The _get_repos_mupen64plus function is used for update checks, and it's called before depends_mupen64plus in this case.
If cmake was removed after installing, the hasPackage would fail and an older commit for the GLideN64 repo would
be used, which would cause an update even if not required.
This is unlikely to happen for end users, but the builder module removes dependencies after installing (so missing
dependencies can be identified). This caused mupen64plus binaries to be rebuilt when there had been no changes.
This check is only needed for older pre-buster distros with cmake older than 3.9 so will be removed in the future.
* dropped hotkey hack patch as now integrated with config option input_hotkey_block_delay with default of 5 (as was our patch)
* rebased disable search patch (code moved back to retroarch.c for now)
* rebased / updated shader path config enable patch (thanks @cmitu)
* revert the default saves/states/sytem path change as we rely on the old behaviour
* default to using ON/OFF text instead of rgui_switch_icons graphics
* added gles3.1/3.2 configure parameters using gles31/gles32 flags
* added gles31 flag to rpi4 system. Mali boards should support 3.2 but need to test before adding
Re-order configure code to avoid nested if statement and use case/esac for setting vars for dosbox/dosbox-sdl2
Only run setBackend configuration for dosbox
Need to do some testing as to whether it's worth using SDL2 for all platforms - but
enabling SDL2 just for kms seems unnecessary - and with sdl2 it should work on mali etc.
The configure_dispmanx_* functions are now just a _backend_set_fuse function which is called
by the backends module for modules that have this. This allows a module to
Previously the module used XINIT: to force X on kms systems (rpi4), but now with the new backend
configuration, we can use the same emulator.cfg launch parameters, but default it to X11 for kms.
Also remove the X11 launch code, as runcommand will take care of this now.
Moved the logic around also, to avoid install related configuration work on removal.
Turns out `-m` is not a shorthand for `-machine`.
Changed the names emulators shown in runcommand selection, the '+' may be problematic when storing the entry in emulators.cfg
There are audio issues on the RPI with the current HEAD version. But also we have had to update this
scriptmodule a fair amount due to upstream changes, and it would be better now to stick to the latest
stable version. Previously there were less frequent "stable" releases also.
The v3.5 version seems to work without issue.
Split off the code to get the various repos depending on platform, adding default branch names.
Add a _pkg_info_mupen64plus function which takes get/newer/check parameters from the main packaging code.
If "get" is called, we check all used repositories, getting the last commit hash of each repo, and make
a md5 checksum which is stored in pkg_repo_extra.
If "newer" is called, we check the remote hashes of all repositories, make a md5 checksum and compare against
installed package pkg_repo_extra.
Switched the hash for GLideN64 for older cmake to a full hash, so we generate the same md5sum also.
If "check" is called (only currently used by our tools module for checking repo validity) it checks all the
repos, outputting any errors, and returns 1 if any of them failed.
This is a lot of extra code, but it's only this module which requires this complexity - we could have just
used the main repository, but as some of the plugins are more actively developed at times, it made sense to
want to update if any of the plugin repositories were changed also. This seemed the best way to do it, but
the additional hook functionality could be used with other modules in the future if it was needed.
* Amiga - new extensions supported by `lr-puae`
* C64 - new extensions supported by `lr-vice`
* 3DO - added `.chd`, supported by the Opera Libretro core; updated license URL for the core.
* GC/Wii - added `.rvz`, supported by Dolphin
* MSX - added `.cas`, supported by both `lr-bluesmsx` and `openmsx`
* PC88 - added `.m3u`, supported by `lr-quasi88`
Users reported that this version runs faster on videocore on the rpi1/2/3
Adjusted ppsspp sources to allow branch parameter and fixed build subfolder to ppsspp
so that paths are the same for md_ret_files, including adjusting lr-ppsspp in the same way.
Older code builds ok with our current ppsspp module script on rpi1/2/3 + videocore gles2,
but not rpi4, which it isn't needed on anyway. This may need changing if we have to update
ppsspp further in the future.
* capsimage version workaround is no longer needed in hatari's cmake
* old capsimage types are no longer needed
* align caps include location with new hatari cmake code
* remove no longer needed __cdecl define hack when building
