 I thought I'd try to get a sense of who's here just as an interested observer or maybe interested in helping with DI and who's a DI developer as everybody knows, I'm not very good with faces and names. So all the DI developers or people who've been involved in some aspects of translating or anything like that, please raise your hands. That's it. That's it. It's just my time. Current or future? Past. Current. Past. Past. How about the razor? You have it in the head. Oh, I'm sorry. I just asked if everybody who was involved in working on DI could raise their hand. So yeah. No. One bitch. Yeah. I get the sense that we don't have as many DI people as we do these metacons. We don't have that many DI people at all. Yes. Translating into DI? So I guess that I also wanted to talk about things like the next release, whatever else we want to bring up, I would really have a big role in this. Who has some actual subject they would like to bring up? We have two different devices, I understand that we have released the device for the test release 3 or whatever came out in a while. So that's for the actual to be released track. Do we have something else where we can throw in experimental changes where we can test out things and maybe back them out if they don't work out? It's probably fine to do that in unstable pretty much. We do split unstable testing so we can still have it working in smaller testing. Although we do mostly do local builds before we commit to standard distributed software engineering. It should be likely that it will be ready for the next release before it's going to start in stable. Do you guys are using SVM, I believe? Well, I've kind of failed to do real planning on that up till now. Basically because the AMD64 archiving integration hasn't gone as fast as with kind of, well, expected, I hope, I don't know, at least it was slower than I expected. Basically because everything is being rebuilt and slowly trickles down into testing and it has to be complete in testing in order for DI to be a feasible release, including AMD64 from the official archive. So that's basically the biggest showstopper. A related issue is that the mirror master list has not been updated yet. No, he's not. Yes, he's working on it today so that's going to happen. The mirror master list is of course essential because after the installer is released the mirror list is basically frozen until the next release. And we do want an up-to-date mirror list with the split that's happened, with mirrors that have dropped architectures with the AMD64 addition to a lot of mirrors. So we want to be able to reflect that. There have also been a lot of progress on a number of issues like the integration of the graphical installer, which we are working on here at that conf, actually. The next day-to-day CD build will be the first one that should have a working graphical installer. We have great progress on apartment crypto, which is the option to install encrypted partition during installation, and a number of other things. So I have not felt real hard pressure to forger this until now. But we should start planning it now. With one problem that I will be taking on a three-week holiday after that conf, so... So that's, well, when I planned it I had no idea that we wouldn't have released the beta 3 by now. After beta 3, I think we should try to slow down new development and start thinking of IC releases. I mean, we're going to be used to... You have some kind of indicative timeline for when you want to stop accepting new stuff. It's part of the official edge release timeline. If you look that up, you will find... Yeah, but there's also the date of, like, when the final Debian installer release should happen for the actualist to be able to have an on-time processor. Unless my memory is really neat. All this was 15 to the end of July. I may have wrong. That's for the pastries, but the final version of the installer is planned for later. Okay. Well, we should try to get your work included at the latest in our sheets to, I guess... I'll do. Yeah. In June? No, that would be like... I would guess, I would guess, August, September, maybe early September. I think we'll get it in about June. Well, my personal time table is to get the AC Linux changes to you guys, stapled by end of June. That would be perfect. Because if any of you guys want to put out on the AC Linux, it would be greatly accepted. So that is basically, I think, what's going to happen for the next beta. And the release after that. For the RC release, we will of course also start pressing a bit harder for translation updates and updates of the installation guide. So that's going to start happening basically halfway through the summer. That means... That's being prepared. I have pending pets that I need to test. To what? To... The main issue really is persisted defile names for our disks. It's at the top, I think. It should be. I've been wanting to do that for a long time, but for some reason or another, I haven't gotten around to it. We need to set up a meeting with Walden. And maybe we could even do this this week. Try to find the time slot where we can set the bias. Sorry? Any more about the status? Currently waiting for X, basically. X. Which is holding up a whole lot of packages. Oh, holding up. And we also... We probably cannot do deny for N64 before we release beta 3. Because too many pets view that are missing. Oh, I see what you mean. Because everything trickles in from unstable. There's not more view than basic because they have to be released. That's the best way to submit bugs for hardware specific things. Like a certain Dell laptop has a SATA drive. And with the Sarge installer, it doesn't load the red ATA for SATA. It treats it as an IDE. It's so slow. So you try the 2.6 with Sarge. I realize this is actually... I was talking about bugs and reporting bugs. So you try the 2.6. And the driver says it works good. But it misses the CD-ROM then. Well, basically, if you want to make bugs for SATA-related hardware. For