status: optimize GRPCStatus() calls

[ash2k]

RELEASE NOTES: none

[arvindbr8]

Hi @ash2k, seems like the GRPCStatus() is only a getter and side effects (if any) arent obvious to me now. I'm curious how did you find this? Has this caused some sort of regression in callsites?

[ash2k]

Hi. It's a getter on gRPC's implementations, but it might be doing something in other implementations. It shouldn't have any side effects, but it may be allocating extra memory (e.g. to construct the object its returning), which is not required to do twice.

It didn't cause any problems, I was just looking at the gRPC code while working on my project and spotted this.

[arvindbr8]

diff lgtm.

I'm less concerned about the double calls since gs.GRPCStatus() is only a getter and does not have other side-effects (for eg, incrementing a counter).

cc: @easwars for another look

[arvindbr8]

closing and opening to re-run new tests

[ash2k]

I don't know why the vet-proto job is failing. I've just rebased the branch but it didn't help.

[easwars]

I don't know why the vet-proto job is failing. I've just rebased the branch but it didn't help.

vet-proto does not block merge. I've sent #6551 to fix the failure in the meantime.