WEBVTT 00:00.000 --> 00:20.000 So, I'm going to quickly introduce myself, so I've been a deviant developer for more than 20 years, which gives you an idea of what my age. 00:20.000 --> 00:31.000 And my work, while I'm getting paid to do, is to, I'm a director at Muzia for, sorry, Muzia. 00:31.000 --> 00:35.000 Do you use the Ecoso? No, it's good, better. Cool, thanks. 00:35.000 --> 00:42.000 So, my page job is to work at Muzia, and this work is completely related. It's not a Muzia project, it's not supported by Muzia in any way. 00:42.000 --> 00:48.000 It is one of my side projects that I'm doing in the evening and doing the weekend. 00:48.000 --> 00:57.000 I have been involved in the rest community for a long time. I uploaded the first version in Debian while back, and it's now in the Buntu. 00:57.000 --> 01:03.000 And I've been a maintainer of some key packages in Debian and the Buntu based on the rest for a while. 01:03.000 --> 01:07.000 So, here is what I'm going to talk about. First, I will ask you some questions. 01:07.000 --> 01:13.000 Then I will present what we have been doing with the implementation of the core deals in rest. 01:13.000 --> 01:19.000 I will do a demo, and then I will make a proposition to the open source community. 01:19.000 --> 01:22.000 So, first, I'm going to ask you some questions. 01:22.000 --> 01:30.000 So, who are here in the room as pung opinion about the CPL, and especially the BSD and MIT? 01:30.000 --> 01:36.000 So, for the recording, it's about like one third of the room, one quarter. 01:36.000 --> 01:42.000 Myself, I don't care, as long as it is open source in the OSI and FSF definition, I'm completely fine. 01:42.000 --> 01:47.000 The project is not about license, even if I'm going to talk a few times about licenses. 01:47.000 --> 01:52.000 Who asked for the opinion about the new project? Not that I'm saying the new project, not the FSF. 01:52.000 --> 02:01.000 So, I see only a few hands. Same. I still contribute to a few new projects like career deals or GCC, also kind of things. 02:01.000 --> 02:07.000 So, Miguel did a talk right before about the Linux kernel and we're writing some part of that in C. 02:07.000 --> 02:12.000 So, now the question is, who asked for the opinion about C? Yeah, a lot more people. 02:12.000 --> 02:22.000 Yeah, yeah. I really don't like C. I think it's going to die, and at some point I think our kids are going to put C in the same category as Fortran and Cobalt. 02:22.000 --> 02:27.000 I think it's time to move on, and this is why we need to start that work now. 02:27.000 --> 02:30.000 And who asked for the opinion about first, positive? 02:30.000 --> 02:33.000 Yeah, many people call. So, you're in the right room. 02:34.000 --> 02:36.000 So, now what I'm going to tell you is this quick story. 02:36.000 --> 02:41.000 So, it has been five years that we all had to go in lockdown. So, it was 60 for everyone. 02:41.000 --> 02:48.000 And some of us, we made some bread, some people, the guy on the right, he was in the LL Ventura. 02:48.000 --> 02:52.000 He did some woodworking, some crazy stuff in with wood. 02:52.000 --> 02:56.000 Myself, I, with my son, we started building also, they go in the boxes. 02:56.000 --> 03:00.000 So, you can see Wally and some Star Wars and some Batman stuff. 03:01.000 --> 03:05.000 But I needed a project in the evening. So, I was like, I want to learn the rest. 03:05.000 --> 03:10.000 West is a material project. I have been surrounded by upstream rest developer, 03:10.000 --> 03:14.000 inside the Paris office of Maria, some several developers. 03:14.000 --> 03:19.000 So, I want to learn it myself, but I'm not a student anymore, and I don't want to do a site project 03:19.000 --> 03:24.000 that it is going to be just sitting on my otherwise. So, I want to do something impactful. 03:25.000 --> 03:31.000 So, I was thinking, what about implementing the corridors in the rest? 03:31.000 --> 03:36.000 Especially, I was curious to know how it was working. I saw it was full of assembly. 03:36.000 --> 03:40.000 But actually, it's a very simple scene most of the time. 03:40.000 --> 03:44.000 And I was very inspired, I've been contributing to LL Ventura, LL Ventura, 03:44.000 --> 03:49.000 for like 15 years, and I was very interested in inspired by the works at Chris Latner. 03:49.000 --> 03:57.000 That Apple, one of the core fundamental of Clang, is if you have different behaviors and CC, it is a bug. 03:57.000 --> 04:01.000 And I was also involved in that project with my debut an hour, a long time ago. 04:01.000 --> 04:06.000 I started with building the debut an archive with Clang, and I reported plenty of bugs upstream, 04:06.000 --> 04:09.000 and I fixed some of those. One of my favorite is this one. 04:09.000 --> 04:14.000 So, some people saw that if you put a dash or six in CC or Clang, 04:14.000 --> 04:17.000 it's going to optimize more than all three or all four. 04:17.000 --> 04:20.000 We even held in the debut and could base a dash on 99. 04:20.000 --> 04:24.000 Someone saw that it would be way more optimized than three. 04:24.000 --> 04:27.000 Of course, it doesn't change anything. 