Tor allows non-ASCII chars in contact line

Tor allows non-ASCII chars in the contact line of router descriptors which

a) breaks tools parsing descriptors and

b) doesn't conform to dir-spec.txt.

Should we disallow non-ASCII chars in the future by being more strict when verifying torrc and when uploading descriptors to the directory authorities? Or should we change dir-spec.txt to allow more characters in the contact line? Or did I miss something?

[Automatically added by flyspray2trac: Operating System: All]

changed milestone to %Tor: 0.2.2.x-final

added component::core tor/tor milestone::Tor: 0.2.2.x-final owner::nickm priority::low resolution::wontfix severity::blocker status::closed type::enhancement version::0.2.1.22 labels

I think allowing utf8 in dir-spec is a wise move. ascii is dying, anyways, and should be (international domains, people who are running relays cannot spell their name out, etc)

Allowing UTF-8 is fine by me, but it needs a proposal. In the mean time, let's leave the implementation alone.

Looks like it's not just UTF-8, but arbitrary encodings that we allow. I found two non-ASCII strings, one of them using UTF-8 and the other one ISO-8859-1:

Florian Birée florian@bireen.name B i r UTF-8 é e 66 105 114 195 169 101

Jürgen Strobel juergen@strobel.info J ISOü r g e n SP 74 252 114 103 101 110 32

That means that we either need to allow arbitrary encodings in the spec, or if we want to use UTF-8, change the implementation.

arbitrary bytes it is. I'll patch the spec.

Trac:
Status: new to assigned
Milestone: N/A to Tor: 0.2.2.x-final
Owner: N/A to nickm
Parent: N/A to N/A
Keywords: N/A deleted, N/A added
Type: defect to enhancement

Today I learned it's not just the contact line, but also the platform line that may contain arbitrary bytes.

Trac:
Description: Tor allows non-ASCII chars in the contact line of router descriptors which

a) breaks tools parsing descriptors and

b) doesn't conform to dir-spec.txt.

Should we disallow non-ASCII chars in the future by being more strict when verifying torrc and when uploading descriptors to the directory authorities? Or should we change dir-spec.txt to allow more characters in the contact line? Or did I miss something?

[Automatically added by flyspray2trac: Operating System: All]

to

Tor allows non-ASCII chars in the contact line of router descriptors which

a) breaks tools parsing descriptors and

b) doesn't conform to dir-spec.txt.

Should we disallow non-ASCII chars in the future by being more strict when verifying torrc and when uploading descriptors to the directory authorities? Or should we change dir-spec.txt to allow more characters in the contact line? Or did I miss something?

[Automatically added by flyspray2trac: Operating System: All]

This conflicts with the spec patch that nickm wrote. See ticket #1270 (moved).

Here's an example of a platform line with non-ASCII chars:

platform Tor 0.2.1.26 on Windows XP Service Pack 1? [workstation] {personal} {terminal services, single user}

P.a.c.k. .1.?. .[

80.97.99.107.32.49.208.32.91

Also see branch the (old) branch dir-spec in my repository.

Trac:
Status: assigned to needs_review

I'm closing this as we're now tracking the spec update in #1720 (moved).

Trac:
Status: needs_review to closed
Resolution: None to duplicate

Trac:
Component: Tor Relay to Tor

Trac:
Reviewer: N/A to N/A
Points: N/A to N/A
Status: closed to reopened
Sponsor: N/A to N/A
Severity: N/A to Blocker
Cc: karsten,Sebastian,nickm to karsten, Sebastian, nickm
Actualpoints: N/A to N/A
Resolution: duplicate to N/A

set to wontfix as suggested by teor: https://lists.torproject.org/pipermail/tor-dev/2017-October/012487.html

Trac:
Status: reopened to closed
Resolution: N/A to wontfix

closed

mentioned in issue #1270 (moved)

moved to tpo/core/tor#1258 (closed)