You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
go mod tidy -e should do whatever tidying it can while ignoring errors and it indeed does it quite well. Unfortunately, when some remote host can not be resolved (either directly or by module proxy), it fails hard, as following:
go: finding module for package funny.host.com/some/unavailable/package
funny.host.com/some/unavailable/package: cannot find module providing package funny.host.com/some/unavailable/package: unrecognized import path "funny.host.com/some/unavailable/package": https fetch: Get "https://funny.host.com/some/unavailable/package?go-get=1": dial tcp: lookup funny.host.com: no such host
I believe that it should not fail hard, but rather treat the network/DNS error as "skippable", in line with all the other possible errors.
For example, in my particular case, the missing packages are not really missing - they belong to the same project, but need to be generated (things like protobuf schema and such). Yet, I cannot tidy the rest of the project without making the remote git repo somehow available.
The text was updated successfully, but these errors were encountered:
I have tested this myself and it's not quite conclusive; it is possible it was some sort of glitch on my side.
Or, may be the error which slips through is not of trivial nature, and happens somewhere between network failure, dns failure or sum db error response. But then, it may not be worth the time to hunt it down.
Sounds like we need to figure out how to actually reproduce these errors. (If we can find a way to at least identify the root cause, then I agree that we should fix them, but I agree that it might not be worth spending a lot of time hunting it down.)