04:27.000 --> 04:33.000 Now, one more question, who knows, in that list of corridors program, who knows one of them? 04:33.000 --> 04:37.000 Well, almost everybody, who knows five of them? 04:37.000 --> 04:42.000 I have a few. Who knows everything here, in that list? 04:42.000 --> 04:47.000 I'm pretty sure nobody, except if there is a new upstream developer, one of my favorite is PR. 04:47.000 --> 04:52.000 Who in the room knows what PR is about? You always have that on your system. 04:52.000 --> 04:55.000 If you are running Linux, I see one guy over there. 04:55.000 --> 05:01.000 It's basically to format the text for printing, actual printing on the printer, like paper and so on. 05:01.000 --> 05:05.000 Some of the comments here are not very useful or used anymore. 05:05.000 --> 05:09.000 They are legacy, but for compatibility reasons, they are still over there. 05:10.000 --> 05:16.000 You're all familiar with those. Like, I selected those offensive ones, or CH-mode, CH-roll, install, and so on. 05:16.000 --> 05:17.000 L-S, M-Cardier. 05:17.000 --> 05:25.000 This is a comment that you folks, if you are running a Mac or Linux, you use those almost every day for your work. 05:25.000 --> 05:29.000 So, it is a full list, 105 comments. 05:29.000 --> 05:35.000 So, now it's, we are five years later, after COVID, five years later of working on that project. 05:35.000 --> 05:39.000 As you can see, I have way more gray hairs than I had five years ago. 05:39.000 --> 05:44.000 We have in the project, this is a screenshot from last week. We have 530 contributors. 05:44.000 --> 05:48.000 How many projects has that number of contributors? Not that many. 05:48.000 --> 05:51.000 We have more than thousands, we have right folks. 05:51.000 --> 05:59.000 And we have almost 18,000 star on GitHub. I know that it is meaningless, but still it's one of the metrics that we have. 06:00.000 --> 06:07.000 And I'm very proud to show that this is a graph of when we run our implementation in Rust. 06:07.000 --> 06:11.000 Again, the C test suite, the C-based state suite with new. 06:11.000 --> 06:19.000 And as you can see, we started, we had 139 only test passes when we started the project in 2020. 06:19.000 --> 06:26.000 And now, we are almost 500. I should have worked as weekend to be at 500, but I didn't have the time. 06:26.000 --> 06:30.000 We are very close to a full completion of the project. 06:30.000 --> 06:35.000 And most of the tests that are failing right now, some of them are in PR and some comments that nobody cares about. 06:35.000 --> 06:40.000 Or sometime we have some weird, I will put ordering. 06:40.000 --> 06:44.000 So, if you delete, if you use RM on the directly and you don't have permission, 06:44.000 --> 06:52.000 new is going to show something first and we show it at the end, so it's going to make a small difference. 06:53.000 --> 06:57.000 So, what we are also doing is we run the test independently. 06:57.000 --> 07:02.000 So, we have that fancy dashboard, which is telling us that, for example, with Scott, 07:02.000 --> 07:09.000 we have one test that is failing currently from the new test suite, but C-H group and the C-H zone and C-H roots, 07:09.000 --> 07:12.000 there are fully compliant with the new test suite. 07:12.000 --> 07:17.000 That doesn't mean that it is fully compliant because it is the new test suite as a limitation. 07:17.000 --> 07:23.000 So, we are contributing upstream to the new project when we see some stuff that are not tested. 07:23.000 --> 07:28.000 But that shows that you can use it in, you have, you can trust the code. 07:28.000 --> 07:30.000 You can trust that it is working well. 07:30.000 --> 07:36.000 And it's, you can also verify that you are not progressing when you are making some changes. 07:36.000 --> 07:40.000 So, the current status, I'm very proud to say that we are production ready. 07:40.000 --> 07:44.000 We support Linux, Mac, Windows, Android, open base day, freebie, SD in at BSD. 07:44.000 --> 07:47.000 If you use Fushia Redox, then Wazam. 07:47.000 --> 07:53.000 So, it's most of those target are in the CI, not all of them because GitHub doesn't provide everything here, 07:53.000 --> 07:56.000 but we have a lot of good coverage. 07:56.000 --> 07:58.000 It is also now used in production. 07:58.000 --> 08:03.000 Most important thing on my laptop, this laptop is running that implementation. 08:03.000 --> 08:07.000 So, it is a kind of production, but we have also, 08:08.000 --> 08:11.000 a collaborator who worked on that project to ship it on an apartment. 08:11.000 --> 08:16.000 So, if you have an apartment's version, it's going to have our implementation. 08:16.000 --> 08:22.000 If you are in, if you are young, you might know about Snapchat and spectacles, the glasses, 08:22.000 --> 08:24.000 they are running our implementation. 08:24.000 --> 08:28.000 I must admit that those two are mostly for licensed reason, but it's working well. 08:28.000 --> 08:30.000 There are certain production. 08:30.000 --> 08:33.000 Microsoft is also using it for VS code web. 08:33.000 --> 08:37.000 It's a lot of article about it because it is written in the rest, 08:37.000 --> 08:40.000 compiling the rest to web assembly, it's pretty easy. 08:40.000 --> 08:43.000 And it is one of the beauty of open source. 