systemd'oh! DNS lib underscore bug bites everyone's

File: 1473000121505.jpg (54KB, 480x451px) Image search: [Google]

54KB, 480x451px

>netflix

>>61549719
temporary patch https://github.com/keszybz/systemd/commit/0926f34
issue https://github.com/systemd/systemd/issues/6426#issuecomment-317203834

>>61549719
>netflix

How bad is the quality control at Red Hat/systemd?

>>61549719
>systemd on gentoo

served him right

>>61550285
how about you stop bitching about proprietary shit

>inb4 muh gnu

>>61549719
does anybody know of a domain name with _ in it?

>DNS resolver in PID 1
Surely this cannot go wrong in any way.

File: 1416643951908.jpg (48KB, 511x512px) Image search: [Google]

48KB, 511x512px

>>61549719
Underscores aren't valid characters in DNS hostnames. The external library systemd uses (libidn2) responds correctly to spec by not processing the invalid request.

So it's not a bug and not part of systemd.

>https://www.ietf.org/rfc/rfc1035.txt
>RFC1035 - DOMAIN NAMES - IMPLEMENTATION AND SPECIFICATION
>section 2.3.1 - any one of the 52 alphabetic characters A through Z in
upper case and a through z in lower case... any one of the ten digits 0 through 9

Its an invalid character sempai.

Granted, it COULD be used in private domains on custom DNS software, seemingly the '_' character shouldnt be recognized as a valid character for public DNS lookups.

Can anyone give an example of a site that actually uses '_' in its domain?

>>61549719

What was Poettering's snarky reply?

>>61550476
>Can anyone give an example of a site that actually uses '_' in its domain?
ipv6_1-cxl0-c088.1.lhr004.ix.nflxvideo.net

Let me quote few lines from the 10 year younger RFC on DNS:
>any binary string whatever can be used as the label of an resource record.
>Implementations of the DNS protocols must not place any restrictions on the labels that can be used
So, not only is _ fine but so are things like Яндeкc.pф or sää.fi for example. I don't have PoetteringDNS so I can't test if it fails on those too or was _ just a special character that caused it to NOT-A-BUG/WONT-FIX

>>61551004
interesting. Which rfc is this?