[tor-bugs] #19170 [Metrics/CollecTor]: make parsing more robust (extra-info)

[Tor Bug Tracker & Wiki]

#19170: make parsing more robust (extra-info)
-------------------------------+---------------------------------
 Reporter:  iwakeh             |          Owner:  iwakeh
     Type:  defect             |         Status:  accepted
 Priority:  Medium             |      Milestone:  CollecTor 1.0.0
Component:  Metrics/CollecTor  |        Version:
 Severity:  Normal             |     Resolution:
 Keywords:  ctip               |  Actual Points:
Parent ID:                     |         Points:
 Reviewer:                     |        Sponsor:
-------------------------------+---------------------------------

Comment (by iwakeh):

 Idea for an "accept non-ASCII characters" solution (follow-up of
 comment:3):

 As Karsten pointed out above (comment:2) this would require changes in
 metrics-lib, but not necessarily in the direction of having to distinguish
 Exception severity.

 Another option would be to make use of the `descriptor.parser` and
 `descriptor.reader` properties and supply a different non-ascci-accepting
 parser, let's call it `LenientParser`, as well as a LenientReader.

 * Necessary change in CollecTor would be to set the `descriptor.parser`
 and `descriptor.reader` properties to the LenientParser class.
 * Necessary change in metrics-lib would be the addition of the
 LenientParser, which consist mostly in providing additional ParserHelper
 methods that accept non-ascii and calling these in the appropriate places;
 most of the code will be the same as in the current, stricter
 implementation.  Also a LenientReader would have to be supplied.

 That way we could switch between implementations.
 Users of metrics-lib would also have another option.

 Another question that would need to be investigated: how will CollecTor
 clients deal with the additional non-compliant data?

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/19170#comment:7>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online