Please read this before reporting a bug:
https://wiki.archlinux.org/title/Bug_reporting_guidelines
Do NOT report bugs when a package is just outdated, or it is in the AUR. Use the 'flag out of date' link on the package page, or the Mailing List.
REPEAT: Do NOT report bugs for outdated packages!
https://wiki.archlinux.org/title/Bug_reporting_guidelines
Do NOT report bugs when a package is just outdated, or it is in the AUR. Use the 'flag out of date' link on the package page, or the Mailing List.
REPEAT: Do NOT report bugs for outdated packages!
FS#7216 - Firefox 2.0.0.3 IPv6 works, except when both A and AAAA DNS records exist
Attached to Project:
Arch Linux
Opened by Anonymous Submitter - Sunday, 20 May 2007, 05:05 GMT
Last edited by Alexander Baldeck (kth5) - Wednesday, 01 August 2007, 22:22 GMT
Opened by Anonymous Submitter - Sunday, 20 May 2007, 05:05 GMT
Last edited by Alexander Baldeck (kth5) - Wednesday, 01 August 2007, 22:22 GMT
|
DetailsHi,
I've recently got a IPv6 6to4 tunnel going again, and have had complete success performing IPv6 functions such as DNS lookups via dig/host, ping6, traceroute6 and tcptraceroute6 (part of the ndisc6 package I've put together over at aur.archlinux.org) However, I'm having a problem where if I put a URL into firefox that I know has both AAAA and A records (e.g. http://www.kame.net), firefox seems to be ignoring the AAAA record result, and connects to the website via IPv4. I've performed tcpdumps of the DNS lookups for these addresses, and see both AAAA and A records being requested and responded with. If I put the returned AAAA IPv6 address into the URL in firefox, that works too (e.g. the below is for www.kame.net, and I then get the dancing Kame image) http://[2001:200:0:8002:203:47ff:fea5:3085]/ I have the network.dns.disableIPv6 firefox option left at the default of false, which I understand to mean that firefox should not ignore AAAA results if they are returned. I've also found if the DNS name only has an AAAA record, I'm also able to successfully connect. I've also tried "option inet6" in resolve.conf, although from the description of what it did I didn't think it would probably help much, and it didn't seem to. I've also done a bit of searching with google and I can't seem to find anybody else describing this type of problem or behaviour. So I'm stumped. The thought I'd had was that maybe the archlinux firefox package isn't "ipv6 enabled", except that that really means functionally IPv6 is always enabled in firefox, just that it will ignore AAAA records when there is also an A record available. I thought that might be some sort of compile option, so I figured my first step would be to lodge a package bug and see what came of that. Thanks, Mark. |
This task depends upon
Closed by Alexander Baldeck (kth5)
Wednesday, 01 August 2007, 22:22 GMT
Reason for closing: Won't fix
Additional comments about closing: seems to be a problem in how firefox handles address resoltion with glibc. upstream problem.
Wednesday, 01 August 2007, 22:22 GMT
Reason for closing: Won't fix
Additional comments about closing: seems to be a problem in how firefox handles address resoltion with glibc. upstream problem.
https://bugzilla.mozilla.org/show_bug.cgi?id=355087
> Necko just uses the addresses in the order in which getaddrinfo returns them...
I have seen that for some sites that have both A and AAAA records, Firefox uses the IPv6 address (even though I might not have a route to it ;)).