You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Given the possible fragility of the build environment for a cross compiler such as this, it seems like it would be really useful to have an automated build available on Docker Hub for people to download (sans Xcode).
The way I would envision it working is that the base docker image that gets automatically published would install using a known distro and known packages that actually work (similar to your instructions in #45), and also does the work of installing osxcross in that environment.
Then a second build script that one would have to run locally would download that image, copy the right xcode bits in the image, and then change the internal docker userid/groupid to match the local user credentials. This image would be saved locally.
Then, there would be an environment script that would launch the final image, and mount a volume inside the image whereever you specified. This would give you a shell into the environment, which then compiles could be executed reliably regardless of how your host OS packages change.
I would contemplate working on this, but I need to package a python product for OSX that has binary requirements, and I don't think pyinstaller would work in a cross-compilation environment.
The text was updated successfully, but these errors were encountered:
Btw, this is exactly the approach taken by xgo for C embedded/dependent Go cross compilation. You might want to look at those dockerfiles for inspiration. I pretty much have everything you need ;)
Given the possible fragility of the build environment for a cross compiler such as this, it seems like it would be really useful to have an automated build available on Docker Hub for people to download (sans Xcode).
The way I would envision it working is that the base docker image that gets automatically published would install using a known distro and known packages that actually work (similar to your instructions in #45), and also does the work of installing osxcross in that environment.
Then a second build script that one would have to run locally would download that image, copy the right xcode bits in the image, and then change the internal docker userid/groupid to match the local user credentials. This image would be saved locally.
Then, there would be an environment script that would launch the final image, and mount a volume inside the image whereever you specified. This would give you a shell into the environment, which then compiles could be executed reliably regardless of how your host OS packages change.
I would contemplate working on this, but I need to package a python product for OSX that has binary requirements, and I don't think pyinstaller would work in a cross-compilation environment.
The text was updated successfully, but these errors were encountered: