dbdeployer icon indicating copy to clipboard operation
dbdeployer copied to clipboard

arm64 downloads

Open jpike88 opened this issue 3 years ago • 5 comments

add some downloads for arm64, for M1 Macs. looks like just 8.0.27 but that's a good start

jpike88 avatar Jan 14 '22 05:01 jpike88

makes sense to define a new OS: "DarwinARM64"

jpike88 avatar Jan 14 '22 07:01 jpike88

I have a couple of problems with this task at the moment:

  1. I don't own an M1 Mac. I can't just add the downloads for the new architecture, as it would require an update in the code, and I would also need to test the new downloads.
  2. Even if I find someone who would let me access an M1 remotely, I am unable to code now and I will be for several weeks, as an accident has deprived me of the use of my dominant hand.

That said, if someone would submit a pull request with the code changes and the downloads, I can have a look.

datacharmer avatar Jan 15 '22 07:01 datacharmer

No problem I could have a crack. Just a question regarding how you want to differentiate apple from intel chips in the list.json?

jpike88 avatar Jan 15 '22 08:01 jpike88

That depends on how an M1 defines itself. The operating system used in the tarball list must be detectable by the function CheckTarballOperatingSystem (in common/checks.go). As you can see from that function, figuring out the right type requires some experimentation. Probably it would be necessary replacing every occurrence of runtime.GOOS into a function that combines OS and architecture.

datacharmer avatar Jan 15 '22 10:01 datacharmer

that looks about right. I ran locally and got these values:

runtime.GOARCH => arm64 runtime.GOOS => darwin

jpike88 avatar Jan 15 '22 10:01 jpike88

I have opened a new branch that should fix this issue. (https://github.com/datacharmer/dbdeployer/tree/arm-64-downloads) It would be nice if someone with access to a arm64 architecture would compile this branch and test the downloads commands: get-by-version, add, add-remote, list, tree

datacharmer avatar Aug 15 '22 12:08 datacharmer

Release v1.69.0 resolves this issue

datacharmer avatar Aug 27 '22 11:08 datacharmer