wiki:WeeklyMinutes20100107

Notes from the BIND 10 call

Attendees

Jeremy
Likun
Feng
Shane
Jelte
Larissa
Michael
Kambe-san
Shawn
Fujiwara-san
Jinmei
Evan

Agenda

  • Roll call
  • AP from last week:
    • Shane to try to make a graph of timeline
    • Shane to give commit access to Francis
    • Jelte to put up a link to description about merge steps on Wiki
    • Shane to update Wiki with deadlines in review process
    • Jelte to write what he did when reviewing Jinmei's code as checklist in Wiki
  • Sponsorship status Y2
  • Face-to-face last confirmation
  • Status Check
    • DNS Message API
    • msgq
    • Configuration
    • Command & Control
    • Statistics
    • BoB
  • Data Source
  • AOB

Action Points

See above.

  1. In progress
  2. In progress [ update: now done ]
  3. Done
  4. Done
  5. Done

Sponsorship status Y2

Shane updating statement of work for Y2, to be used to sign up current sponsors again. One new sponsor (although unsigned). Hopefully Y2 SoW will convince more sponsors.

Face-to-face last confirmation

2010-01-25 to 2010-01-29 ... book your flights and pack your bags!

Status Check

DNS Message API

Jinmei: Name class committed. Other classes... work ongoing... in middle of development. Possibly reviewable version in a week or so.

Jeremy: parkinglot... convert to code in trunk?

Jinmei: unknown... depends on parkinglot

Shane: don't think we will convert parking lot to production version

Jelte: think we should do that...

Shane: how much code are we talking about here?

Jelte: not much!

Jinmei: makes sense merge to merge part of parking lot to trunk, but not related to DNS Message API

Jelte: already thinking about doing it because doing data source stuff in parking lot! do we need all of the changes in DNS message API?

Jinmei: need all of it.

Jelte: once whole message API is done, not a lot of trouble to merge in to parking lot branch

Jinmei: maybe we can do it in the face-to-face meeting

Jelte: or maybe we can have a working version in trunk by then

Jinmei: we should do coverage test for name API now that it is in trunk AP Jeremy to make coverage test automatic

msgq

done but needs review

Michael will create ticket for this

Shane: will pick reviewer

Jinmei: maybe data source in more important?

Feng: we can review code

Shane: done!

Jeremy: separate into review for daemon and API?

Michael: it's a bite-sized chunk, so one review

Configuration

Jelte: in parking lot branch, a version that works, but it is far from ideal... but it works! Let it hang there for a while.

Jelte: did more work on specific data element stuff - also used by message API

Jelte: nearly 100% code test coverage for C++ part, documentation, also ready for review

Jelte: for bigger, general configuration "protocol"... no progress

Shane: go ahead and create a ticket for that and fire that off

Shane: Michael probably best to review?

Michael: Maybe better to have someone who has nothing to do with it?

Jelte: maybe... but almost completely new

Command & Control

Likun: sorry... past 3 weeks have been on other stuff... plus missed all meetings since meeting before Christmas! And couldn't log in with Jabber.

Likun: now subscribed in GMail.com, so this should work

Likun: will try to finish code before face-to-face meeting

Statistics

Shane: Jeremy has been writing this... URL into jabber, not sure if it will necessary to have a statistics gathering daemon

Kambe: making daemon of SNMP stuff

Shane: can you commit to experiments branch?

Kambe: not yet, I am researching

Shane: okay, looking forward to it!

Fujiwara: need to make time for BIND 10 statistics. Will like to make time before face-to-face meeting

Shane: sounds very good. if you want to talk to someone about it, Jeremy has spent the most time thinking about it, otherwise talk to me

Fujiwara: will read Jeremy's ideas

BoB

Nothing done

Data Source

Michael: most things are on the list, need to put together feedback from Jinmei and Shane. Hopefully will have something about that today. Like 3-layer approach ("pure abstract", "common functionality", "concrete").

Shane: maybe set up a separate call to talk about the issues that came up on the list?

Michael: want to be very simple, agile thing

Shane: so not IETF-style (discuss, discuss, discuss), but just code it and see what happens?

Michael: running code!

Shane: looking forward to the 1-second startup!

Jelte: I have been working on implementing something like the straw-man proposal in the parking lot. Nothing concrete yet, but a call about this soon might be very useful.

Michael: Tuesday next week?

Jelte: If the conversation on the list continues, then we can wait until Tuesday, otherwise sooner. Michael: okay, Tuesday

AOB

Jeremy: was Jelte working on blog article about coverage testing?

Jelte: not working on it, was suggestion... could write up something about it... but have another idea that I want to write first

Jeremy: will do code covering blog article in 2 weeks maybe... then I'll have an automated thing that people can visit


Jeremy: since we have the first thing put into the trunk, should we ask people to start looking at it, since it is in trunk

Jelte: not much to look at


Jeremy: we have a few command-line tools. Has anyone started working on manual pages for them. If not, I'll put together man pages based on reading the source code and running the tool

Shane: yes please!

Jeremy: doing it from perspective of the code containing user documentation. Embedding the documentation in the code and generating man pages that way.

Jelte: Does that work?!?

Jeremy: I'll look into it and figure it out.

AP Jeremy to start this with bind-control tool

Michael: document describing wire format from control channel is no longer in sync with the code

Jeremy: okay will work on that too.


Larissa: BIND 10 has any plans to addressing DNS6to4?

Shane: no

Michael: what do they want done?


Jinmei: deleted some old branches... is this the expected way to clean up a completed branch? I did this because it will be noisy if we have all the branches otherwise.

Jelete: in favor of deleting branches we are not using any more

Jinmei: drawback is that we cannot quickly review it via Trac subversion plugin

Michael: once they are no longer touched and maintained, it's old code. So make sure we don't lose anything, but delete otherwise. Would rather notsee 20 branches for historical agreement.

Last modified 8 years ago Last modified on Jan 12, 2010, 4:10:34 PM