08:43.000 --> 08:47.000 It's probably elsewhere, and if you use that in your company, 08:47.000 --> 08:49.000 and you didn't tell me before, please let me know. 08:49.000 --> 08:53.000 I'm always excited to know when people are using our code. 08:53.000 --> 08:57.000 And it's almost all Linux distro. 08:57.000 --> 09:00.000 Some of them, at least one, that I know, 09:00.000 --> 09:05.000 the third point to us is now shipping by default with that implementation. 09:05.000 --> 09:09.000 So, before that took me well already covered some part, 09:09.000 --> 09:12.000 but here is my point of view of why it works. 09:12.000 --> 09:15.000 Why it works for this? 09:15.000 --> 09:22.000 We have, we are seeing that rest is replacing C and C++, almost everywhere. 09:22.000 --> 09:25.000 So, I'm completely biased. I work on Firefox. 09:25.000 --> 09:28.000 It has been my work for 11 years and 10 days. 09:28.000 --> 09:32.000 So, I was part of the release management team, 09:32.000 --> 09:35.000 when we started shipping rest in two Firefox. 09:35.000 --> 09:37.000 I'm completely biased on the browser part. 09:37.000 --> 09:40.000 We know that Chrome is also following our leading database, 09:40.000 --> 09:42.000 a Linux kernel and menu operating system. 09:42.000 --> 09:46.000 We know that Microsoft is using it for Windows. 09:46.000 --> 09:50.000 Going to study reviews, but rest is very good for security, 09:50.000 --> 09:52.000 for parallelism, for performances. 09:52.000 --> 09:55.000 And I'm going to show you that in the demo. 09:56.000 --> 10:00.000 But I like also to say that we are not replacing the simple implementation, 10:00.000 --> 10:02.000 because they did a poor job in terms of security. 10:02.000 --> 10:04.000 They did an amazing job. 10:04.000 --> 10:08.000 They almost don't have any security issues in their codebase, 10:08.000 --> 10:11.000 they are doing amazing work, and it's super safe. 10:11.000 --> 10:14.000 So, they're not about being better in terms of security. 10:14.000 --> 10:18.000 As I'm playing several times before, it's not about license. 10:18.000 --> 10:21.000 I'm not interested, I know that some people care a lot, 10:21.000 --> 10:24.000 but it's not why I'm doing it. 10:26.000 --> 10:28.000 Second part. It's very portable. 10:28.000 --> 10:30.000 Some people are surprised when I say that. 10:30.000 --> 10:32.000 Unlike Java, it's really really portable. 10:32.000 --> 10:34.000 So, I mean that if I write a code, 10:34.000 --> 10:38.000 usually I'm almost certain that it is going to work well on Windows and Mac, 10:38.000 --> 10:40.000 and Android, which is very surprising to me, 10:40.000 --> 10:42.000 given the complexity of everything that we do, 10:42.000 --> 10:44.000 and everything is platform-specific, 10:44.000 --> 10:47.000 but things to the ecosystem and things to the quality of rest. 10:47.000 --> 10:53.000 The cost of parting a change to an operating system is very, very small. 10:54.000 --> 10:57.000 And I know that one is up to debate. 10:57.000 --> 11:00.000 Myself, I'm on the side of AMAZE, 11:00.000 --> 11:03.000 so if there is a great doing library doing that work, 11:03.000 --> 11:05.000 I'm going to use it, I'm not going to implement it. 11:05.000 --> 11:09.000 We have like 200 dependencies in the rest of our materials, 11:09.000 --> 11:11.000 because we don't want to be implementing. 11:11.000 --> 11:15.000 I do understand that there is always a supply chain attack risk, 11:15.000 --> 11:17.000 but it's a risk that we are willing to take, 11:17.000 --> 11:20.000 and they are more and more tooling around that to protect ourselves. 11:20.000 --> 11:24.000 And if you did some Windows Mac and Linux development, 11:24.000 --> 11:28.000 in your career, you know that parting new dependency on a different, 11:28.000 --> 11:30.000 another operating system is a nightmare. 11:30.000 --> 11:33.000 And the rest is usually as easy as typing, 11:33.000 --> 11:35.000 cargo, add, and the name of the crate. 11:35.000 --> 11:39.000 Usually everything is managed by for you by the build system. 11:39.000 --> 11:42.000 I'm a developer and I'm a geek, 11:42.000 --> 11:45.000 it's also super interesting to learn a language. 11:45.000 --> 11:49.000 Like it's part of what makes our job super exciting, 11:49.000 --> 11:52.000 is learning new things every day. 11:52.000 --> 11:54.000 And as I'm getting older, 11:54.000 --> 11:55.000 I'm my son is ten years old. 11:55.000 --> 11:59.000 I also think about what we are going to live to the new generation. 11:59.000 --> 12:00.000 There's a new generation. 12:00.000 --> 12:03.000 They don't want to do cobalt, see, or for twins. 12:03.000 --> 12:04.000 They want to do fancy stuff. 12:04.000 --> 12:07.000 Like rest, we have to go or cut in. 12:07.000 --> 12:10.000 And I think it's a good investment for the future 12:10.000 --> 12:14.000 to start now investing in the new technology and the new languages. 12:14.000 --> 12:17.000 And the new approach for the future of the Linux ecosystem. 