unidasm has access to all the cpu cores again. The
interface to the disassembly method has changed from byte buffers to
objects that give a result to read methods. This also adds support
for lfsr and/or paged PCs.
- memory_translate now returns an address space number rather a boolean flag, permitting addresses in part of one space to map to an entirely different space. This is primarily intended to help MCUs which have blocks of internal memory that can be dynamically remapped, but may also allow for more accurate emulation of MMUs that drive multiple external address spaces, since the old limit of four address spaces per MAME device has been lifted.
- memory_translate has also been made a const method, in spite of a couple of badly behaved CPU cores that can't honestly treat it as one.
- The (read|write)_(byte|word|dword|qword|memory|opcode) accessors have been transferred from debugger_cpu to device_memory_interface, with somewhat modified arguments corresponding to the translate function it calls through to if requested.