Sarge, don't bother. No, but the same issue is with the Edge installer. With the Edge installer, it should work. This is the CD-ROM. Yeah, but it should work if you look at the wiki page on SATA tapping support. And if you follow the workaround, that's used to dare it should work. To answer the actual question, the best way to call the bug is to go ahead and submit a full installation report with all the details. And make sure that it's good enough for somebody to say it. The installation report is in the slash root. Is that weird? It's not there. You can type the report bug installation. Basically, the missing CD-ROM is a known issue. The kernel team has made the decision to wait for activation of SATA taping support until the upstream kernel maintainers feel it's ready. A lot of other distributions have decided to enable it earlier, which is their choice. We are dependent on the kernel team for this. It's not our decision. It's not the installer team's decision. As of 2.6.16, in the live ATA dot... Oh, God. There's a variable. Enable a taping. I think as of 2.6.16, that was still set to zero. Yes. I have written an email to Jeff Garzen. He said, pressure him to change it to a one. He had promised to do it by 2.6.16. And now he's promising to do it by 2.6.17. I'm wondering if we could help if we pressure him to change that. I've been testing at least in our hardware, and I've had no trouble. And I asked him, what do you need done for you to change that from zero to a one? He said, well, don't worry. In fact, he says, don't worry. And he will do it by 2.6.17. So I'm wondering if we could help if we were to try to pressure him. Or just go ahead and change it. Ignore the fact that he didn't change it and change ourselves. I can't hear you. Maybe somebody from the kernel team wants to speak up on this. I can't speak for everybody in the team, but my opinion should all be done. It's done upstream first. So yes, if you want to get it done, talk to Jeff Carson because he's the way to go. He's going to do it 2.6.17. I don't know what more we can do about it. Okay, but we don't want to change ourselves. No, not unless he does. And what you want to do is we're considering right now to show 2.6.16 an edge, which would mean the current way plans are and they could change. It wouldn't be on by default. But 2.6.16, Adrian Bunk is supposedly going to be maintaining a branch 2.6.16 that has security releases and backward features. Okay. So Jeff Garzik getting that change into Adrian Bunk's tree is the best pattern. Okay. All right. Okay. Maybe if somebody, if you feel like it, send an email to jeffgarzikitsud.com and tell them we need to change that. Yeah. Let me go ahead. Let me tell you that from a debit installer team point of view, we would love to see it enabled because it's going to save us loads of buck reports for edge. So if that helps the CRL team to decide if it helps you to pressurize. But on the other hand, if we get the same amount of buck reports because it's not ready yet, then, well, leave it off. Thank you. That's giving me a lot. Well, it's enabled to 2.6.17 and never in the 16 branch is possible to backport it because theoretically, it's stable there. We can backport those changes before we freeze. Okay. Thank you. Discuss the number of base of the DI team last year when you first started worrying about this question. Worrying about what? Worrying about the size of the DI team. Yeah. You can worry about... People drifting away. Yeah. I think you began to perceive worrying about one year and a half ago. Yeah. Worrying about a few months before that it will be the priority. And it came obvious that the number of people involved in the DI was decreasing. And actually, I think we can say that it's about less than 10 people probably actively collaborating with the DI and worrying more like 5 or 6. I think it's a lot of people. Yeah. I don't count translator. No. Not because it doesn't count. But I'm saying this. For example, Max is working on a permacrypto. Yes. And that's all. And I'm very proud of him as well. So that's two people really active but only on... Yeah. And also we're talking about several is only working on drug stuff, for example. What I think we are missing are people who know who have a good picture of the whole thing. Actually, I think that we only have both of you and for a week only. Yes. This is little bit... Not enough. And there's a few areas that are basically not being maintained at all. Basically permanent and that's safety. Yeah. Which are worrisome because they are so basic to a good installation. Yeah. We hear a lot of how bad these spotmen are. Some people don't really like it but we don't see that much how it's doing when we're actually. So it's all a little worse to a certain degree. There's a lot less incentive to try that. Yes. On it. And I think that you could probably go back and find the same thing happened. There was an initial search probably much smaller. It's still... And then it just trickled off. And then it just fell off basically and it was no more who bought it in minutes. Which we don't want to have to. Yeah. For most areas there's still a lot of features that can be added and real improvements for usability that can be made that could be very interesting to work on. So after what's now two years for me I find it's still very challenging to work on the ISO. And I would hope that more people will trickle back in. Yeah. Of course. Yeah. What's more of a comment. I tried helping with the Debian installer team and I found the overhead of trying to work with the Debian installer team took more time than it had available. You knew we were just speaking up today. Yes. Well, at that point in time which was well before the surge release things were being updated faster than I could pass them and so by the time I developed a pass that worked on Spark the code behind it had changed totally. The Debian boot mailing list has so many messages that trying to keep up with that is takes, you know, if