Tom Rybka

I conferred with @comius and I'm going to take point. As I understand it, it doesn't merge cleanly in our repo so I anticipate some back and forth to get...

Internal checks mostly look good. AFAICT no adverse breakages, but I'm re-running flaky failing tests anyway. I don't have a good solution yet for the failing test. I'm wondering if...

Yes that seems reasonable. If you want to do that I can merge your changes, otherwise I'll do something tomorrow.

Can you try rebasing? There are some non-trivial changes to input discovery here that I think you should probably take into consideration. Notably, the most recent change to CppCompileAction was...

Ack, thanks. I've re-imported, and re-running our Global Presubmits. I should have results tomorrow (need to run off-peak hours as it invalidates all the caches). Will post an update then.

Mailed for review internally -- local presubmits all green. Global presubmit hit a snag and didn't run. Scheduled another one for this evening, will try to check it early in...

To clarify, the ndk_cc_toolchain would need a new feature declaration, somewhere around here: https://github.com/bazelbuild/rules_android_ndk/blob/7e9abc68cf9d8016c8cfbe52ed3b1f8e68edbaad/ndk_cc_toolchain_config.bzl#L601 It would look like this: ``` # Disable fallback native deps linking in android binary rule...

> Hm I think it's not quite so simple. I have a draft PR in [#111](https://github.com/bazelbuild/rules_android_ndk/pull/111) that doesn't quite work as expected: I agree it's not quite so simple. I...

Selfishly, we have on the order of a dozen+ cc_toolchains that will need to add the `"system_include_paths"` feature for backwards compatibility until we can spend the time to test the...

Sorry for the naïve question, but do you have a concrete example or sample toolchain? It seems quite odd to me to have output directories like bazel-out be present in...