Closed xplshn closed 5 months ago
BigDL uses the .json at each $ARCH and $ARCH/Baseutils. If the format is different across architectures it makes it difficult because I'd have to do jank such as checking the architecture at run-time.
Once again I urge you to only use https://bin.ajam.dev/METADATA.json This file remains immutable and is never overwritten in case any of the other json data are bad/empty/invalid. Only when every other source is validated, the merged METADATA.json is created. It's likely that you experienced https://github.com/Azathothas/Toolpacks/issues/10 because you keep relying on unreliable sources. Also, I am planning to deprecate hysp's metadata entirely. No more description etc in the short term (you may still use it, it's just that I won't be fixing any bugs or maintaining hysp aymore) , but in the long term, I plan to add a better system which would be automatable and not rely on github's api for description etc.
I would still recommend the solution of creating a repo for bigdl bigdl-registry
and workout formats on your own as discussed in https://github.com/Azathothas/Toolpacks/issues/18#issuecomment-2109586322
I would still recommend the solution of creating a repo for bigdl bigdl-registry and workout formats on your own as discussed in https://github.com/Azathothas/Toolpacks/issues/18#issuecomment-2109586322
I don't understand Github Actions at all, if I knew how to make a GH action that runs every a few hours and adds the missing Description/Version/Updated field to METADATA.json
that'd be great
EDIT: I just realized; This may be a bug not in my program...
Once again I urge you to only use https://bin.ajam.dev/METADATA.json
Because, I AM using METADATA.json. After reading all of my source code I couldn't find what caused that http request. (In termux I had an error message when trying bigdl saying: "https://bin.ajam.dev/arm64_v8a_Android/Baseutils/METADATA.json 404")
Now it works fine
Currently testing. (It may be available but inaccurate at the time of you seeing)
Further, if I were to generate a tree, which format would you prefer (be easier to parse, for eg, just download this file and extract the dirnames directly)
or if you want to rely only on the single root metadata.json look for those whose size says
-1 B
for example: