Closed Bug 545866 Opened 15 years ago Closed 9 months ago

DNS: Need to query arbitrary DNS record types in Mozilla

Categories

(Core :: Networking, enhancement, P3)

enhancement

Tracking

()

RESOLVED WONTFIX

People

(Reporter: BenB, Unassigned)

References

(Depends on 1 open bug, Blocks 2 open bugs)

Details

(Whiteboard: [necko-triaged])

Attachments

(1 file, 2 obsolete files)

Current status: To look up hostnames, Mozilla simply uses C lib function getaddrbyname() from the OS. That's enough so far, and a good idea. However, there are other situations where we need to query other record types like SRV (bug 14328), TXT or MX (bug 342242). We need an IDL interface and implementation in Necko to do that, usable from C++ and chrome JS. Implementation: We don't need to write our own DNS resolver for that, the OS provides another function to make such lookups, namely dnsquerya() (Windows), res_query() (Unix) and friends. The patch in bug 14328 uses them, but restricts the API to SRV and TXT. The API in this bug should allow to query any DNS record type.
Assignee: nobody → ben.bucksch
Status: NEW → ASSIGNED
Started working on it. API changes are done, implementation follows.
Attached patch WIP v1 - API changes (obsolete) — Splinter Review
Depends on: 551519
Attached patch WIP, v5 (obsolete) — Splinter Review
Attachment #436833 - Attachment is obsolete: true
Attached patch Fix, v6Splinter Review
Attachment #437194 - Attachment is obsolete: true
Priority: -- → P2
Blocks: 551519
No longer depends on: 551519
Blocks: 563958
No longer blocks: 551519
Blocks: 265226
Blocks: 495623
Blocks: 589537
What does it take to get this patch committed?
It's a "work in progress", not done yet.
Anybody feel free to adopt the patch.
No longer blocks: 265226
@Ben: thanks for your work on this. How would you characterize the state of your patch now - feature incomplete, needs testing, etc?
I don't know if this was mentioned in any of the other related bug-reports, but there is a paper on adding DNSSEC-checks to applications here: http://conferences.npl.co.uk/satin/papers/satin2011-Hardaker.pdf it points to an existing patch on this page: https://www.dnssec-tools.org/wiki/index.php/DNSSEC_Applications I guess the patches are in the SVN-repository: https://www.dnssec-tools.org/svn/dnssec-tools/trunk/dnssec-tools/apps/mozilla/ It got the last updates for Firefox 3.6. I haven't checked if it works for 4 or applies to the current code.
Comment on attachment 437201 [details] [diff] [review] Fix, v6 FWIW, IIRC, this was mostly working, but had one of the new, included tests failing unexpectedly. That would need to be investigated.
Attachment #437201 - Attachment description: WIP, v6 → Fix, v6
:Fallen volunteered to take a look. I'm personally swamped at my day work currently.
Yes, I'll take a look and see what the state is on this. Calendar could make good use of this, detecting the server for a domain using DNS SRV and TEXT entries as in http://tools.ietf.org/html/draft-daboo-srv-caldav-10
Assignee: ben.bucksch → philipp
All, thanks for the patches - please see my comment in bug 14328 about how these two mugs fit with other work we're planning to do for DNS. Thanks!
Blocks: 735215
ping - can I ask if this needs anything to be landed? Is there something I can help agitate for? As an XMPP person I would love to see this land so that Thunderbird can get proper XMPP client support. thanks!
Hey bear, thank you very much for the offer! The patches are pretty much finished. As far as I remember, there was one of the new testcases that I wrote which sometimes failed. I didn't have time to investigate, and I stopped. Of course, the patch also has to be unbitrotted. The code is should be all written and finished, though. If you would like to adopt the patch and drive it to commit, I would be very grateful! Ben
(In reply to Mike Taylor [:bear] from comment #15) > ping - can I ask if this needs anything to be landed? Is there something I > can help agitate for? I think the most recent sign of activity on a related bug was in bug 735967, so it may be where you want to start if you want to find who would be most likely to make things move forward if you pinged them.
Hi all, Florian is right - bug 735967 has a framework patch uploaded for feedback. Please take a look and comment. Hoping to make some more progress on this in the next month.
Blocks: 787369
No longer blocks: 735215
Blocks: 861622
sorry, as much as I'd like to I don't think I will be getting around to this any time soon.
Assignee: philipp → nobody
Status: ASSIGNED → NEW
Whiteboard: [necko-backlog]
Priority: P2 → P1
Priority: P1 → P3
Compare bug 787369, which has a DNS.jsm that directly calls the native "resolv" library.
Severity: normal → S3

Bug 1852900 has covered this for RR records for https; extending that to cover the request from this bug seems like a quite useful improvement

See Also: → 1852900, 787369
Whiteboard: [necko-backlog] → [necko-prioity-review]
Whiteboard: [necko-prioity-review] → [necko-triaged][necko-priority-review]

Webextensions are no longer allowed to do arbitrary dns resolutions, so wontfix.

Status: NEW → RESOLVED
Closed: 9 months ago
Resolution: --- → WONTFIX
Whiteboard: [necko-triaged][necko-priority-review] → [necko-triaged]
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: