BIND 10 #547: Update of statistics daemon for HTTP/XML reporting: implement

BIND 10 Development do-not-reply at isc.org
Mon Apr 11 09:48:25 UTC 2011


#547: Update of statistics daemon for HTTP/XML reporting: implement
-------------------------------------+-------------------------------------
                 Reporter:           |                Owner:  vorner
  naokikambe                         |               Status:  reviewing
                     Type:           |            Milestone:
  enhancement                        |  Sprint-20110419
                 Priority:  major    |           Resolution:
                Component:           |            Sensitive:  0
  statistics                         |  Add Hours to Ticket:  0
                 Keywords:           |          Total Hours:  0
Estimated Number of Hours:  40.0     |
                Billable?:  1        |
                Internal?:  0        |
-------------------------------------+-------------------------------------
Changes (by naokikambe):

 * owner:  naokikambe => vorner


Comment:

 Can I ask you more?
 Replying to [comment:10 vorner]:
 > No need to. I already went trough the code, so splitting it up now is no
 longer needed, it would only waste your time.
 OK. Thank you!
 > Well, with these things, there's no correct or incorrect, it's just how
 we feel about it. It surprised me, so I asked ‒ if it is on purpose, then
 it's OK.
 OK, we put it in the same directory, but we will split it if needed in
 future.
 > I mean, if I was a user, I would care about if it runs or not and how to
 ask it for my web page with statistics. Where or how it gets the data,
 it's none of my problems. So, why would any other user want to care what
 is the internal format that goes from one place in the BIND10 system to
 another? I fear that the man page is too long and confusing users about
 too much detail.
 >
 > This information is, of course, valuable as developer documentation. But
 man page is for the end user.
 I understood that so detail isn't needed for the most end users. I'll
 change nothing.
 > But that's not true. No user edits the .spec file, it's configured
 trough bind control.
 Further, for avoiding such end-users' misleading about command-line
 options, do we need to append the following lines into the CONFIGURATION
 AND COMMANDS section in the manpage?
 {{{
 You can change the default setting through bindctl. You cannot change it
 with command-line
 options. Please see also bindctl(1) for how to change.
 }}}
 > Well, the limitation probably will not be hit, but there's no technical
 reason for this exact number. The code and hardware can surely handle much
 more. So it only makes the code little bit more complicated and there's no
 benefit for the user (it can, in theory, make him unhappy actually), so I
 don't see much reason to have the limitation there.
 OK. I will remove that.
 > Just `raise`, without any parameter. It will just re-raise the original,
 unmodified, exception.
 That's right. I'll change it like your suggestion.
 > If you are aware of it and think it's not so big problem to complicate
 the code, then OK. But it would deserve a FIXME comment in the code at
 last, to note that the issue is known.
 OK. I will append a FIXME comment around the while loop in the code.
 > They don't really need to be reusable, but again, a comment with warning
 about it might prevent someone from trying to reuse it and failing.
 I will write some notes on these mock-up codes.

 Regards,

-- 
Ticket URL: <http://bind10.isc.org/ticket/547#comment:11>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development


More information about the bind10-tickets mailing list