.NET Core Runtime ARM32 builds now available

原文地址：传送门

.NET Core Runtime ARM32 builds now available

The .NET Core team is now producing ARM32 builds for .NET Core Runtime 2.0+, for both Linux and Windows. These builds are great for using on Raspberry Pi. The builds are not yet supported by Microsoft and have preview status. Please give us feedback.

Docker ARM32 images are also available at microsoft/dotnet. See microsoft/dotnet Docker Hub README adopts segmented OS/Arch style for a related announcement.

There are two good sources of .NET Core ARM32 samples that you can use to get started:

• .NET Core on Raspberry Pi
• .NET Core Docker Raspberry Pi Samples

Discussion

You can share your thoughts on this change or see what others are saying at:

• dotnet/coreclr #13369

Details

• dotnet/coreclr label:arch-arm32

The team is producing Runtime and not SDK builds for .NET Core. As a result, you need to build your applications on another operating system and then copy to a Raspberry Pi (or similar device) to run. The samples provided above show you how to do that.

You can get builds from the following links. Official builds are not yet provided.

• .NET Core 2.0.0 ARM32 build
• .NET Core 2.0 servicing builds
• .NET Core 2.1 daily builds

Note that ARM32 is sometimes referred to as armhf, which refers to hardware floating point capability.

You can remote debug Linux ARM32 applications from a Windows, macOS or Linux machine.

The SDK has not been made available for ARM32 to yet as it doesn't fit within a reasonable performance envelope (space and time) for ARM32. We will resolve this although it isn't a top priority currently (unless feedback suggests otherwise). Our belief is that building on a desktop or build machine will be both more efficient and a reasonable work around to doing development on the Pi for the time being. We also wanted to prioritize the runtime being available since it's necessary for anything to work.

A more near-term need is unit testing on ARM32. There is no replacement for running tests on device. We intend to publish short-term workarounds for this and work on a longer-term solution that is easy to use.