I was actively trying to do that that would take a lot of time and most of the messages aren't of particular interest to me and so somebody trying to help with a small portion of the Debian installer winds up getting locked out by the overhead I would like to be able to find out say a week or two before the beta release is scheduled so I can try the installer and see if it works on Spark. All of that doesn't scale very well because if we have to deal with 13 how many architectures two weeks before the release not working then we don't have a release. Well, and I can I don't really have time to do a test every week. Well, since the charge release there has been one really major change, only one really major change and that is the switch from 2612 to 2614 which involved the removal of DevFS and the complete integration of UDF. Yes, which doesn't really affect architectures the same stuff that's talking and DI beta 2 basically didn't couldn't find my hard drive at all it did not find Spark. But that was a result of the switch to UDF because that meant a drop of this cover which previously took care of that and it was on my to-do list and I've made sure that the installer itself recognized it but there was still the open issue of doing it after the reboot as well which has not been fixed I understand it's been fixed but identifying issues like that is very important I have only an ultra 10 so I cannot test all the SBUS stuff so getting semi-regular installation reports about the status and maybe the same issue should I deal with monthly or Yes, that would be correct Absolutely That's fair Two questions from the School of Linux that I overheard is that the automatic partitioner we have our own and it's I believe that Petr told me that the code was not so it would be nice to change it but something has to be done and I don't know how that is integrated The other question is how to make the installer or describe how that people can make more standard distribution using re-use the installer as is you know changing anything about the installer to plug it in because we have students asking for a project that could involve making some special case of DBM on some special case equipment So these two questions I want to address Because taking the first one has been too long so I may be out of date but as far as I know the issue is that it works very similarly to part-man in automatic mode but it doesn't can also be used in part-man and so on for some of the next ones but we don't necessarily calculate partition sizes using the exact same algorithm the things that are tuned specifically for school so you know how they work if you switch over using part-man it's going to change you might end up with different ratios of sizes, different provisions and that kind of thing so I think it's better to switch over at some point than just flipping the switch and switching over I will publish which Petro said that so you see as this pressuring helpful event around because we haven't enough people there ourselves but if you want to get maybe some changes in to Petro support school Linux I guess you should not wait too long to start working this to start the transition The thing is that I have to start arguing on the school list I have some friends there with me I will all the questions Custom Debbie distribution increased seating used in custom UDFs internals which probably it doesn't cover it doesn't cover building current UDFs so all the issues if you want to know how to do it just mail chemistry you need to write something they would also be talking also about maybe pulling some of that back back and forth so that people can build their own UDFs easier than having somebody else build a full kernel image that dredging out the parts that you want UDF maybe think we The main thing is that because of the changes from this cover to UDF the backports were very major so you really had to jump into the installer and backport a lot of the changes we've been doing after search to get search installing something later than 2612 right? Many points are taken from H in installer but maybe I miss some points so I received some reports I can't install my UI image but I have no time to look at it I guess also that's the most issue if it's hard to try to split the list or try to have an announcement and maybe go forward with splitting the box out of the way The same problem was glass, a couple of years ago I tried to get a lot of them and was scrolling the 1 million messages of the list and also I failed to tell it to understand how did I work trying to do anything so I was explaining to you again recently and I think it's clearer now than it was a couple of years ago My problem at the moment can we do installs to empty devices yet? What do we do? Part of that needs to be done right? Assume we have to do some work on part and we do a lot of things and the other problem is that the board doesn't have the kernel in the mainline which I think makes a big difference as long as you build a model you can play from experience So there is a mechanism to stuff my kernel in just kind of a main kernel but it's still built in the items that will work and there's a way of doing that as well and this is something that would be to take the AES image it's actually called AES and it's just missing any protocol and all you should be building all of it I wouldn't have much time to devote to this but if there wasn't a next list saying there's a review list please give it a go that would help we have a lot of things that we would thank you for thank you I think it would also be easier to do more regular updates on things that have changed do a quick announce to that list if it's about 20 minutes we should be able to get 40s to subscribe to it instead of being subscribed to that input but I think we want to do that on the list that we know probably first day late afternoon I assume you'll teach me the I's so I'm going to go here I'm developing a guy that's actually interested in helping working