12:17.000 --> 12:22.000 And the computer science system ecosystem in general. 12:22.000 --> 12:24.000 Now I'm going to do a demo. 12:24.000 --> 12:27.000 I'm going to do some product placement for some of the cool stuff 12:27.000 --> 12:29.000 that we do at Medziab because everything is intertwined. 12:29.000 --> 12:32.000 And if I know about those tools also because I work at Medziab 12:32.000 --> 12:34.000 with the folks who develop those things. 12:34.000 --> 12:35.000 So please bear with me. 12:35.000 --> 12:36.000 It's a demo. 12:36.000 --> 12:37.000 It's never perfect. 12:37.000 --> 12:40.000 And I need my notes. 12:40.000 --> 12:41.000 Here we are. 12:41.000 --> 12:48.000 So yeah, we can see cool. 12:48.000 --> 12:51.000 So I'm using Alacrity Terminal Invest. 12:51.000 --> 12:53.000 And so point is Starship also Invest. 12:53.000 --> 12:56.000 So most of the tools that I'm using on Medziab. 12:56.000 --> 12:59.000 I'm packaging the Medziab and they are almost all I have to do 12:59.000 --> 13:02.000 the box with some customization to make them nicer. 13:02.000 --> 13:05.000 So you all recognize the pattern. 13:05.000 --> 13:07.000 So I've got a cargo.tml. 13:07.000 --> 13:09.000 So what I want to do is to show you, for example, 13:09.000 --> 13:13.000 that I'm using the best implementation of that. 13:13.000 --> 13:17.000 And if you are from Australia, you will appreciate 13:17.000 --> 13:19.000 an opportunity that we are telling news 13:19.000 --> 13:21.000 that you are spelling color wrong. 13:21.000 --> 13:23.000 Color should be spelled this way. 13:23.000 --> 13:27.000 But we are helping you to fix your English to write way. 13:27.000 --> 13:32.000 So what I want to do now is to show you one of the, 13:32.000 --> 13:36.000 so we take the liberty at times to, to different 13:36.000 --> 13:39.000 I was self on the new implementation. 13:39.000 --> 13:45.000 So what I'm going to do is we added a command, an argument called dash dash progress, 13:45.000 --> 13:47.000 where you have the progress bars. 13:47.000 --> 13:51.000 And if you have been using CP, you don't have that. 13:51.000 --> 13:56.000 And it can be very frustrating because you never know how much time it takes. 13:56.000 --> 14:01.000 I have not done the implementation. 14:01.000 --> 14:05.000 It's one of the other maintenance, third, third of the clubs. 14:06.000 --> 14:08.000 And it's pretty easy to add that. 14:08.000 --> 14:13.000 And if you see other places where we could add it, don't hesitate to let us know. 14:13.000 --> 14:17.000 We have also some algorithms that we are shipping. 14:17.000 --> 14:19.000 No, it's sorry. 14:19.000 --> 14:22.000 The demo effect that I said. 14:22.000 --> 14:31.000 So we are also, for example, we have that as algorithms that we are shipping. 14:31.000 --> 14:34.000 Of course, it's totally because it's waiting for something 14:34.000 --> 14:35.000 on the city in. 14:35.000 --> 14:38.000 But we are also taking the liberty to add more stuff. 14:38.000 --> 14:44.000 So what I want to show you also now is how easy it is to run test. 14:44.000 --> 14:46.000 So the command, cargo test. 14:46.000 --> 14:49.000 Sorry, I'm going to clear the screen, capabilities, cargo test. 14:49.000 --> 14:50.000 We all know how it works. 14:50.000 --> 14:51.000 We have dash dash feature. 14:51.000 --> 14:53.000 We say I'm running a unix. 14:53.000 --> 14:56.000 Because we have also unix dash SE Linux. 14:56.000 --> 14:57.000 I'm a debyan user. 14:57.000 --> 15:00.000 So I don't have any CD news and I don't care. 15:00.000 --> 15:01.000 But some people care. 15:01.000 --> 15:04.000 So we have also instead of unix. 15:04.000 --> 15:07.000 We have Linux, underscore SE Linux or something. 15:07.000 --> 15:11.000 And here I want to run only the test on sort. 15:11.000 --> 15:15.000 So I'm building the dependency. 15:15.000 --> 15:17.000 It's going to take a bit of time. 15:17.000 --> 15:18.000 It's not in with this. 15:18.000 --> 15:19.000 So here it is. 15:19.000 --> 15:20.000 I launch all the tests. 15:20.000 --> 15:22.000 And as you saw, it only took one second. 15:22.000 --> 15:24.000 And, sorry, one second. 15:24.000 --> 15:27.000 And one point two seven second. 15:27.000 --> 15:30.000 To run the test suite for all sorts come in. 15:30.000 --> 15:34.000 If I were doing that with a new, it would take 20 seconds. 15:34.000 --> 15:35.000 Something like that. 15:35.000 --> 15:36.000 And it's much harder. 15:36.000 --> 15:39.000 So it's very easy when you want to test your changes. 15:39.000 --> 15:40.000 Because you can quickly on the events, 15:40.000 --> 15:43.000 or test on the commands that you're attaching. 15:43.000 --> 15:44.000 So this is easy as this stuff. 15:44.000 --> 15:48.000 So what I want to show you now, it's my laptop falls. 15:48.000 --> 15:51.000 Of course, it's a demo. 15:51.000 --> 15:53.000 Let's move to a different terminal. 16:00.000 --> 16:07.000 You can see, yeah, perfect. 16:07.000 --> 16:11.000 So, ah, it's probably because I have something mounted. 16:11.000 --> 16:13.000 So it's why it's long. 16:13.000 --> 16:17.000 So on my system, I'm basically to run my implementation. 