Oscar Bonilla

In our case, we have some `copts` in the `objc_library()` rules that include calls to `$(location)`. Specifically, `-I$(location headermap)` which the rules expand to the location of a generated header...

We wrote our own test runner for [bluepill]() called [`bptestrunner`](https://github.com/MobileNativeFoundation/bluepill/tree/master/bptestrunner). We use it something like this ``` load("@bptestrunner//:bluepill_batch_test.bzl", "bluepill_batch_test") bluepill_batch_test( name = "Some.xcscheme", test_targets = [ "SomeTests" ], config_file =...

You could, but in our case, we run `bazel-remote` in a distributed fashion. That is, each host that runs `bazel` has its own local copy of `bazel-remote` that proxies to...

I think the problem is that Deep copy should keep resolving the symlink until it gets to a file that is not a symlink no? So maybe a loop [here](https://github.com/otiai10/copy/blob/51bf83a372fd4023d4dde7ada7f7c4565c6d9af2/copy.go#L159)?

I think for shallow copies creating broken symlinks is the right answer. I believe that's how rsync works.

No, this only shows the problem, it doesn't have a solution. I haven't had time to figure out how to fix it.

> Does this part check the internal process state of the simulator to determine if its ready to run tests? We don't really know what this does... seems to work...

I managed to reproduce this locally in the test suite by enabling the Address Sanitizer for the `BPSampleApp` and running the UI test for `bp`. I think the bug is...

Are you sure you built the app for testing? The runner is generated by Xcode.

Thanks for picking this up! Looking forward to the PR :-)