Bug 1059122

Summary: libc getaddrinfo fails for hostname containing only decimal digits
Product: [Fedora] Fedora Reporter: Ariel Glenn <ariel.glenn>
Component: glibcAssignee: Carlos O'Donell <codonell>
Status: CLOSED NOTABUG QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: unspecified Docs Contact:
Priority: unspecified    
Version: 20CC: codonell, fweimer, jakub, law, pfrankli, spoyarek
Target Milestone: ---   
Target Release: ---   
Hardware: x86_64   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2014-01-31 12:28:35 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Attachments:
Description Flags
test-getaddrinfo.c does what the name says none

Description Ariel Glenn 2014-01-29 08:32:34 UTC 
Created attachment 856951 [details]
test-getaddrinfo.c   does what the name says

Description of problem:

getaddrinfo fails with an error for a hostname that consists only of digits 0-9.  Docker for example generates such hostnames for its containers (it generates hex strings but some of those turn out to contain only 0-9).

Version-Release number of selected component (if applicable):

glibc-2.18-11.fc20.x86_64

Steps to Reproduce:

1. add the lines to /etc/hosts:
172.17.10.53    blobber
172.17.10.54    836937931829

2. compile the attached c program (just sets up a struct addrinfo and calls getaddrinfo() and run it

Actual results:

$ ./test-getaddrinfo 172.17.10.53 80
Success!
$ ./test-getaddrinfo blobber 80
Success!
$ ./test-getaddrinfo 172.17.10.54 80
Success!
$ ./test-getaddrinfo 836937931829 80
getaddrinfo: Name or service not known

Expected results:
success for all four runs

Additional info:
this also breaks getaddrinfo() in python.
Comment 1 Siddhesh Poyarekar 2014-01-31 12:28:35 UTC 
Numeric arguments are considered to be IP addresses and not valid host names.  The amendment to RFC 952 (with RFC 1123) may allow for all-numeric host names, but that conflicts with this documented documented IP address format:

http://pubs.opengroup.org/onlinepubs/009695399/functions/inet_addr.html

There is an explicit note in RFC 1123:

           If a dotted-decimal number can be entered without such
           identifying delimiters, then a full syntactic check must be
           made, because a segment of a host domain name is now allowed
           to begin with a digit and could legally be entirely numeric
           (see Section 6.1.2.4).  However, a valid host name can never
           have the dotted-decimal form #.#.#.#, since at least the
           highest-level component label will be alphabetic.

which technically allows us to look at a number as an IP address and if that fails, fall back to a hostname lookup.  This would be very confusing behaviour though, so I don't think it's a good idea to change the current state of the art.

So to conclude, please don't use only numeric hostnames.