wiki:WeeklyMinutes20100311

Notes from the BIND 10 call.

Attendees

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

Y2 Meeting

Will be at the CNNIC offices in Beijing, 2010-04-26 to 2010-04-30. I'm not sure if we need letters of invitation, but I think CNNIC can arrange that for us. Please check to see where the nearest place to get a visa is - if you are far away we may want to do this at San Francisco.

IETF 77

Shane will present BIND 10 at IEPG.

Kambe: Maybe Fujiwara-san will attend.

Jinmei: Fujiwara-san's name is on attendee list.

Feng: Someone from CNNIC will attend.

Shane:

Next Week's Meeting

USA has time change to "daylight saving time". Call time will remain rooted at UTC, so an hour later for people in the USA.

Y1 Release

Shane: Are we going to review the first year code? If so, we need to make that a priority.

Jinmei: I think we should do that. Maybe "lighter", but someone else should look at the code.

Shane: I agree. Maybe we should list the Subversion number in the review ticket?

Shane: Are there a bunch of tickets waiting for review now?

Jinmei: Jeremy originally made a list of code that should be reviewed.

Shane: Code not ready for review?

Jelte: Still working on unit test for one specific file. Already noticed a few things I would like to change.

Shane: Code ready for reviewed?

Jelte: Yes and ticketed, with Subversion numbers.


Shane: Code not ready for review?

Feng: All code is in ready for review.

Shane: Ticket for that already.

Feng: Can create ticket tomorrow.


Shane: Likun also?

Likun: XFRIN yes. XFROUT needs to have change from authsrv.


Shane: Boss process needs tests. No other code changes "good enough".


Shane: Code not ready for review?

Evan: Everything working on is already in review.

Shane: Everything ticketed?

Evan: Jinmei began thorough review, wrote lengthy e-mails about components. Then he realized we don't have time for review on that level, then I saw changes to the code, nearly all of which were good. Jinmei is reviewing the code by modifying it. Is that okay?

Shane: I think that's okay. The main point is to have two people look at the code.

Evan: There are some bugs left to fix, and I'm not sure where Jinmei is in his work. We seem fairly close to being able to merge if that is acceptable.

Jinmei: What we are doing is a short-term workaround, due to the limited time of work for the release. In the spirit of "at least 2 people have looked at the code" it should be okay. I think this is not actually a complete review - some parts are not reviewed. I think it is okay for year 1, but we need to perform another review cycle after that.

AP Shane to update backlog to note additional review cycle.

Shane: After the bug fixes, ask Jinmei to have a look, and then get merged into reviewed tree.

Evan: Okay.

Jeremy: Review trunk does not exist. I can start creating it - the only thing I knew to put into it was libexeption.


Shane: Code not ready for review?

Jinmei: Message part of DNS library, including EDNS0 is totally unreviewed. Feature wise it is almost complete, but I need to add more tests and it is not documented at all. I am wondering how much of documentation we need to provide for Y1 release. In any case, the code needs to be reviewed.

Jinmei: I am currently looking at authsrv code, and I am now fixing some errors. After that I think we need to get it reviewed by someone else.

Shane: Having running code is more important than documentation. At least for things like the message API.

Shane: Also have code waiting on review?

Jinmei: I don't think so. But I should at least provide more tests before the message API is ready for review.

Jinmei: Some of the RData code may need to be reviewed by someone else.


Shane: priority needs to be to getting code so it can be reviewed.


Shane: Jeremy, do you have work you want someone to review and look at as well.

Jeremy: Working on documentation, but probably doesn't need review. I'm going to try to get the Makefile to build the Python wrappers for libdns, but that doesn't need review either.


Shane: Larissa and Shane have been working on the BIND 10 communication, including press releases and presentations. Jelte is working on a "demo" that we can give for curious people.

A.O.B.

Jeremy: Really long release checklist, and I need to talk to Shane about what we are skipping or rushing. What we are not going to do.

Shane: We can talk after this call.


Jeremy: We're wanting to release some software that none of us even uses. We need to start using it today. Maybe this is on all of us.

Shane: I can commit to that for one of my servers.

Jinmei: For what purpose? Semi-production like serving personal zone?

Jeremy: Whatever you feel comfortable with.

Shane: I have a test domain I use for mail, I'll set it up for that.

Jinmei: I'd like to do that, but my server is primary authoritative for some of the zones. No xfrout so it's not easy for me to replace that. If the purpose is to use it in some practical environment, without the necessary features it is not realistic for me.

Jelte: I have the same problem. I am now thinking if we want to look at the features we need to implement first, we should look at what we can use for our own setup. I only have one IP address...

Shane: We could run something on bind10.isc.org.

Jeremy: Yeah that would be good.

Shane: I'll set it up on my own server, then have a look at the bind10 site.


Jinmei: I'd like to be sure that I didn't miss the AP.

Shane: I thought it was more important to look at the release first.

Jinmei: Maybe we should do it now.


Jeremy: A couple weeks from now when we start working on 2nd public release, we need to prioritize some things. The main thing I see right now is debugging. We have a lot of places where people are adding in debugging (like logging output or stderr output), so we need to define how we are going to do it and convert everything to that. I think that is a really critical priority.

Shane: I agree. Can you put that on the backlog page.


Jeremy: For our work for the next week. Do we have specific things defined that we must do. I can think of a few things, but do we have a list?

Shane: I guess the answer is "no". We should also talk about that...

Jeremy: If we have a list, even if it is short, it will make sure everyone focuses on those things. That way we don't get sidetracked.


Jelte: If you're going to run this as a real system somewhere, can we run this on port 53 without running everything as root? Is there some place that drops privileges.

Shane: No there is not, and we do need that.

Jeremy: As a workaround, you can use a packet filter to rewrite 53 to 5300.

Shane: I think we would be better waiting on this, rather than doing it in a way that may not be correct.

Jinmei: Should include this information in the release notes, and come up with a solution after that.

Jinmei: Maybe the most fragile part is the authsrv component. We can at least setuid() after allocating privileged port. Maybe easy to implement after year one, with a "-u" option. Can that be an option for the Year 1 release.

Shane: That is certainly possible and well understood. But less high priority than the other work we have.

Jinmei: I agree.


Jinmei: When I reviewed C++ code, I often found a common pattern of errors, which is exception unsafe code. One common pattern is to allocate new things by new and then do something that may throw an exception. So, I would like to propose that we not allocate things by new as long as we rely on exceptions. We should include a specific point on this in the code review.

Shane: I agree. Can you update the code review guidelines to say that?

AP Jinmei to do that.


Larissa: I'm working on various stuff around the deliverable. There are a few possible blog articles, if they could be done around the time of the release.

Shane: press & blogging and such doesn't all have to happen on the day of release.

AP from previous weeks

Jeremy to update blog article to include directory moves and publish it

Jeremy: Added it to the new web site. I don't know what happens next. I had it e-mail Shane and Larissa.

Shane & Larissa: We didn't get e-mail. Will report.

Shane to try to get logging for Jabber

Not yet, will ping.

Michael to write blog entry about SQLite3

Not here yet.

Last modified 8 years ago Last modified on Mar 12, 2010, 9:21:48 AM