[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: [tor-bugs] #19842 [Metrics/metrics-lib]: offer a `LenientParser` option with metrics-lib

#19842: offer a `LenientParser` option with metrics-lib
---------------------------------+-----------------------------------
 Reporter:  iwakeh               |          Owner:  karsten
     Type:  enhancement          |         Status:  needs_information
 Priority:  Medium               |      Milestone:
Component:  Metrics/metrics-lib  |        Version:
 Severity:  Normal               |     Resolution:
 Keywords:                       |  Actual Points:
Parent ID:                       |         Points:
 Reviewer:                       |        Sponsor:
---------------------------------+-----------------------------------
Changes (by karsten):

 * status:  new => needs_information


Comment:

 Hmm, I don't know how to proceed with this issue.  A few thoughts:

  - We should only use new properties for new implementations of a parser
 or reader, not for a config option.  If we want a config option, let's add
 a method to the descriptor source.
  - `DescriptorReaderImpl` internally uses a `DescriptorParser` for parsing
 descriptor.  If we just want to switch to another parser, let's create a
 method `DescriptorReader#setDescriptorParser()`.  Or let's use
 `DescriptorSourceFactory` internally to create a parser instance rather
 than instantiating `DescriptorParserImpl` directly.
  - The case of non-ASCII characters in statistics lines seems rather
 special, and doesn't seem worth adding an option that 99.9% of users won't
 care about.  We could generalize and provide an option to not fail
 descriptor parsing if a single line cannot be parsed, or something like
 that.
  - If non-ASCII characters in statistics line are spec-conformant, let's
 fix metrics-lib by default.  I don't recall whether that's the case.  But
 if it's a spec violation, let's treat this case as any other spec
 violation.
  - How much do we still care about this issue?  Is there a current use
 case that would justify putting in the effort?  If yes, okay.  But if not,
 let's just close this ticket.

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/19842#comment:1>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online
_______________________________________________
tor-bugs mailing list
tor-bugs@xxxxxxxxxxxxxxxxxxxx
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs