Craig Tiller

Drive by comments: 1. there are no tests - we're in the process of removing iomgr and so any functionality that does not include tests I'd expect to be abandoned....

Lines of code probably doesn't matter here at all... what does is being able to reason about the code without having to do combinatorial thinking around ifdefs. As a concrete...

I dropped one comment, but architecturally this seems fine to me. @Vignesh2208 and @drfloob should review the iomgr changes and plan for their integration with event engine. @jtattermusch should review...

Under iomgr should be good yeah. On Thu, Sep 15, 2022, 5:02 PM Andres Beltran ***@***.***> wrote: > ***@***.**** commented on this pull request. > ------------------------------ > > In src/core/lib/iomgr/tcp_server_posix.cc...

Note: if we remove the ResetDefaultEventEngine here we end up with a thread leak issue.

@veblush probably best for you to make a call on where to take this

looking at the bloat numbers it seems like we're pouring too much code into AutoLoader and not enough into shared backends in the cc file

Just a drive by... there's a lot of systemic bloat here, anything prefixed with `grpc_core::json_detail`: ``` +292% +2.75Ki +292% +2.75Ki grpc_core::json_detail::AutoLoader::Insert() ``` That's probably fine to take on because, being...

@yihuazhang @ZhenLian @markdroth I don't think I'm sufficiently up to date with the TLS creds model to be an approver here, can someone else take that role?

yeah... things are getting a bit messed up here. mpscq includes log.h for GPR_ASSERT (since GPR_ASSERT depends on being able to call gpr_log)... fix_build_deps has gone and grabbed one of...