If we extract in-place, doesn't that remove the possibility of conditionally skipping the thinning in the future? Or would we be relying entirely on xcode to skip

If we extract in-place, doesn't that remove the possibility of conditionally skipping the thinning in the future? Or would we be relying entirely on xcode to skip

The existing UnpackIOS assemble target called from the Run script build phase (pre-compile) copies the Flutter framework from the artifacts cache into the mode-specific directory like build/ios/Debug-iphoneos/, but skips if it's already there (unless the artifacts changed). Then the Thin Binary Xcode build phase (post-compile) always copies from there into the app framework directory Runner.app/Frameworks and is never skipped. So it's a new copy of the engine artifacts fat framework with all its archs on every build. So destroying it in-place doesn't matter since it will be re-copied on the next build.

We could skip thinning in this new target by validating the input framework already only contains the requested archs. I can look into that performance improvement in a follow up PR, I was mostly just trying to refactor this one out of the bash script without changing too much.

Actually come to think of it the App.framework probably never needs to be thinned--it was built with the same requested archs. I'll also look into that in a new PR.