on something you're just interested in helping out not just working on it I'm going to see if I can support it I only have 8 to 6 or whatever I can help with I will do that I can't do exotic architecture I would like to test it I want to see how it works you know what I mean I love it I'm trying to do the same I didn't want to do pure is there anyone else we have lots of boards that we produce really we support it on our board but nearly all of that I don't have any specific plans of things that I'm interested in working on but occasionally I'll try out the installer and there'll be some glitch and I have trouble finding where to try to track it down just because I don't understand the architecture very well so I'm interested in learning just to fix smaller bugs well that's just mentioning how I start the problem is that you get to act many you have to go I think so you should say well I've been warning other people don't start with translations because my dad don't release manager oh thanks for all that so far but debugging DI is basically quite easy for most components are almost completely scripted so you can basically just set my necks and see what happens it will all show up in the sixth wall there's a few components that are written in C so that's a bit more involved but most you can mostly narrow it down fairly even part 1 I think part 1 scared me away it originally was that one of the problems that I was trying to fix first time I ever checked out the repository it was something in part 1 yeah the whole thing is like this I don't know I may have to do talk about apartment based structure and stuff like that sometime in the future maybe we have to foster them or something so for those of us who are currently in Boston no but they will tell you proceedings and slides so I no no yeah using the the installed base this is something I have had on my mind myself using Debian installer without the installer part like ever since the NSLU2 installer came out it was like hey it already has SSH just put somebody there and took all the installer stuff out isn't the boot to live CD basically Debian installer without the installer it starts out with parts of Debian installer it does run major portions of Debian installer you know I think it was different parts but for all of us it was just selection life CD there is a live CD wiki project going on I don't think they are directly involved in using the Debian installer for it I'm doing this and the problem is that Ubuntu is casper and casper is not that perform it's not that fast as we need it and we need to rewrite that and be working we don't want a CD but that should have support on installing Debian off the CD on good to that I personally have another aim to do live CD or live system the current Debian installer is great and I personally don't have the aim to replace it because why do you trust the live system and Debian install you can just launch Debian install it's faster but it's going to be awesome I would love to have something based on Debian installer rather than topics that I carry around so I could you know, infect the machines as I come across them and still have something that I can use to rescue my box from anything to screw up LVM for example well, there is always a risk switch just for LVM that's the problem and even topics doesn't have LVM to do that I'm sorry to interrupt but I'm just checking out the stream and it's barely audible it would be very very helpful if people here could be talking like three or four times rather it would make that audible to the people outside the problem is when I put the level up you hear the machines exactly the highest level from the fans and the machines usually blocks the speakers out can you pass around the light can you have a lighter? Joey, speak up you also you are proud of you yes, I would like to, with respect to the life against the installer problem share the experience in it first, we had like install CDs after that we started making live CDs with the packages that went to installer with the lightest release of Gento it was like it came with a really interesting option the full CD was basically normal live CD in space for packages just like two mechs for the installer and there was an option to like reverse install from live CD to a package and then install that package so that could be maybe interesting for the devian installer for live CD install I don't think so it's waste of time, waste of performance it will create a decent live CD which you use only for live CDs to run a system without installing it or you create an install CD you install from it but otherwise you would be forced to have a larger medium and a DVD to just to just install the most packages you have a DVD package and that's in my point of view it's just interesting that's actually the point of like you don't need a larger medium because you take everything from live CD the live CD can then you would have to then you would be forced to install all these packages and you were later to be able to install all these very live CD and there are the data as to the one is there are a lot of packages which cannot be installed at the same time and the second one is as I said then you must have all those packages installed on this very live CD and that would blow it up you just have to have the base packages the essential packages on live CD and the rest you can go from a repository but just like sharing my experience I think basically as the DI team we are not that interested in doing live CDs because basically we have our hands for doing the install on proper so if that is going to happen we need really need extra capacity the only reason I asked is because I mean you might be worth at least putting a new version of the graphic on smaller live CD or something like that so you can still install the CD if you really want to if you have a network connection I just wanted to make sure that we were able to do that also the problem with the packages that you have in live CD