16:17.000 --> 16:19.000 I'm just avoiding the pass. 16:19.000 --> 16:22.000 So if I'm typing the full pass to the implement, 16:22.000 --> 16:24.000 I have a full increment. 16:24.000 --> 16:27.000 So you can see here that I'm running not the latest version. 16:27.000 --> 16:29.000 I'm using the package one in Debian. 16:29.000 --> 16:32.000 In Debian, we still have the 9.5, which is not the latest, 16:32.000 --> 16:34.000 which is important because they improve some performances. 16:34.000 --> 16:38.000 Anyway, ah, I have on my system, 16:38.000 --> 16:41.000 ah, all the books that Maria wrote. 16:41.000 --> 16:43.000 But because I speak in English, 16:43.000 --> 16:45.000 I'm going to use six years, an example. 16:45.000 --> 16:51.000 So I'm using a W-L and we can see that we have all six 16:51.000 --> 16:56.000 year books about 127,000 lines. 16:56.000 --> 16:59.000 So now what I want to do is to do some benchmarking. 16:59.000 --> 17:01.000 So I'm using hyperfine, which is a worse program. 17:01.000 --> 17:07.000 One more to do some easy benchmarking between a command line program. 17:07.000 --> 17:11.000 So let me type it from start. 17:11.000 --> 17:16.000 Debug, corality, sort, Shakespeare. 17:16.000 --> 17:18.000 Of course, I need quotes. 17:18.000 --> 17:21.000 And now I'm going to give the signal. 17:21.000 --> 17:24.000 So if you know about stress, you are telling to yourself, 17:24.000 --> 17:26.000 that guy on stage is stupid. 17:26.000 --> 17:27.000 He doesn't know what he's doing. 17:27.000 --> 17:32.000 He's using the Debug build of stress, which is often very slow. 17:32.000 --> 17:34.000 But it's part of the demo effect, right? 17:34.000 --> 17:35.000 So I'm running. 17:35.000 --> 17:39.000 So what if you're fine is doing is running 10 times the program. 17:39.000 --> 17:43.000 And as you can see here, that part, you have the result. 17:43.000 --> 17:47.000 And we can see that even have our Debug build of sort. 17:47.000 --> 17:54.000 So sorting all the line in Shakespeare books is 1.45 times faster 17:54.000 --> 17:57.000 than the new implementation. 17:57.000 --> 18:02.000 It's because we are using some rest week and work makes that super easy. 18:02.000 --> 18:07.000 So now because it takes forever, I already build with this bug. 18:07.000 --> 18:13.000 A village build of that. 18:13.000 --> 18:23.000 So instead of doing Debug, I'm just going to use a village build. 18:23.000 --> 18:26.000 It takes a few seconds. 18:26.000 --> 18:31.000 And we can see that we are six time faster than the new implementation. 18:31.000 --> 18:41.000 So now it is when I start doing project placements from the staff. 18:41.000 --> 18:44.000 So I'm going to use a tool that one of our colleagues developed, 18:44.000 --> 18:47.000 called simply, if you are into profiling performances, you should know 18:47.000 --> 18:50.000 but simply, if you don't, you are making a mistake. 18:50.000 --> 18:57.000 So what I'm going to do is I'm going to ask simply to do a profiling of the sort itself. 18:57.000 --> 19:01.000 So activities, it's analyzing. 19:01.000 --> 19:06.000 Now I'm going to go in Firefox and I'm going to open the Firefox profiler. 19:06.000 --> 19:09.000 It's one of the things that we developed to improve the performances of Firefox. 19:09.000 --> 19:13.000 So we can, I'm not going to go into many details about the few, 19:13.000 --> 19:16.000 but we can see the performances, we can see the values thread. 19:16.000 --> 19:22.000 And what I like is graphic or flam, flam, whatever in English. 19:22.000 --> 19:27.000 And so I'm going to do, I'm going to look at the profile. 19:27.000 --> 19:29.000 Of course, I prepared everything before. 19:29.000 --> 19:30.000 So I know what I want. 19:30.000 --> 19:36.000 But we can see that here, I'm spending 30, 29% of the time, 19:36.000 --> 19:39.000 just to do a resize of the vector. 19:39.000 --> 19:42.000 And if you went to university and studied that, 19:42.000 --> 19:45.000 it's one of the basics exercise. 19:45.000 --> 19:48.000 So why are we spending that long on that stuff? 19:48.000 --> 19:51.000 So I can double-click because of the res magic. 19:51.000 --> 19:54.000 I have access immediately to the source code. 19:54.000 --> 19:56.000 I'm not here to do a demo of the profiler. 19:56.000 --> 19:58.000 I'm going to do one more product placement. 19:58.000 --> 20:00.000 Some of my colleagues are going to present certain the web 20:00.000 --> 20:02.000 perform later today. 20:02.000 --> 20:06.000 But I can see that here, I'm spending most of the time in buffery size. 20:06.000 --> 20:12.000 And I'm like, oh, yeah, it's probably not the right thing to do. 20:12.000 --> 20:15.000 So I'm going to go back to the terminal. 20:16.000 --> 20:18.000 Just to show you Shakespeare. 20:18.000 --> 20:20.000 Here it is. 20:20.000 --> 20:22.000 The end of the bottom. 20:22.000 --> 20:23.000 The end of the last book. 20:23.000 --> 20:29.000 So what I want to do is I'm going to do a backup first of, 20:29.000 --> 20:32.000 so that we don't have to be compile. 