 I think I've removed 50 of these. Could you use a bin enemy? No. There's no bin enemy for... I can definitely... My desk... Sorry? We could actually ask if it's working now. Is it supposed to? Well, there is now an odd all... Yeah? Yeah. Wow, that would be great. That's awesome. Anybody who wants to rebuild those packages in advance, so they don't... Yeah, I mean, refreshing the packages is awesome for the idea. Yeah, because they are really, really old. The packaging architecture will be... The whole game to depth should be refreshed. Yeah. And add the header line, the control file for testing. For the test use. I think we do need to spend some effort doing this because those old packages are being dried, since squeeze or whatever. Yeah, so much for removing packages. Let's move on to Ruby 2.2. So we have the... Sorry, did you talk about Ruby Archbishop? No. So Archbishop has been unbending upstream for years and unfortunately not everything has been switched to something wrong, including especially from the documentation generator. So there is a bug upstream event? Yeah, there was no activity. Was open about three years ago? Amazing. Everybody seems to... Please send patches. Yeah, thanks. I don't know. I don't know enough about hprecord or... I think we have added this note that is deprecated inside the Debian package. I did it, I think so. Or you did it, I don't know. Yeah, but we need to fix it. Yeah. Because otherwise we're going to remove lots of stuff. And the issue that it uses hprecord to generate the HTML, which is really not the common use case. So in North Faro there was... We migrated away from hprecord to North Oguiri some time ago, but it was mostly parsing HTML in test switch. So I'm not... If that was the case, I could give an example of the exact committee that did that and we could do some ideas, but it's not the use case that we're using it for. I guess, yeah, we should try to do that. And you said fast access is only used by... We should probably do that. Now we're reaching point two. We have a transition for a few months already. It's pretty much done. It's at 80%. It's missing switching the default on unstable and rebuilding some stuff that builds against the default like subversion and some other things. But then, if you see 5K, they won't talk. So it would probably be a good idea to help with this 5 transition. There are four or five Ruby packages in the list, looking for stuff to do. So that's that. I think we should add a note, yeah. Thanks. I think one package is also affected by another transition. Yeah, there are a few collisions. Not only one, a few more. But the GCC5 collides with everything. And also collides with the transition of cabinet, I don't know what package it is. Yeah, but at least here it's ready. So you can already use Ruby 2.2.4 by installing the Ruby package for experimental. And you can test, if you install Ruby-all-dev from experimental, it will also stop building against 2.1. Which is not going to be the case when you use a build or whatever you use to build a McLean CH root. And that's not like I'm using Ruby 2.2 as default but I'm still building for 2.1 as well. But then, I didn't see any problems at all. I use it every day with Vagrant and the development I've been doing just works. Then we can plus cross-stretch. I think it's pretty clear that we probably want to ship 2.3 with stretch since it comes out on Christmas. Then 2.4 comes out. If everything goes like Jesse was, 2.4 will ship one month after the freeze. So, there is nothing written there. When does freeze of stretch must off? They didn't announce it officially but I expect to be something similar than last time. In July or something else. Was it in November? Ah, the freeze. I mean, we should be ready by this time next year. 2.3 by 2.4 and everything migrated. And then I need people to help with tracking the transition. I mean, maintaining the interpreters is not that much work. But then tracking down these broken packages and transitions takes most of my time. Maintaining the interpreters is usually simple. And since I'm getting involved in other stuff, it would be nice to have more help to when they will be and do everything else that comes with it. Like, bugging people to fix that shit. Can you explain a little bit your workflow when you are, for example, what you did during the 2.2 transition? Okay, I actually started taking notes. So, there is a page on the wiki where I documented how the transition works. So, we do the transition in phases. Phase one is just a protein interpreter to unstable. You don't need to wait anything for doing that because it's not going to be the default. Then phase one is enabling both interpreters. Then this needs a little of coordination because as soon as you enable a new interpreter, things might start failing to do. So, you want to do an experimental upload first and then do a test rebuild of everything to see if the stuff works. Then there is a fixing phase, making sure stuff builds with both interpreters. And then phase three, which is the pointer view. Actually, we are on phase two now. And phase three is making the switch. It's switching the default and disabling building against the old one. And then phase four is rebuilding the stuff that always built against the default. That's subversion. What else? X, MMS. God knows why. And 20 to 30 packages that always built against the default. But at this point, we should already have tested on experimental. It should be just a matter of sending, being on the news and waiting for the build. And then phase five, we just removed the old interpreter and being on the new everything to adopt the support for the old interpreter. So, we actually came up with this plan in the very first movie screen in Paris two years ago. I mean, it's a way to never break and stable while the transition is happening. I think this time testing is mostly fine. We don't have break issues. Except for the packages that are actually broken. Yes, we have issues. Please help with that for the next time. Then there is aspect three. I think we should probably have all the FTB fast bugs we should have by now. I think the reproducible builds people feeling a lot more of those. It's getting less and less, I think so. But the others, the package remaining are very hard to fix, I think so. Either you look yesterday