search

knulli-cfw / distribution

Main KNULLI distribution
https://knulli.org
Other
347 stars 25 forks source link

Compile issue #75

Closed bladeoner closed 2 months ago

bladeoner commented 2 months ago

Knulli build version

40-dev-4b893c35df 2024/07/11 08:29

Your architecture

RG35XX H

Issue description

I have tried to compile the latest source, but it shows an error. The images are created and can be flashed, but the database sources under scraper seem to be missing.

IMG-20240711-WA0010~2

Detailed reproduction steps

I have performed the following steps:

Details of any attempts to fix this yourself

Tried to compile twice, first time with the mentioned source code. The second time with the mentioned source code, but I updated the batocera-emulationstation package to the latest version, both attempts the same results and error.

Details of any modifications you have made to Knulli.

I just downloaded the source and started to compile it in a docker container.

Logs and data

python /build/package/batocera/emulationstation/batocera-es-system/batocera-report-system.py /build/package/batocera/emulationstation/batocera-es-system/es_systems.yml /build/package/batocera/emulationstation/batocera-es-system/systems-explanations.yml /build/package/batocera/core/batocera-configgen/configs /tmp/br_systemreport_3516111/configs Traceback (most recent call last): File "/build/package/batocera/emulationstation/batocera-es-system/batocera-report-system.py", line 218, in EsSystemConf.generate(args.yml, args.explanationsYaml, args.configsDir, args.defaultsDir) File "/build/package/batocera/emulationstation/batocera-es-system/batocera-report-system.py", line 80, in generate emulators = EsSystemConf.listEmulators(arch, system, rules[system], explanations, config, defaultEmulator, defaultCore) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/build/package/batocera/emulationstation/batocera-es-system/batocera-report-system.py", line 182, in listEmulators raise Exception("default core ({}/{}) not enabled for {}/{}" . format(defaultEmulator, defaultCore, arch, system)) Exception: default core (libretro/flycast) not enabled for h3/systemsp '+' exit 1 make: [Makefile:821: target-post-image] Error 1 make: Leaving directory '/build/buildroot' make: [Makefile:104: h700-build] Error 2

bladeoner commented 2 months ago

I think I found the anwser, I will close the issue.

jbbandos commented 1 month ago

@bladeoner Would you mind sharing the answer? I am having the same issue.

bladeoner commented 1 month ago

Hi @jbbandos , I didn't manage to fix this error because the compiling is working. I didn't look into it anymore.

I think it's also not easy to fix.

jbbandos commented 1 month ago

@bladeoner But you now have compilation working? Because for me it stops here, after generating only the rg28xx image. I am unable to finish building the other images, as it always ends the same way: `+ cp /h700/target/usr/share/batocera/batocera.version /h700/images/knulli

bladeoner commented 1 month ago

What you can do is use the script in the following PR: https://github.com/knulli-cfw/distribution/pull/98

You can not compile all devices directly after each other. You need to compile them one by one.

jbbandos commented 1 month ago

I Still have the same issue at the end, but at least have the "batocera" 1.7GB boot image. Checking the configs, this seems to be an issue only with the validation tests, as FLYCAST is explicitly disabled for h3 and bcm2836 in config.in: select BR2_PACKAGE_FLYCAST if !BR2_riscv && \ !BR2_PACKAGE_BATOCERA_TARGET_BCM2835 && \ !BR2_PACKAGE_BATOCERA_TARGET_BCM2836 && \ !BR2_PACKAGE_BATOCERA_TARGET_BCM2837 && \ !BR2_PACKAGE_BATOCERA_TARGET_RK3326 && \ !BR2_PACKAGE_BATOCERA_TARGET_X86 && \ !BR2_PACKAGE_BATOCERA_TARGET_H3

And of course, we're building for h700, not H3 or bcm2836, so no idea why the tests are being done for those two

jbbandos commented 1 month ago

@bladeoner I ended up editing scripts/linux/systemsReport.sh and changing line 3 to read ARCHS="h700" . Like that, no other arch is tested, no more spurious errors. There is probably a better way to do this, filling that ARCHS line automatically with the arch you've built...