is that in many cases it's a live CD that is customized so you have tripled down all the translations that are not related with specific languages so you can only use it for those languages and repackaging those packages doesn't get to you the original package itself it gets you something totally different but that's one of the key points why I wanted to have an official live CD for Damian because we do not alter any package all packages are on the live CD install are unaltered Damian packages are sent at the moment so this point is not really valid but I would suggest to say that it could be technically made to use that live CD which can have an installer on it but that should never be as a proceeding installation as a proceeding for the current Damian installations because they are much more valid I tried to find the current do we have it here on the local network because I can find some daily builds from 5th of May should we ask for it to be mirrored here would it help? it's a bit hard to do because you could ask for the beta 2 to be mirrored probably that would make sense for the first two CDs and the first DVD or maybe only the CDs but it's going to take up a huge amount of space and better ways to get them if you want to get them for all architectures and if you are not going to do for all architectures how do you decide again it's about you could do IT86, AMD and PowerPC I guess could it make sense to have the net install images that would be small yeah if we want to do testing here is that what makes most sense to install from the local mirrored I've been doing the business card image which is even smaller it's self-contained I've done some business cards which works in trying to make it yourself I just noticed that it took me an hour to download that 20 megs of the net-purposing that fast all the time trying to do it early in the morning or late in the evening in the user proxy I think it has a couple of lines together instead which is why it makes sense another thing I wanted to ask do you have I have understood that Joey it makes some automatic tests that you are I do which I don't have a lot of time to maintain the machines and it's a lot of times to keep six or seven different architectures running and testing that doesn't that's all it is it works if somebody's interested in working on a setup for it I've also successfully this is a test setup I've actually been I was told where I actually set it up in Brazil to do testing which is mostly useful to if anything major in this you can't test everything I was actually asking explicitly of the automatic instance but more generally about the coverage somehow possible to for example see configurations have been tested the longest time I go or something like that you can go look at installation reports and see if people have tested say a RAID or an LAM installed lately but there's just so many permutations exactly what everybody installed and stuff like that so there's no real way to do it there's probably too many different combinations to use it still is our main source of information about issues the installation reports it's very important that we keep receiving them even if we don't don't always answer them we still read them all and we kind of see patterns developing two people report the same breakage that's obviously not some mistake they made or a broken image or a hardware major but it can be nice for the testers to have more coverage than for other people to see these patterns I don't know should it be some kind of automated metadata gathering so basically you can ask for people to test something if you know it I think what you're asking for is some sort of machine parsable installation report there's so much information random information that I don't know if that would be feasible and often the real problem is this kind of hidden in a fake description of what people are seeing and you have to have some experience with reading installation reports to even identify what the issue is it's it's one of the hardest parts of working on the installation actually looking at the installation reports it looks to me as though it's about a structure as it can be for the information that you're providing it just doesn't look to me like there's too much room for making it more structured so that it could be more parsable one of the ideas I've had on that is maybe you could send a program and for example based on the architecture you choose you get only a limit list of the installation methods that are supported so you could like to do a more targeted installation reports and also for instance you could get the stages in the installation in the right order which they aren't for some installation methods currently so in that way it could be improved and also if you look at the top-up presentation that was given today it would be very nice to integrate with something like that in the installation report that basically means that we would need an installation report tool instead of just using report back we would need a dedicated tool probably wouldn't it possible and kind of easy to at least for network based installation for the installer itself report bugs with the current state and logs without having to repeat the bad report on the webform, email or whatever you're saying it actually has an error yeah I think like if there was an error due installation or whatever or if there was an error but the problem was with the start it can keep logs on the same place in its priority states yeah I mean right now if there's an error you can break out and start the web server's building install or download the logs yeah it definitely would be possible to make it send stuff to the BTS so I think we'd probably just get overwhelmed in reports maybe in the BTS in the right place to send it to some kind of network tracking system network tracking you'd have to cut something and let you look at poorly patterns so you have individual reports so we need to add some some filtering right so I mean you