20:32.000 --> 20:37.000 Of the debug bills, target the bug, corretiles, 20:37.000 --> 20:41.000 previews for first-dem demo. 20:41.000 --> 20:44.000 So it's super clear. 20:44.000 --> 20:47.000 And now I'm going to show you one of the patch. 20:47.000 --> 20:49.000 It's certainly, I don't want to put it in the tree, 20:49.000 --> 20:52.000 because I don't like and say it, but it's one of the tricks 20:52.000 --> 20:54.000 that you use to do an initialized memory. 20:54.000 --> 20:56.000 There isn't any safe at the bottom. 20:56.000 --> 20:58.000 So it's not some things that we are going to influence 20:58.000 --> 21:00.000 our better ways to do that in the rest. 21:00.000 --> 21:02.000 But it's just for the demo. 21:02.000 --> 21:05.000 So what I'm going to do now is to apply the patch. 21:05.000 --> 21:07.000 I'm going to do a rebuild again. 21:07.000 --> 21:11.000 Well, I will show you how quickly this to run the test. 21:11.000 --> 21:16.000 So I'm running the test to make sure that I don't request. 21:16.000 --> 21:18.000 I didn't request the test. 21:18.000 --> 21:19.000 Cool. 21:19.000 --> 21:28.000 And now what I want to do is to run sample again with a new cut. 21:28.000 --> 21:30.000 So I'm opening the profile again. 21:30.000 --> 21:33.000 Graphic of Flammer Core. 21:33.000 --> 21:34.000 And here it is. 21:34.000 --> 21:37.000 We see that the sort, the weak to buffer now, 21:37.000 --> 21:40.000 is we present only 5% of the time. 21:41.000 --> 21:46.000 Now what I want to do is to do some profiling of that. 21:46.000 --> 21:48.000 So my laptop is not plugged. 21:48.000 --> 21:52.000 So on the power, so it takes about 2 minutes to do a rebuild 21:52.000 --> 21:53.000 of an in-release. 21:53.000 --> 21:55.000 It takes 1 minute when it is plugged. 21:55.000 --> 21:58.000 So I will not, we are not going to wake 21:58.000 --> 22:01.000 together for 2 minutes for the compiler to finish the job. 22:01.000 --> 22:03.000 So I already prepared the build. 22:04.000 --> 22:09.000 So I'm going to run hyperfine. 22:09.000 --> 22:12.000 And now what I want to know is, 22:12.000 --> 22:18.000 I don't have completion in that case. 22:18.000 --> 22:24.000 Let's do the debug build instead. 22:24.000 --> 22:27.000 So I have the debug build that I prepared. 22:27.000 --> 22:32.000 I have the new one. 22:32.000 --> 22:36.000 We are still faster than the regular implementation. 22:36.000 --> 22:40.000 So the good target, debug, call it kills. 22:40.000 --> 22:47.000 Previews for them demo, Shakespeare. 22:47.000 --> 22:50.000 So now if your fine is going to run it on the free binary. 22:50.000 --> 22:52.000 So the one without my files, the one with a batch, 22:52.000 --> 22:55.000 and the new implementation. 22:55.000 --> 22:57.000 Ah. 22:57.000 --> 22:58.000 Cool. 22:58.000 --> 23:00.000 So which one it failed? 23:00.000 --> 23:03.000 I probably made a typo in what is the command. 23:03.000 --> 23:05.000 What? 23:05.000 --> 23:06.000 Sort? 23:06.000 --> 23:07.000 Oh yeah. 23:07.000 --> 23:09.000 Merci. 23:09.000 --> 23:10.000 Wow. 23:10.000 --> 23:12.000 Merci. 23:12.000 --> 23:13.000 So demo, that's fine. 23:13.000 --> 23:15.000 Also why can we record it next time? 23:15.000 --> 23:16.000 Yeah. 23:20.000 --> 23:21.000 So here it is. 23:21.000 --> 23:23.000 We can see that the new, so the first one, 23:23.000 --> 23:26.000 that one is the one that I just patch and we compile. 23:26.000 --> 23:30.000 It's faster than the ones that we have in three currently. 23:30.000 --> 23:34.000 And it's twice faster than the new implementation. 23:34.000 --> 23:36.000 So what I'm showing is not to show off. 23:36.000 --> 23:38.000 It took me like 30 minutes. 23:38.000 --> 23:39.000 It's not very hard to do. 23:39.000 --> 23:41.000 And the profiler and also new tools. 23:41.000 --> 23:45.000 The rest tool helps a lot to debug and identify 23:45.000 --> 23:47.000 those kind of issues. 23:47.000 --> 23:50.000 And as I was saying, it's tools that we develop for five or four years. 23:50.000 --> 23:53.000 And it is why five or four is now as fast as Chrome. 23:53.000 --> 23:55.000 It's thanks to all those tools. 23:55.000 --> 24:00.000 But I'm also about five or four and go back to the presentation. 24:00.000 --> 24:03.000 So I did the demo. 24:03.000 --> 24:08.000 So we have been held by the Southern Tech Fund in 2023, 24:08.000 --> 24:12.000 with the other maintainer and Google helped us as part of the Google 24:12.000 --> 24:14.000 summer of code 2024. 24:14.000 --> 24:18.000 And as I was saying earlier, we have proposed CI on most of the 24:18.000 --> 24:21.000 supported platform that everything, but most of them 24:21.000 --> 24:26.000 get up is doing a great job to provide the banner in GitHub 24:26.000 --> 24:27.000 action. 24:27.000 --> 24:29.000 Our code coverage is amazing. 24:29.000 --> 24:30.000 Almost 80%. 24:30.000 --> 24:33.000 And if you have played with code coverage, if you are both 24:33.000 --> 24:35.000 A.T., you are usually very happy. 24:35.000 --> 24:36.000 Here we are. 24:36.000 --> 24:39.000 Even happier with almost 90. 