 Pudet, sveče. Sveče či je, da je več vseč neba. Zelo sem jaz. Prezeno, da je, da se več na projekte grab, v kratku, na vseči, in v prejzavom placu. Mi je Daniel Kipper, imam prijevstvo, in sem seboj, in jaz za grab upstimov. Oko povedam, da bomo povedati in bilo odreč tudi, da bomo povedati, z vsem izvok, da je tudi pravda. In da bomo povedali in se povedati na naši trakujskih pomečnih problemov, kaj smo tudi v projekti. Včas bomo povedali za vse input. Nezostaj je to človek, da se pošli, da sem zelo vsek, vsek in sem pošli hov glasba. Zelo sem vsek in služite ampak in všeč vsek, da sem pozdravil naprati in postožiti več svečno. To je sveč nekako, da pošli, da sem vsek svečnja. Alex, skupaj na odvajnega, grupne odvajnega na druh komponijo, zelo se je sreč na vrstru. Vladimir je zelo vzdušen, in da so nekaj nekaj nezaj, da se nekaj nezaj, da se nekaj nezaj, in da se nekaj nezaj, da so nekaj nezaj, da se nekaj nezaj, da so nekaj nezaj, da so nekaj nezaj, Skupaj, da je zelo zelo še naj breakdownovine. Zelo se prišlo na dnev. In zelo se še izgleda po vso povrstke. Kaj je zelo, da je vso povrstek? Kaj je zašlič, da je zašlič, da je zašlič? Kaj je zašlič, da je zašlič? Tako, da je zašlič, da je zašlič. Nene, da je zašlič, da je zašlič. Srečno, da je zašlič. Kaj je zašlič? Kaj je zašlič? kako se počuje, da se pričeši na početku. Tudi je pa tudi tudi počet, pričeši in počet. Tudi je pričo še vršati in tudi pričeši na početku. In je to možno, da so bilo početk, da se pričeši. Tudi je pa vse začetko, da je initardija, imači. Vse je pripravil od Robin in Matju. Grapprope je začelnje z Nikolače. XFS je začelnje začelnje inovi. To je vse zelo, da je to vse. Vse je začelnje, da je to vse. ButterFest got some quite interesting new features. There was a lot of work happening for right five and right six supports. This work was done by Gefrado. ButterFest also got ZSTD support from NIC who works for Facebook. We also worked on some security features for Grab. First of all, Grab got various files framework. This feature was initially prepared by Vladimir, but he was busy with other stuff, so I took over the project and added some fixes and cleanups and some extensions. We got this at the end of last year. On top of this we put UEFI Secure Boot Shim support and UEFI TPM Y.2 and 2.0 from Matthew Garrett. It looks that some people are testing this feature because we got recently some fixes for TPM driver. Jürgen Gross from SUSE added PVH support. Currently it works. It was tested by one guy, so it seems that it is working. Of course, on top of that we had a lot of fixes and cleanups for Grab. What's happening? Currently we quite closely cooperate with Trench Boot Project. Trench Boot Project is the goal of this project is to use all security features which are currently available on X86 platforms. I'm thinking about UEFI Secure Boot, TPM, TXT, et cetera. This is currently working progress. We have all features which are needed for this project in Grab, but we are also working on Uroot changes, Linux changes and other module which is needed to have this project working. You can take a look at GitHub page and you can get more information about this project if you would like to get involved, you are welcome. As I said at the beginning, there is still work happening around IEEE disk driver which is done by Eric. Saidly, he is involved in other stuff for our company. Alexander from SUSE is working on RISC-FIRE for RISC-GV. It depends. And Colin Watson from Ubuntu posted recently some patches for easier integration of GNU Leap with Grab. Andrei from ARM reposted patches for DHCP support because currently Grab does not support DHCP, just only BootP. This is quite surprising, but it works in that way. So we would like to improve that. This work was taken over from Andrei who is busy with other stuff. Never ending story, EFI MNP driver. From time to time, people come back to us and say that currently existing implementation for SNMP doesn't work and it shouldn't work. So we would like to get this driver. This will solve most of Boot issues on EFI platforms, over network, of course. But we would like to live as an option SNP support. So main support will be MNP and then if it doesn't work then Grab fall back to SNP. And Grab to point of for RISC. There is more and more questions coming to us. I had a lot of discussion about that yesterday. Initially we thought that we would like to have some features complete in our gap upstream but it looks that people expect earlier release and do not care about complete features right now. In general, as I can see, it is more expected that we will be releasing in time then with future complete things. So probably I will discuss with other my tenders today and we will take decision shortly and for sure we will announce this on GrabDevel. What are our pain points right now? We try to increase purchase review throughput and decrease response delay. This is quite difficult for us because not so much people are reviewing purchase and mostly I. So this is difficult and I would like to ask people to take a look at purchase at the area of expertise. This would help much. Even we are thinking about splitting responsibility to different areas and to get some submiterence or something like that. We will be discussing that. Anyway, I think that situation last year improved. As you may saw, we got many new features and we will be working for sure on improving the situation. Another project which kick off with Alex, this is reduction of number of purchase, custom purchase needed by various distros to make GrabUpStream usable. The worst situation currently is in Red Hat at Fedora. They have about more than 200 purchase, custom purchase with various things and during the FOSDEM conference we discussed that, we go through all of them and some of them looks quite interesting for the project to get upstream but some of them are distro specific so for sure we will not take them. And this also is quite close to improving overall cooperation with distros. We would like to be closer to the distros, to the distros mitanner and we would like to get more input and more input and more help. During, after introduction of very famous Firmock we had huge fallout because we touched a lot of things and by mistake broke some other platforms and features and distro that we need some kind of script or some kind of solution which will help to discover at least build problems. So currently I'm going on the script which can be used on my platform but I'm going to upstream it. It should be quite simple and easy to configure so this way almost everybody on own platform will be test all, all some platforms. Of course I'm also trying to introduce support for Travis and also it will be upstream. Another issue which we have is that some people start posting the purchase on the mailing list. I've or other people asked for some changes and after that people disappear and drop this work in the middle. This is bad because first of all these people waste their time. We waste our time on reviewing the purchase and we also lose quite nice features, fixes and cleanups and additionally because due to our workload we are not able to take over this work very often so simply this harms our community in a lot of ways. So please treat us seriously and do not do that. Another thing which is also related to development it looks that some people doesn't read our comments carefully. I