Use the X86_64 variable instead of AMD64 for this architecture in cmake.
For our upstream binary builds, use -march=core2 -mtune=rocketlake for
X86_64 on macOS, -march=core2 -mtune=znver3 for X86_64 on non-macOS (not
currently used) and -march=pentium-mmx -mtune=core2 for the future
Windows XP 32 bit MinGW build.
Signed-off-by: Rafael Kitover <rkitover@gmail.com>
Instead of suffixing release zip files with 32bit or 64bit, add an
architecture suffix, which would be one of:
x86_32
x86_64
arm32
arm64
Allow codesign to fail on mac without aborting.
Disable the interactive pause for gpg signing because there is no
passphrase anymore.
Signed-off-by: Rafael Kitover <rkitover@gmail.com>
m4: 1.4.18 -> 1.4.19, remove patches.
libxml2: 2.9.12
wxwidgets: master (3.0.5.1 on mac for 10.7)
Run script with brew bash.
Unlink brew wxwidgets before build and relink after because it conflicts
with the built version, this needs to be fixed.
Signed-off-by: Rafael Kitover <rkitover@gmail.com>
Add indirect dependencies that cmake complains about and remove ccache
because it currently fails to build on NixOS unstable.
Signed-off-by: Rafael Kitover <rkitover@gmail.com>
The present ini variable is from vbam 1.x and is behaviourally backwards where setting disablemmx to true actually enables mmx, 2.x builds should not be using the ini from 1.x for various reasons.
I don't know if this is right or not, but I am running an Arch system and tried running `./installdeps` and nothing happened. So I went into the script and figured that it got stuck in the `elif [ -f /etc/os-release ];` block. This file exists on Arch as well, so the block is executed, but arch is not listed in the `case`, so the script exited without doing anything. And it's only checking for pacman after that, so it never ran `arch_installdeps`.
Don't know if this is the 100% correct fix, but it's working for me now and should also catch cases like that in the future.
Add commands to delete the ssh-agent service, otherwise the installer
throws an error.
Add instructions for setting permissions on authorized_keys and private
key.
Signed-off-by: Rafael Kitover <rkitover@gmail.com>
- Minor refactoring, some variable name and type changes
- Remove unnecessary checks in clock latch
- Fix stuck in whitescreen after Hudson startup logo
- Fix Gamepak error message at initial load of game without prior rtc
data available to load
- Save the bare-minimum variable needed for rtc (saves only
Basetime/lastTime)
Savestates are broken due to the change in savestate struct size for
HuC3. Since save/load function also includes save data for other
non-active mappers, the change in size breaks every savesfiles being
made.
This PR refactors the HuC3 clock data struct to be separate from the
main struct, and then only added to save/load functions when HuC3 mapper
is used. This still breaks previous HuC3 states but other mappers should
now not get affected
Solution:
Savestate/LoadState should only include mapper data of current active
running rom, not all mappers even if they are inactive. This will break
every savefile but it would be better for the long run when changes are
neede to be done.
Rafael Kitover