24:39.000 --> 24:43.000 We have a lot of jobs in the CI to make sure that if you 24:43.000 --> 24:46.000 enable the adoption ends at one, it's going to work. 24:46.000 --> 24:49.000 If you disable that one, and the other one is going to work. 24:49.000 --> 24:51.000 And we also contribute to have three. 24:51.000 --> 24:52.000 So we report bugs. 24:52.000 --> 24:55.000 We send patches when they are doing something wrong. 24:55.000 --> 24:57.000 We collaborate with them. 24:57.000 --> 24:59.000 It's a very friendly collaboration. 24:59.000 --> 25:03.000 So it's not we are not fighting. 25:03.000 --> 25:05.000 And one of the cool stuff that I play with, 25:05.000 --> 25:07.000 so Atmosier used to manage the fuzzing teams. 25:07.000 --> 25:09.000 So I know one was two things about fuzzing, 25:09.000 --> 25:11.000 and they told me when I discussed with them, 25:11.000 --> 25:14.000 they told me you should fight differential fuzzing. 25:14.000 --> 25:18.000 So what we are doing is we have a fuzzing that is going 25:18.000 --> 25:21.000 the fuzzing, the grammar of the command. 25:21.000 --> 25:25.000 So for example, it's XPR for expression. 25:25.000 --> 25:27.000 So you can do some stupid math. 25:27.000 --> 25:30.000 And it's going to generate something that it is valid, 25:30.000 --> 25:31.000 and sometimes invalid. 25:31.000 --> 25:34.000 And what we are doing is that we are asking the fuzzing to generate 25:34.000 --> 25:37.000 a lot of things that are relevant and irrelevant. 25:37.000 --> 25:40.000 And then we run the command into both implementation. 25:40.000 --> 25:42.000 Howard and the Gnu. 25:42.000 --> 25:44.000 And we verify if S.A.D. out. 25:44.000 --> 25:46.000 So the standard output, S.A.D.R. 25:46.000 --> 25:49.000 and the exit code are similar. 25:49.000 --> 25:53.000 In that case, we can see that the Gnu error message 25:53.000 --> 25:55.000 was not integral argument. 25:55.000 --> 25:59.000 And our world was expected an integral occuren. 25:59.000 --> 26:00.000 So it says the same thing. 26:00.000 --> 26:02.000 But we consider that as a bug, 26:02.000 --> 26:05.000 we want to be really drafting with placement 26:05.000 --> 26:08.000 to the later in terms of error message. 26:08.000 --> 26:11.000 And differential fuzzing for that is amazing. 26:11.000 --> 26:14.000 If you have an implementation and you have a target 26:14.000 --> 26:19.000 that you want to reach, differential fuzzing is really one of the 26:19.000 --> 26:22.000 silver bullets that you can use in security, 26:22.000 --> 26:24.000 but in compliance also. 26:24.000 --> 26:27.000 But they know about the quantity because I'm talking about the past 26:27.000 --> 26:30.000 and because you are all occurs and you all know about 26:30.000 --> 26:32.000 free software and you are all most of you probably are 26:32.000 --> 26:34.000 contributing to open source. 26:34.000 --> 26:36.000 I'd like to tell you about what we have been working on. 26:36.000 --> 26:39.000 So there is an old Chinese proverb that I love. 26:39.000 --> 26:42.000 Let's say that the best time to plan to three was 20 years ago. 26:42.000 --> 26:44.000 And the second time it's now. 26:44.000 --> 26:48.000 And you can apply that in your life with many different ways. 26:48.000 --> 26:52.000 And the corridors have been here for 55 years, 26:52.000 --> 26:54.000 if you think about it. 26:54.000 --> 26:56.000 I have friends who worked as software 26:56.000 --> 26:58.000 and we have that. 26:58.000 --> 27:01.000 And if you can look at the first commit of the corridors 27:01.000 --> 27:03.000 that was done for Unix. 27:03.000 --> 27:08.000 It was June all in four months, five months, 27:08.000 --> 27:10.000 it would be 55 years. 27:11.000 --> 27:13.000 And yesterday when I worked on the slide, 27:13.000 --> 27:15.000 it was like, oh, it's actually a book. 27:15.000 --> 27:17.000 The timestamp a book was six months before. 27:17.000 --> 27:18.000 So it's not a coincidence. 27:18.000 --> 27:22.000 And those things are really back from the beginning of our 27:22.000 --> 27:23.000 industry. 27:23.000 --> 27:25.000 And I was thinking about it. 27:25.000 --> 27:27.000 It's like how the corridors can disappear. 27:27.000 --> 27:30.000 Like in which world are we going to get rid of L.S. 27:30.000 --> 27:34.000 M.C.R.C.H.Mertiation on Artist on Mac and on Linux. 27:34.000 --> 27:35.000 Probably not. 27:35.000 --> 27:38.000 I think we are all going to be retired when those tools are going 27:38.000 --> 27:41.000 to disappear if they disappear one day. 27:41.000 --> 27:43.000 And those tools still continue to evolve. 27:43.000 --> 27:45.000 Like if you look at the new projects, 27:45.000 --> 27:47.000 the new implementations are taking your options, 27:47.000 --> 27:49.000 they introduce new algorithms, and we are doing the same. 27:49.000 --> 27:52.000 And there is no reason why those tools should not evolve. 27:52.000 --> 27:54.000 At some point we should have the discussion. 27:54.000 --> 27:56.000 Does PR still make sense to have that in three, 27:56.000 --> 27:58.000 probably not, but it's a different discussion. 27:58.000 --> 28:00.000 And by the way, it's not a Chinese proverb. 