BSDTalk Interviews Joris Vink

I'd just like to say in comment to the fact of using OpenBSD as a workstation, It's true what he said.. OpenBSD makes a awsome and powerfull desktop system as well as a server system. I'd like to thank the developers for making a awsome OS.

I also frequently use OpenCVS for projects, though sadly i've been leaning alot of my code over to SVN as that seems to be more stable in a production enviroment, Though I use OpenCVS at home.

:) Have a good day everyone.

Where's the imprint please?
STOP those MP3 interviews just because of your own laziness hidden under
the trendy coat of iPodizm (read: do the real, written stuff again).

I really love to READ about events and happenings... i think the most of us do.

This goes up through 7:08 or so. I apologize for misspellings and transcribing errors. Apparently this is one of the many many things I'm bad at.

BSD Talk 050:
"Don't tell anyone I'm free... Don't tell anyone I'm free..."

WB: Hello and welcome to BSD talk number 50. It's Tuesday June 6, 2006. I do want to apologize about the audio levels on the last podcast. I should have devoted more time to getting the levels right. It sounded not so bad in a very quiet environment, but I understand that in less than ideal environments it was difficult to hear me or the interviewee was too loud. So I'll try and pay more attention to that. My apologies.

WB: Alright, well on to today's interview. Today on BSDTalk we are speaking with Joris Vink. He is an OpenBSD developer. So welcome to the podcast.

JV: Well I'm glad to be there.

WB: So I had a couple topics I want to ask you about, and I guess I'd like to start with the hackathon that happened recently for OpenBSD. Maybe you could let me know, uh, how that went for you.

JV: Well for me this was my second hackathon and from the both of them this was the coolest one for me because I got a lot of work done on OpenCVS. And it was really nice to catch up with all of the other developers and you know have fun and drink beer. And well it was fun in general basically because we got to [??]leach[/??] everyone there to suspend their lives for a week like Theo said to develop and stuff. And that was really neat and the mood that was just surrounding the hackathon was just awesome. And it was really really cool to be there for the second time.

WB: Were there more people there this time than last time?

JV: I really don't know. I think there were a little less people than last time. I think last year there were about 60 developers and this year maybe around 50, but I'm not sure though.

WB: I guess the hackathons are where a lot of work gets done, people focusing on specific tasks and as you mentioned you did a lot of work on OpenCVS and recently I saw an email from you on the OpenBSD list describing some of the success you've had with OpenCVS. Maybe you can describe that.

JV: Well the thing was 2 weeks before the hackathon basically happened I got so sick of hacking on the current OpenCVS code that I actually started talking to Theo to see about rewriting everything. And he kinda threw me some ideas on how to implement some stuff. And one of those ideas really caught on to me and I basically decided to partially rewrite OpenCVS. And you know like I said the two weeks before the hackathon I had so much stuff working that when we arrived at the hackathon basically what I did was commit everything. And everything went into really fast pace. You know we got commit working, we got update working, checkout status, all the basic commands that you use on a normal repository for example. There are working less than 3 weeks old basically. And the funny thing is we've actually got a lot more things done in like 3 weeks than OpenCVS has been able to do in 2 years. So that was kind of fun really.

WB: Prior to OpenCVS you actually had done some work on RCS, correct?

JV: OpenRCS, that's right. Although I have to say OpenRCS was actually forked from OpenCVS code. And OpenCVS development stagnated so we were talking to Theo and he thought it was a good idea to start OpenRCS first you know get more BSD licensed programs into our tree and stuff like that. And we had the base, we had the rcs [something] from OpenCVS and well since then evolved from there. And OpenRCS was actually done before OpenCVS. And it's pretty much a full replacement for the GNU RCS suite. There's some features that we still lack though. Like committing to branches which is kind of important but we still haven't gotten around to doing it actually, but we'll get there.

WB: Now is OpenCVS using OpenRCS to manage its on disk files? The diffs and everything like that or are they pretty much developed seperately?

JV: Well, originally like I mentioned, OpenRCS was a fork from OpenCVS. So we kinda were using the RCS parser in between, and then when we find out that wasn't really a good idea we completely forked OpenRCS. So there are some similarities in our code between OpenRCS and OpenCVS. But that's just in the rcs parser. Basically because I mean it would have been kind of stupid to rewrite an rcs parser solely for the purpose of OpenRCS.

WB: Now with OpenCVS you said that you've had some success with the checkouts and commits. Are you hosting the OpenCVS code within OpenCVS yet or are you still using GNU CVS for that?

JV: If I were using OpenCVS to commit to the OpenBSD repository Theo would probably kill me right now. But, um, I do manage all my own local repositories with OpenCVS already right now, so.

WB: That puts a lot of weight on your shoulders if the hope is eventually for OpenBSD to use OpenCVS you could end up being either a hero or somebody that people are trying to hunt down and get their hands on. Depending on how well it works.

JV: Yeah that's kinda true. I'm actually hoping it will be the first one and I'll be a hero because well I don't want to run away from Bob and Theo because they're really violent. I'm just kidding. Yeah it's real yeah. It's a really huge responsibility when we actually switch to OpenCVS because I mean if something screws up then the RCS files are bogus and everything messes up then well we're kind of screwed if you know what I mean.

WB: Although I guess unlike other source code repositories the OpenCVS is using text files ASCII so at least there's some hope of recovery.

JV: Yep, RCS is basically just a text file that describes all your versions like it has an administrative section that describes the HEAD revision if there's a default branch, it describes the description of the RCS file has branches and stuff like that. And basically all the deltas that describe all of the revisions are just plain text. And then the revisions themselves are just normal RCS diffs that you can apply against each other although sometimes the delta texts can be binary as well which is somethiing that we actually um well we're how can I say this OpenRCS right now handles binary files real good because Nilo actually committed something recently that was able to handle binary delta texts in the RCS file. OpenCVS right now cannot work on binary files which is something we really need to fix.

Out of interest, is there a regression test suite for Open{CVS,RCS} ?