Isn't mdns the only service discovery implementation?  I wonder if enable_local_discovery without enable_mdns makes any sense.

I didn't write any of this code originally, so I'm not super knowledgable about all this.

FWIW, in chrome/common/features.gni, we have: enable_service_discovery = enable_mdns || is_mac

and in net/features.gni, we have: enable_mdns = is_win || is_linux

I haven't checked all the uses of the two variables to figure out when it's appropriate to use one vs. the other.

I forgot to mention the real motivation comes from being able to build with printing turned off. Previously, the build config that would do that is:

enable_print_preview = false
enable_basic_printing = false
enable_mdns = false
enable_pdf = false

But now this build config does not work anymore. Doing some local hacking, I came up with https://chromium-review.googlesource.com/624562 to get it working again. It's probably not right, but I'm just throwing the CL out there for those interested.

Bin, can you PTAL?

Mark, if enable_service_discovery = false, do we disable both DIAL and mDNS discovery, or just disable mDNS (DIAL still work, dual discovery still work, Cast devices will still be available).

No, I don't think we should change the meaning of the flag.  Setting it to false should just disable the local discovery backend, which is mDNS. I guess we need to guard some of the source we moved that depends on enable_service_discovery.

I don't really know the history or motivations for these flags, maybe reaching out to the local discovery OWNERs (gene@ and vitalybuka@) might shed some light.