can write what in shell actually I think that the installer does have an SAP client written in shell I just don't think we're not using it right now because it sucks so you can write it better or it doesn't do really nice one thing I noticed I found I had a half brick at the store you know I got started but then got stuck so you can fall out into the shell but I couldn't find anywhere any docs which told me what rooms I was allowed to type and they brought the DI equivalent to go and get me a package it is because there was something missing so I wouldn't know I was quite sure I could fix this manually because it was incredibly hard to work out how and I don't know if that document existed and I couldn't find it but it's simply not written in it existed completely well so I'm thinking about the I think it would be good to have that document for preceding to because people who are preceding need to know how do I tell how do I tell the installer to install the package for example it's always not going to get longer into the shell to fix it it needs to be able to find a way so that would be an improvement it's going to have to be painless to make I think I'll offer a few pages on that and try to you know through to get through at least the most important tools that you have then we can maybe have the installer to bring you out when you start the shell or something like that it's also partly covered in my workshop oh is it? yep and it's also partly documented in preceding we do mention how to install and the problem is that they don't always work from the shell we need to document what works from the shell I'm answering you about the DSP right now we are using version 2 because of the size I checked it, it's 40 kW version 2 140 kW version 3 and the main problem are the floppies yes I thought about it and discussed it with Tempura who is an h3 3 maintainer would it be possible that the installation uses version 2 but it installs version 3 I mean here's the problem on the h3 clients they're in situations where they can't install version 3 or they can't install version 2 or version 3 doesn't work it's really good to have the h3 sustain tool for both things it's just the same as it tries to use the same problem as it installs so you don't have to deal with all of the issues but then we will stay with about what we obviously can't say with DSP 2 for a review I'm not really sure what the floppies will look like for 2.6 as far as how much free space they'll be and do you know not at all but I know it was a problem because it couldn't fit DSP on it but that was on the good floppie right, it's a real floppie that's the issue and it will be possible if that's what it takes there was probably after you prompted him Andrew started a new threaten that he would about DSP client through another option around this yeah, the nano the problem the thing is that we are now also having the option that you can proceed using the dhcp server so you define which file to download in your dhcp server and that requires options that are not available in the lightweight dhcp clients that are out there so they would need to be extended to offer that to be useable for di in the current situation it's sort of like a way to see how what happens with 2.6 floppies and just see how the space situation looks and it's probably dropping to 4 floppies so you don't have to worry about that and then we can look at the scene you know if we want to do what to use this dhcp3 in the install we will need to get the unit check out the slides again periodically check it to make sure it still wasn't fit I don't know why I still figured out what the other 100k does but if you can't even pile them out that's why that's what I understand you really need access just to get all these things if you're trying to do public things you don't go down with this pile you need some substance yeah you're always out of time I'm sure I guess we have all nine if you want there's one thing I think the handcuffs and unfortunately it's going to be too late because obviously from what I understand dhcpH it will be based on the 2617 1626 everything today southbridge in it which I think the saddest south bridges that I'm concerned about I am currently at a taffy boy I think I just looked it up and I think it was already changed which one did you look at get so it should be changed to 17 okay alright there are some networked the graphic drivers are going to be tricky because there are going to be some drivers the best approach for me is for the audio drivers the drivers and the DRM 1000 lectures for those before sometime in October they've been committed later very easy for us unfortunately I have nothing now hopefully across my fingers either a daily build or a point you guys have at and it's a trivial backport and it's even 1000 drivers if you would file a bug you can catch file or a pointer alright the thing I'm concerned about later this week he he and Keith I think Keith helped in with the kernel but it is loaded into the actual device so imagine it being like a bunch of programming for a field program like a FPGA or a programmable logic like that kind of firmware that's part of each one of them that's open source and that can go into kernel it's just that the binary of resistance including that in the main line of the distribution and I guess I'm going to be in the extras or the actually I think that people are out of the distribution of these you can you can be redistributed you just have to redistribute with the license of the document that has to be redistributed something more in combination of input products therefore it's not distributable the other thing I wanted to tell you I'm going to be up front and I'd like to be honest and I'm not going to say it through it that's the IPW 945 that is the network driver for the Centrino Duo laptops which is the 945 GM 945 mobile platform that as I said it has a binary firmware it also has a binary daemon the front I knew this guy binary daemon that runs as root