understand that some patches are quite long and you can miss some comments but very often I see that some half of the comments are taken into the account and some comments are skipped silently. I understand that some of them may not agree with my colleague's comments so if you do not agree just complain but if you think that it makes sense please take into the account and please post the purchase with all things fixed. Because once again these also waste our time and doesn't help to improve the project. And at the end I would like to ask you for some input to the project first of all what do you think about current work? What we are doing well? What we should improve? And do you think that our priorities are the same as yours? And we ask for some help because as I told you this is difficult work and there is a lot of work which should be done and for sure we as a maintenance we are not able to do all of this work. I am waiting for your questions. Partially we have a back website which is integrated with GNU and some of this work which should be taken over is on this backlist. But as I told you this is difficult for us to maintain something like that because we are currently focusing on let's say taking the code and reviewing the code and this is difficult. I think that it would help to have a wiki. For sure. Any other question? We don't have to write a new work configuration in case when you install a new kernel. Is that something you want to pick up or is that something you don't care about? Yes, sure. The question was what was our feeling about bootloader specification. Is it correct? Yeah. Okay, thanks. When we were looking at the Red Hat and Fedora patches we saw a lot of patches for bootloader specification. This is not our target in the new feature but for sure we will be working with that because at least yesterday one guy was asking about that. So I think that we will be working with that. But Vladimir, I think that we can say more about that. The PowerPoints we checked basically you don't want to have the list of kernels anymore because that makes it impossible for you to sign the config. And so I am working on some config which will describe the way to boot without listing every kernel. And then there is something that you would be able to sign and so on. And that is something that I hope can be adopted by several distros. Because right now the current config file it is good if you want to write to the plugins program. That complex logic and so on is very useful for this. But for normal user it is kind of over well. Another question? Yes? Actually more of a comment. I have got a couple of patches in Chrome. I can say that the development process at least for me as a distro person has clearly improved recently Thank you very much for your effort in going to the people. Ok. It is much easier to get things in than it used to be and it is definitely great to see what is going on in good constructive reviews of people. Thank you very much for that. Thank you. Specifically I was saying that it is a bit different because I have told you that it was my fault that it was so difficult to get batches because when I was a student it was kind of okay to find enough time for it but then with work and family it became way more difficult. It is hard. We all understand that. That is really the same course. Thank you. Thank you. And there is a big stack of webhack for door patches and all of the distros have a stack of patches. It is going to take time to go through and get all of those things merged and hopefully they don't diverge too far that we can get more common stuff into core growth. The question... The question... The big batches which are carried by the distros are secure goods and that is something that Daniel took care of. The question was about taking all the specific patches which are in distros. I would say that the worst thing is... Let's say the worst thing is Red Hat and Fedora. I think that Debian and other distros has much less distros-specific patches. Our goal is, as Alex told, we would like to compare this list which we discussed before the FOSDM and compare it with other distros. At first we will be taking the most important patches and the smallest one because I think the smallest one it will be much easier to get and quicker to get to the upstream. So this is the goal but as you said this is a long-term project and I suppose that it will take one year or something like that. It depends how many people will be involved in that and for sure we would like to ask for help distros maintenance in this thing. First of all we will initialize the discussion and then I hope that we will speak the work among these people. In a specific enterprise platform in which you would break other platforms that they try to have basically and doesn't care about and they try to contact the original author he is available you can speak with him but then they disappear when it's about fixing their patches. OK. Is there any other questions? Yeah I think this is let's say we would like to let's say coordinate this work. I'm not going to report this patches let's say. To be fair there will be a lot of patches at least in the end which are all upstream already. So it's like because there hasn't been a release for a long time. For example all this spark stuff this has already been in upstream. OK. Is there any other way to explain is it often the patches to fix some specific issue on specific platform so maybe we should break everyone else and then they don't want to suppose that sometimes they are often surprised outside deadlines so they prefer to quickly find something for the search work rather than It's the exact same issue as the kernel but in the kernel they managed to push the patches upstream. Why don't they do that on the work? That there is always a reason and the reason could be because maybe that's the reason I think that Alex's point is good and we will be discussing that I think. 250 patches, that's a lot of backwards. I think if we start getting to time based releases where distros can actually base their packaging on upstream releases that are a lot more reason then you'll start seeing a lot of those patches going away and it will be a lot more clear what's actually distros specific and which is just backwards and sometimes if you're backporting patches then you're only getting part of it and there are other changes that you need to throw in additional patches to make those. Yeah, I know what you mean. Just adding on that. 5 minutes left and other questions? I think the questions, sorry I will repeat it because we are recording there was a question because I understood correctly what we are going to do before the next release and what we are going to take in the next release. I think that we'll be discussing this today and we'll take decision quickly because I think that this is the problem for people so I would like to solve this quite quickly. But give us some time and we will post shortly our decision on the grab deval so please please stay tuned. I'm not sure that I understand the question could you repeat? I think that at least some of work can be done by Travis and I told that I'm going to take a look at Travis too. This is the goal. The question was about CI testing and integration and does Travis will help us to test grab. I think that it will help. At least some. Ok. Anything else? Ok, thank you for coming.