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
The commit d96a3a2 makes hostname entries in local database lowercase forcibly during go1.6 development. I don't think it's a good idea because /etc/hosts-like local database is a last resort for name resolution and https://tools.ietf.org/html/rfc4343 never says that "don't preserve case on input and output" for some reasons; perhaps it was for internetworking with old-plain systems and it is just for future use.
I don't see any rationale for the change in d96a3a2, so can you please restore the behavior that preserves case on input and output?
Note that we know a few name resolver libraries don't preserve case of hostnames and DNS labels, for example libc for OS X/iOS that supports both unicast and multicast DNS, but I still don't think we need to do case conversion for now especially with inconsistency that providing case-converted hostnames and case-preserved DNS labels.
The text was updated successfully, but these errors were encountered:
I think I've already described that that commit fixed a bug in a wrong way. Like handling DNS labels, we need case insensitivity comparison with preserving case on input and output, not case conversion on input and output.
Lowercased host is only preserved in the internal cache for comparison. It is not mutating anything. Could you explain the case with a snippet where it breaks things? I guess I am missing something.
The commit d96a3a2 makes hostname entries in local database lowercase forcibly during go1.6 development. I don't think it's a good idea because /etc/hosts-like local database is a last resort for name resolution and https://tools.ietf.org/html/rfc4343 never says that "don't preserve case on input and output" for some reasons; perhaps it was for internetworking with old-plain systems and it is just for future use.
I don't see any rationale for the change in d96a3a2, so can you please restore the behavior that preserves case on input and output?
Note that we know a few name resolver libraries don't preserve case of hostnames and DNS labels, for example libc for OS X/iOS that supports both unicast and multicast DNS, but I still don't think we need to do case conversion for now especially with inconsistency that providing case-converted hostnames and case-preserved DNS labels.
The text was updated successfully, but these errors were encountered: