Nice! LGTM (fix the build break with the extra return)

@jhowardmsft - Do we want to remove the event log logging hook?

@jterry75 don't think we should remove the event log hook yet. We can register a service to cause these ETW events to show up in event log, and then removing it's a good idea, but we should do that first in case anyone relies on it.

@kevpar - I agree for Docker but this is new in CD anyways. Should we just effectively 'never add it'?

@jterry75 We might want to check with Nick first, I think event log -> fluentd might be their plan for logs.

Nice! LGTM (fix the build break with the extra return)

Removed. (Odd that lint didn't flag it on my dev box, but whatevs....)

@kevpar - I agree for Docker but this is new in CD anyways. Should we just effectively 'never add it'?

@jterry75 Let's take any action on this as a follow-up. Not related to this PR itself.

Looks like the vendor on linux is broken with this change.

These files were modified:
 M vendor/github.com/Microsoft/go-winio/internal/etw/etw.go
 M vendor/github.com/Microsoft/go-winio/internal/etw/eventdata.go
 M vendor/github.com/Microsoft/go-winio/internal/etw/eventdatadescriptor.go
 M vendor/github.com/Microsoft/go-winio/internal/etw/eventdescriptor.go
 M vendor/github.com/Microsoft/go-winio/internal/etw/eventmetadata.go
 M vendor/github.com/Microsoft/go-winio/internal/etw/eventopt.go
 M vendor/github.com/Microsoft/go-winio/internal/etw/fieldopt.go
 M vendor/github.com/Microsoft/go-winio/internal/etw/provider.go
 M vendor/github.com/Microsoft/go-winio/internal/etw/providerglobal.go
 M vendor/github.com/Microsoft/go-winio/internal/etw/ptr64_32.go
 M vendor/github.com/Microsoft/go-winio/internal/etw/ptr64_64.go
 M vendor/github.com/Microsoft/go-winio/pkg/etwlogrus/hook.go

The command "../project/script/validate/vendor" exited with 1.

@crosbymichael Fixed now. Some Windows-style line endings had crept into go-winio. Updated that repo, cut a new release and updated the vendoring commit again.

Codecov Report

Merging #3036 into master will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master    #3036   +/-   ##
=======================================
  Coverage   43.87%   43.87%           
=======================================
  Files         102      102           
  Lines       10903    10903           
=======================================
  Hits         4784     4784           
  Misses       5384     5384           
  Partials      735      735

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 0769763...d83e4e9. Read the comment docs.

Overall, this looks good. I think we can move the initialization in an init() just to clean up the main.

Overall, this looks good. I think we can move the initialization in an init() just to clean up the main.

@crosbymichael Yes, makes sense. Updated.

LGTM