[metrics-bugs] #30255 [Core Tor/sbws]: Add additional bandwidth file headers in sbws 1.2

[Tor Bug Tracker & Wiki]

#30255: Add additional bandwidth file headers in sbws 1.2
---------------------------+-----------------------------------
 Reporter:  teor           |          Owner:  (none)
     Type:  task           |         Status:  new
 Priority:  Medium         |      Milestone:  sbws: 1.2.x-final
Component:  Core Tor/sbws  |        Version:
 Severity:  Normal         |     Resolution:
 Keywords:                 |  Actual Points:
Parent ID:                 |         Points:
 Reviewer:                 |        Sponsor:
---------------------------+-----------------------------------

Comment (by karsten):

 Replying to [comment:7 juga]:
 > Replying to [comment:5 karsten]:
 > > irl and I discussed the topic of additional bandwidth headers as part
 of #30216, where we're trying to write a parser for these files.
 >
 > i'm unfamiliar with metrics ecosystem and i thought that implementing
 the parser in stem (#29056, #30160) would avoid you to re-implementing for
 metrics. Is there a way to avoid that?.

 Nope, there's no way to avoid that. The deployed Tor Metrics services are
 all implemented in Java, and we're currently starting to re-implement
 parts in Python. There will be a transition phase in which we'll need to
 write parsers in Java and Python.

 > > When reading this ticket and its subtickets I'm slightly concerned
 that you might consider adding spaces to new header lines.
 >
 > No, we should not add spaces. I made the mistake of writing the url of a
 wip branch in a ticket that should not have been reviewed yet.

 Not sure which URL or branch you refer to, but glad to hear that you agree
 about the spaces.

 > > For example, operating system and contact information sound like they
 might contain spaces.
 >
 > For the operating system, this is what the command suggested by teor
 would generate: `Linux-4.9.0-8-amd64-x86_64-with-glibc2.7`. It doesn't
 contain space.

 Okay!

 > For the contact information, teor said:
 >
 > > Let's not specify another unstructured contact string, please.
 >
 > And imo, it should just be an email address, if we add contact
 information.

 I'll leave this to you. I mostly wanted to know about the spaces, because
 that's what affects me when parsing the format.

 > > But as far as I see it, that means trouble! One reason is that
 `bandwidth-file-headers` in dir-spec doesn't permit spaces in `KeyValue`
 parts, so that it wouldn't be possible to copy over header lines to the
 vote.
 > >
 > > Does this make sense?
 >
 > Yes

 Cool!

 > > In fact, if it does, would you mind clarifying in the spec that header
 lines always have the format `HeaderLine ::= KeyValue NL`? That would also
 simplify the parser, because we'd then be able to distinguish header and
 relay lines by the number of space-separated `KeyValue` elements.
 >
 > Hmm, that makes sense, but it's not true for the first line `Timestamp
 NL`. Maybe we can just clarify that.

 Yes, I think it should be possible to clarify that. After all, that's a
 special line anyway.

 Thanks a lot!

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/30255#comment:9>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online