and yes, sometimes you need to be tested with a test. Did you try to understand? Yes. The problem is that the whole is also removing that application warning. And then the diff is very huge. And then changing tests and you don't know about the software is I think hard to make because you have no reliable tests and you don't know if it's really okay. You are only checking if the tests are not failing but is the output really correct. So it's... A good example is the YARG which is the stream is very strict on keeping the version 2 aspect. Because I do not want to modify the test which is very huge. And last time I checked it would involve like thousands of changes in the test so it's really difficult on the outside to do things. So maybe we should just disable the test and replace this by super... Yeah, that would be a good idea. It would be very difficult to avoid and then maybe getting a batch with a huge batch against it that's sweet. It's not a good idea because it's probably going to break in the next... Break in the next absolutely to inspect every time and this makes packaging new upstream horrible. Any of those are left? Is it a lot? The UDD database is so much behind I have no doubts about how many bugs are left open. It looks like there are probably 50 or so 50 of these bugs. But I fixed something like yesterday or something there already. There are 20 already in testing on there. Ah, okay. But also it's not one of your approach is disabled street by street. Right. I think that will be a good solution. I have tests and I'm not that much involved on having an upstream batch. It's funny because sometimes a few wants to use the very latest breakers and then sometimes a few wants to use the very old thinning also breakers fighting the world. Transition to Rails 4.2 So I'm coordinating this with Praveen by email. We have a wiki page. Works well. We track what's ready where. It's only missing mostly making sure... updating RedMind to the new upstream and make sure it works. The new RedMind upstream already works so it's probably going to be okay. It's mostly done. Some things I upload experimental like new versions of Sprockets and Sprockets Rails and probably a few of those things are pretty much tied with Rails and actually before I updated Turbulence Rails didn't work at all. In this case I do reroute package test for Rails which is very nice. So it just creates a new application generates a resource and model control and test and stuff and runs the test for that. So that helps a lot to catch those problems. Even on CI when some dependency gets updated and breaks stuff. This one is each lab. So we just noted the other day that we are at 94% of the dependency for each lab and I think I think I can say that on record it's in a public way. He got the each lab people to fund two months of his work and that's probably going to happen soon. I think he's working already on something because he sent me a mail. Yes. He has some of the parts that are stashed but not that finished yet. At least there's upstream sourcing that he proposed to me. No. I'm not sure about packaging. I didn't do it. But he's also finishing the diaspora effect. Too many things at once. The diaspora is even for diaspora there's actually packages. Not only the dependency. And then there's also on that front there's also open build service is being packaged. So the actor and the angel they are from Palabra. They have OBS package for Jesse. They are working. They had some problems to update to the latest upstream and I gave some help to them in the last days and then I encouraged them to put that in the archive and have OBS. OBS is very nice and if anyone knows it it's a build system created by the open source it builds at the end and whatnot and it's very useful it has things like if you upload something it reviews all the best dependencies for that. It's very nice. And then we can think of world domination plans using OBS CI Yes, OBS. Pretty soon we're going to have four rails application. I plan to have no Sphere package as well for the way I'm going to do that. There's a related issue with the rails assets package. No, so OBS. So I noticed that there are more and more rails asset packages entering archive that are maintained by the team because of this web application we have and I think that we have at the moment no documentation to understand that they are not quite like usual gems so at least it would be nice if people who already package things like that would add some documentation in the wiki and also I noticed that some of these packages are a huge portion of it is in fact Javascript and they are shipping libjs something by new packages in the Ruby section so maybe we should extend the Ruby policy but it's not. Available yet to also understand. I think we need more packages the application available in the archive Those would actually go into the JavaScript team. Yes, but is the Javascript team interested in maintaining the very short bit of Ruby? Is it something similar with Python? I guess it's the same issue. Remember they have something called where a bunch of IPs appear for Python packaging whatever Yeah, because of Changa Well, do we have people in the JavaScript team? Do we want to propose a blog about that? Maybe it isn't too late at this point to try to get the people that have to deal with that to figure out how to do it If I know that the Python people are doing I would say that Perl people are probably doing it as well because of those other frameworks Can you try to do that? You can just send an email to that conference girls right now and then we can find the time to sit together So there is removing the transition of packages? The list is outdated I generated yesterday Yeah, I did Oh, I did And I also got some packages that will be removed or are removed by Christian I think so I will re-generate the list and take care of them to get the doubt So you can Maybe you can get together some FTP marches just remove all those Just close all those books It's probably actually very fast at doing the R&D Yeah, just file the back in Yeah, it takes about two or three hours for the packages One question about these packages are not maintained by the team I think the maintainer is very slow response about these two packages Can we potentially try to make it a network? Does it check the