On 13 Oct 2017, at 09:17, nusenu nusenu-lists@riseup.net wrote:

Hi,

I'm trying to find out if non-ASCII characters are allowed in descriptor data (and what onionoo would do with non-ASCII currently displayed as "?").

According to the spec non-ASCII chars are not allowed: https://gitweb.torproject.org/torspec.git/tree/dir-spec.txt#n218

I searched trac because I remembered there were some ASCII related tickets:

https://trac.torproject.org/projects/tor/ticket/1258 closed due to https://trac.torproject.org/projects/tor/ticket/1720 (which has nothing to do with non-ASCII) https://trac.torproject.org/projects/tor/ticket/1270

Is this summary correct?

Currently non-ASCII is not compliant with the spec but your consensus is that non-ASCII chars should be supported but no one got around to patch the spec and implementation.

Non-ASCII characters are supported in the contact and platform fields. Some contact fields use non-ASCII for accented characters, and relay operators want us to continue to support them.

https://gitweb.torproject.org/torspec.git/tree/dir-spec.txt#n2011

There is a ticket about requiring ASCII in relay descriptors, which we should wontfix. There is a ticket about requiring ASCII in extrainfo descriptors, which we closed because we fixed the underlying memory disclosure bug.

T