renderer backends (in SDL). This allows to select software mode (not
recommended) or Direct3D/OpenGL in Windows.
The app icon is now loaded in Windows.
Cleaned up the VideoDialog UI, removing references to double-buffering
and OpenGL-specific settings.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2867 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
options have changed names. In particular, all tia-related options
now start with 'tia.'. This includes most of the old options that
started with 'gl_'.
Changed the names of all tv-effects options from starting with 'tv_'
to 'tv.', to match the name changes elsewhere.
Some work on the documentation to reflect the latest changes.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2858 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
than before.
Fixed compile issue for joystick support (thanks to email from D. Church).
Joystick support still isn't tested, so no guarantees it will work.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2857 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
to SDL2. For now, there are many things broken, namely keyboard handling
and fullscreen modes. But the launcher does show up, allows to start a
game and enter/exit the debugger, etc.
The code will only compile on Linux for now, and sometimes maybe not even
then. Expect breakage on a regular basis over the next month or so.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2854 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
build files (sorry, there's no PPC support for Stella 4.0 using
SDL2).
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2851 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
SDL-specific code into its own EventHandlerSDL2 class. This is part
of a larger reorganization of the codebase, to completely remove
SDL-specific code from src/emucore, and make it easier to port to SDL2.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2850 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
into FrameBufferSDL2 only. This will make conversion to actual SDL2 code much
easier, since it won't be touching any of the core emulation code in src/emucore.
Next TODO is something similar for EventHandler (which uses SDL heavily and is
too tied to a core class).
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2846 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
break. First thing to do is update the year. Happy new year to anyone
reading these commit logs.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2833 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
have been changed to 'SDL2'. From this point on, the OpenGL-specific
code will be removed, and replaced by SDL2-specific hardware-accelerated
code.
Cleaned up (mostly) the API of FrameBuffer vs FrameBufferSDL2. Since the
FrameBuffer class is an emucore object, it should have no knowledge of SDL,
or any particular rendering toolkit for that matter (just in case we want
to move to SDL3 eventually :) ).
At this point, all checks for OpenGL have been removed, so if you don't
have it on your system, Stella will fail to compile and/or run.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2832 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
centering has been removed (SDL2 provides this natively).
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2831 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
in line with the other ports. Besides, the code now also supports
64-bit Windows, so the name was incorrect anyway.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2830 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
software rendering support. I've been wanting to do this for years,
and finally the time has come. So for now, you need OpenGL for Stella
to build and run.
Expect major breakage over the next few months, as I port to SDL2.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2829 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
ever want to see the TIA, not the entire debugger area too.
This will be fixed for Stella 4.0, once I remove the @#*%$-ing software
rendering support. That part of the code keeps coming back to haunt me
for years. It will be *SO MUCH EASIER* to deal with these things when
everything is a FBSurface that can be stretched/scaled as necessary.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2828 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
of bold and normal fonts within the debugger UI. Because of the # of changes
this requires, this is as far as I'll willing to go with UI configurability.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2820 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
I wasn't completely sure how the file was structured, but it's starting to
make a lot more sense now. I think in the end, it will be better to use this
file as much as possible.
Bumped version #, and starting all over again.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2815 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
number.name convention (which seem to be pass #'s for DASM). Initially
having this info in Stella was simply visually inconsistent, but now
it also interferes with recompiling when saving the disassembly (IOW,
the data put into the symbol file by DASM can't be fed back into DASM
again in these cases).
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2803 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
tab area and RAM area now take advantage of extra space.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2798 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
to the screen (inactive pixels are 'transparent' in the UI). In the
case where there is only one possible object below the register, a
'blank' uses the underlying colour (ie, in normal priority, a blank
PF is 'coloured' using COLUBK).
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2797 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
COLUBK instead of always using colour 0 (black). This is more
informative, as it displays how the registers actually work when
being drawn to the screen. Thanks to Tjoppen of AtariAge for this.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2796 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
symbol names in the DASM symbol file). This partly fixes a deficiency in
symbol files, whereby two names could have the same address (one has an
address, the other as a constant), and Stella would use whichever one
occurred in the file last.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2795 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
values, which messed up code running from RAM (ie, address $80 became $8000!).
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2790 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
extended abilities (see changelog and docs for more info).
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2783 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
previously on my local system, not in the repo), and fixed some minor
UI layout issues.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2778 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
on exit; it attempted to close the debugger dialog instead!
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2775 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
it was confusing which one was being called.
Fixed output in debugger bankstate info; in some cases, hex numbers were shown
when decimal was required.
Fixed regression wrt the last TIA change (incorrect framerate counter); the
PAL autodetection was failing in certain cases.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2774 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
Once a frame has started, it means a new frame should be added. This
fixes a bug in the debugger UI whereby the frame count showed the previous
value, since sometimes you could enter the debugger (with a cond BP)
before the end of the frame was technically reached.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2773 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
This is the beginning of a major cleanup of the bspf.hxx file, which
has gotten unwieldy over the years.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2772 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
bases into a Common::Base class. Previously, this functionality was
spread across several classes, and used different approaches to formatting.
While the code still mixes C++ streams and C-style sprintf's, at least
it will now be easier to modify it all in one place.
Related to the above, added ability to use upper or lower case
characters for HEX output in the debugger. This is toggled by the
new debugger prompt command 'uhex', which is also tied to a new
commandline argument, 'dbg.uhex'.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2770 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
the CPU registers (A/X/Y/PS) on ROM load.
Added 'INTIM Clks' to the debugger I/O tab, showing the number of clocks
before the current INTIM value decreases by 1.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2768 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
in decimal, not hex.
All DataGridWidgets (ie, most of the inputs in the debugger) have more
strict input filtering, allowing to use $,#,\ specifiers for different
bases, as well as restricting input based on the specifier used (ie, if
you've entered a '\', only '0' and '1' are allowed, etc).
Updated libpng to latest version.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2767 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
will be used, but at least we can improve it a little).
Eliminated the 'alignment' dimension in the various TIATables masks,
reducing the size of the arrays by a factor of 4. I could never
figure out what alignment meant, until I looked at old TIA code in
the repo. It seems that originally, there was an optimization
in the code to fill the array on 32-bit boundaries, instead of the
current 8-bit boundary. As a result, the masks had to be defined
as 32-bits, or 4 groups of 8-bits. Ah, that's where the 'alignment'
comes from.
Related to this, the colors and pointers in the TIA class are now
8-bit as well. Essentially, the TIA class has been doing extra
work to align everything to 32-bit and never actually using the
results. And it's been this way for 4+ years.
Older state files will no longer work, so the version # has been
bumped.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2765 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
Cleaned up the class comments wrt ScummVM; the attributions only needed to
be in the base classes, since all the derived stuff was written later.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2763 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
in Linux/UNIX. This bug has been present since 3.2; too bad I didn't
find it 24 hours ago, as it could have been included in 3.9.
Bumped version #, and starting all over again ...
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2762 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
Added option to save snapshots using either internal database name
(the default) or actual ROM filename.
Extended FSNode API for parsing filenames and adding/removing extensions,
which was previously being done in different ways throughout the codebase.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2753 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
since it was getting too unwieldy for 'Config Paths', and didn't properly
belong there anyway.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2751 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
stephena