package names? I think I did something Yeah, but there is no upload I think so on the image size about years So perhaps we should hijack those packages Solverage Yeah Take them to the team and put them to the uploader Yeah, I think I don't know, we should write an email or something else, but there is an open back and there is no response You want to do that? Yeah I will write an another email and say we will take over this to the Ruby team we are staying in uploader field and we will do this in four weeks if you don't respond to that back or something else So the maintainers have time to respond I think you can skip to waiting I think in this case you can skip to waiting Okay You can just upload the largest delay queue you have Yeah, delay 15 I think it's the largest one Alright, that's a good idea If you don't do anything in 15 days Actually, there is nothing to do with those packages Are you still done? All of them But there are few Yeah, but those are team packages Oh, there are these more, right? Okay, I didn't look at those You have to talk these less right here Yeah, these were really crazy Okay, so then you take care of that Yeah, I will take care of that The maintainers and the uploaders too So there are some packages left that are not translated These ones are not even migrated to the new name Yeah, these are So we need to first delay transitioning packages and then so we need to do this after stretch release So we do not get these out before the stretch It's really... Provides work for that but I don't think it works so accurately Yeah, but I don't have... It's not a problem for me to have it in stretch release and some sort of decision-making Okay, you are going to do that as well? I stopped the tasks but I have not finished it Maybe next spring because it's the easier topic I mean I supported another work implementation like you How did that? Right now I am maintaining the Ruby in the Java team I don't have an easy way to run to break the Ruby So I was wondering if there is any plan or how I can coordinate with Ruby team to have a way to switch I am not interested in switching the Ruby command but I want a way to run general Ruby software with another interpreter Okay, so I only have the need but I have no do anything in that regard to fix it for the next In the past, the Ruby was only there to satisfy that but I think that the Ruby now is good and it's an important implementation to run I don't know what commands we have I think it's hard Can you use Ruby at the end? If you only need it on the user level then No, but I think it's weird that we have J-Ruby in the archive and I can use Ray or something like that I can import but I can run some but you can use Ray-Ruby- as Yeah, I can use that I think it's a bad idea to switch the system Ruby We use it to have user being in Ruby and then we just switch the Shebangs back to user being Ruby be safe so you can do like this should work? Yeah, that works Other than that So maybe I just use the s-flag and document it in J-Ruby? Probably Do you want to use Ray with J-Ruby using the system Ray? Yeah And J-Ruby will use the system Ray library? Yeah It depends on the number and Ruby in binaries We should break in the integrator Yeah, we should, yeah, yes No, we have snow packages Okay The interpreter seems to rate the number and user being rate is from the rate package because it can have a look at Ruby stand alone I don't know if you have said so It's a package I need to be able to run Ruby without any of the damage For instance, when I am sending upstream patches I want a clean environment like they would have it does the opposite but they treat with changing the path to make sure in this case I also have a rate there Oh, okay So you can have which is something like RBM does with little wrapper programs or whatever I'm going to try that You are going to call the right interpreter for this Okay Those scripts are really trivial Actually, if you have dash s then you can make those scripts all just If you want we can give you another new binary for Ruby stand alone source package and have Ruby dash 3 Okay, I'm going to try and continue in the Not sure if it's the should be able to do Yeah, but one is something like one, two comments Yeah, you can run one single comment and then you have a share where everything is given Sorry, what's on you? Ah, you are Miguel Nice to meet you We have five minutes left Is there something more important than the next team sprint? What attempt of the Ruby sprint? Is there any interest in the Ruby sprint? No, sure Yeah, sure So we have one, two, three, four, five Do we have volunteers to organize in there? Which part? Outside of Paris Paris people are Who's from Europe? I think it's that you come to Europe Yes But if you Are you going to be in Europe? Like, when did we do the sprints in February? We should sprint Do the sprint after the release of Ruby 2.3 Right Yeah, early next year So I'm in Europe You are in Europe even? Sorry Sorry We gave it all to you You need to share a flag Yeah, but If you guys want to spend a week in Brazil during the summer We can do that That's something that you said in the video talk that we do all the sprints in Europe And actually Of course, it's easier from a logic for the view, but then maybe we are reinforcing what's already I mean, you should look at it I've always talked in my Debian Talks in Brazil How I showed the Debian developer map And then I showed it In Europe, you can barely see the ground And There's some If you guys want to go to Brazil Maybe we can check If Brasilia has a Brasilia has probably a good connection with Europe Do you know if there are direct flights to there? There is South to the right Are you guys up to it? For you, it's not so bad to get to Brazil What is it in Europe? If it's in Brazil, there is a good chance Okay, let's do it Okay, let's see how we do it We have to check practical stuff So we are going to be tickets to Brazil in a server Well, the question is about the land But last time we actually did the free days to score a little bit Maybe once to do a longer since you have to fly on the range 5 days or something Yeah, probably 5 days in the week so I have to go and fly Yes I will just sleep during this Okay Yeah, I think we are done We are out of time now, thank you everyone for your help Let's have fun watching the video