on the machine what it is is called a regulatory enforcement daemon it is mandated by the FCC and or your country respective regulatory body that's responsible for radio emission regulations and that somehow the way I understand it where you are you can transmit here but I'm sorry you cannot transmit here it is a binary I know it's going to be a quality contention or it's the infant's idea what it is I will take back into my drip report the fact that we cannot get those drivers into making what's happening now, sorry I see no problems with those I think that's all I have do you remember anything else I mentioned upstream or patch what's the deadline for me to get that in check the release schedule there's a freeze date for all the convoys I'm not sure it is I'll top my head before August I'll give you up to three I completely imagine with the 266 idea that we have now there may be an option that we will backport stuff like that even into edge after it's been released so I just stayed in contact with you by email no, the current team current team yeah, I'm on that I have a question actually for the current team representative has there been any decision by the team about binding laws in the current sources I don't know what the latest was, I mean, Dillinger did a kernel non-free package a while back and all of these are still maintaining that I saw Yuri do some stuff about how to change which may be but I'm not sure what the source was but the decision that was taken was that the main the kernel images in main they won't contain binding blocks I'll let you know for sure they will not contain non-redistributable blogs that we can't anyway but what about binding blocks for which we have no source I can't answer, I don't know this binary is redistributed yes if we refining the sources for the binding block is also a point of contention there are some interpretations where we say that if we software that might be required also the license for the binary blob would have to meet the Debbie and Free Software guidelines Debbie she should be able to modify it that's the question modify it and use it on other hardware etc one person at a time I have bad hearing what did you say I said that the license should not require to acknowledge it so you shouldn't have you shouldn't have to click and okay but in software or whatever but if I say that if you use the firmware in an hour you are acknowledging the license so at the end you don't have to no I have to adjust you on that in Norway it's not allowed to enforce a license whatever you have to be actively agreeing on a contract if it's valid or else it's not valid that's the law so if other countries accept and this is interesting that Intel that is an American company that has a license that is not validated by the local law and I have this from the government in Norway it's not valid if it's breaking the law in your country and you just can forget it so it will be ignored so to tell another story and I as a not law person I have asked the lawyers about this so it's pretty devious what Intel are doing so you should just I have to say this you should just file a claim to the consumer authority so you can change your license probably in Norway because you're breaking the law I will be complaining in my trip report I will be complaining in my trip report do that and you are agreeing with us personally I have but as they represent Intel I would just take our yeah I understand so it sounds to me like those drivers that would have to be submitted to the contrib section and is that a separate submission process or do I still submit it to the colonel are you talking about the wireless drivers yes that require the blocks yeah I think that in 221 on there those things are already in contrib okay but any changes to them if they're a separate email list whether a group is maintaining those packages he's the one who gets the upstream sources and packages them up and puts them in contrib so your bug tracking system is typically your approach for asking for updates in general so whoever maintains those packages is the person who would get that by reporting they have not maintained their own fact but as Intel provides newer snapshots of the drivers which to make it in mainline some months later I'm continuing to adopt them for testing if you want to keep track of certain packages there is a mechanism for the package tracking system where you can subscribe to your email and every time a new version an update of that package happens you'll get an email okay that's probably the easiest way of keeping track of any update in the unit I think we will pre-ball over time and pre-ball a lot of such things so maybe we should just close yeah I guess one thing that I did not mention is that DI doesn't have any support right now for installing non-free drivers so people have to do it all on their own after the installation and if anybody wants that to work better that's a big chunk of work that's going to be done still I can't hear you I'm sorry, I said that the Deputy Installer doesn't support right now installing non-free stuff that's part of the installation so if it's in Deputy and non-free you have to go get it yourself and make it work for yourself and so there is the possibility of making DI have better hooks for non-free stuff so that it could actually install it it's just a question of somebody becoming motivated enough to go do it okay there is some progress on it though last time Lars started he is actually also working on the kernel side to implement the separation and well basically it makes not that much sense to do it in the installer before that is ready because you have to know what your environment is and we need a little bit of help from our FTP masters to provide us with some Deputy Installer sections in the archives for experimental so we can actually test this and develop it further and we will be waiting for that for a while now so thanks everybody for coming and we are going to be talking on Thursday so