28:00.000 --> 28:02.000 I looked at that to people in my presentation, 28:02.000 --> 28:03.000 but it looks cool. 28:03.000 --> 28:05.000 So that's why I use it. 28:06.000 --> 28:09.000 There is a page explaining that it is not a Chinese proverb. 28:09.000 --> 28:12.000 It forms the U.S. in the 50s or something like that. 28:12.000 --> 28:14.000 But I like it. 28:14.000 --> 28:17.000 In parallel, we have been working on fine materials and different tools. 28:17.000 --> 28:20.000 So those projects have been less of a focus of us. 28:20.000 --> 28:22.000 But we have people who did a lot of great work. 28:22.000 --> 28:26.000 Here, we are running the B.F.B.F.S. 28:26.000 --> 28:30.000 Test suite against our implementation of the fine materials. 28:30.000 --> 28:32.000 The comments that everybody knows. 28:32.000 --> 28:39.000 And you can see that we are not fully compliant, but good in us. 28:39.000 --> 28:44.000 And now what I like to do is to do the same with most of the essential packages on the 28:44.000 --> 28:47.000 Debian and Ubuntu, because it's the distros that I know best. 28:47.000 --> 28:50.000 We can ask the similar conversation with other distros. 28:50.000 --> 28:53.000 Let's hope you love distros and Debian, but seriously. 28:53.000 --> 28:57.000 And there is no better time than now to start that work. 28:57.000 --> 29:00.000 So the essential packages, you use them every day. 29:00.000 --> 29:02.000 You know about them, not all of them, obviously. 29:02.000 --> 29:04.000 But they are still very important. 29:04.000 --> 29:08.000 So this is a comment on Debian or Ubuntu to get the list of essential packages. 29:08.000 --> 29:11.000 The essential packages, if you want to create a meeting. 29:16.000 --> 29:18.000 I don't do Debian anymore. 29:18.000 --> 29:19.000 Oh, yes. 29:19.000 --> 29:21.000 It's more than one that's going to be there. 29:21.000 --> 29:25.000 When you create a CH route and Debian or you boot a very tiny Debian, 29:25.000 --> 29:29.000 it is a package that Debian needs to run. 29:29.000 --> 29:31.000 So you all know about those ones. 29:31.000 --> 29:34.000 They are easy, free, key, LPS.watch. 29:34.000 --> 29:38.000 In this package, prompt PS, we have also some weird stuff. 29:38.000 --> 29:39.000 Like PWX. 29:39.000 --> 29:42.000 Who knows about PWDX is doing? 29:42.000 --> 29:45.000 I implemented it in a few months ago. 29:45.000 --> 29:50.000 It's going to tell you the current working directory of a PIG. 29:50.000 --> 29:53.000 You want something at least today at first time. 29:54.000 --> 29:57.000 T-load, slapped up, and so on. 29:57.000 --> 29:58.000 Util Linux. 29:58.000 --> 30:01.000 So you know about DMSLs, get up, our link, and so on. 30:01.000 --> 30:02.000 There are some stuff. 30:02.000 --> 30:03.000 Those ones don't ask me. 30:03.000 --> 30:04.000 I don't know what they are. 30:04.000 --> 30:05.000 I didn't pre-per for those ones. 30:05.000 --> 30:08.000 I'm cookie and CH-woo. 30:08.000 --> 30:10.000 I guess CH-woo, I guess CH-woo, yes. 30:10.000 --> 30:11.000 CH-woo, I don't remember. 30:11.000 --> 30:15.000 It could change the behavior of the kernel when you are out of memory. 30:15.000 --> 30:20.000 Then we have some comments that we don't use in Linux in real life. 30:20.000 --> 30:24.000 For some reason, I should discuss with people in Debion Y. 30:24.000 --> 30:27.000 It is installed by default ACR. 30:27.000 --> 30:31.000 It's one of the cool stuff that we don't use at the end. 30:31.000 --> 30:33.000 We have also BSD utils. 30:33.000 --> 30:36.000 So there are some most famous green ice to change the priority. 30:36.000 --> 30:39.000 We have also Austin. 30:39.000 --> 30:41.000 Login. 30:41.000 --> 30:45.000 And you didn't see system D because I'm afraid of lernart. 30:45.000 --> 30:46.000 I don't know if it's in the room. 30:46.000 --> 30:49.000 I don't want to see them before. 30:49.000 --> 30:52.000 More seriously, system D would be a beast to be right in rust. 30:52.000 --> 30:55.000 And it would be way too complex. 30:55.000 --> 30:58.000 We already started all those projects. 30:58.000 --> 31:00.000 They are already live on our project. 31:00.000 --> 31:01.000 It's called U-tills. 31:01.000 --> 31:03.000 I think it's universal utils. 31:03.000 --> 31:04.000 It doesn't matter what it means. 31:04.000 --> 31:06.000 It is a name of the project. 31:06.000 --> 31:10.000 We have already a lot of people who started contributing to those projects. 31:10.000 --> 31:14.000 We also leveraging what we have done for the core materials. 31:14.000 --> 31:18.000 So there are many, many functions that we have in the core materials. 31:18.000 --> 31:23.000 That can be used for the programs as part of that world. 31:23.000 --> 31:28.000 So when you need to monitor the system or when you need to look at the permission, 31:28.000 --> 31:31.000 we have already also function. 31:31.000 --> 31:34.000 We started some of some of the scum end. 31:34.000 --> 31:36.000 I've been fully implemented. 31:36.000 --> 31:38.000 Some of them are just one line. 31:38.000 --> 31:40.000 It really depends. 31:40.000 --> 31:43.000 And it's very easy to plug new tools into that ecosystem. 31:43.000 --> 31:45.000 And we have a lot of good feedback. 31:45.000 --> 31:47.000 So if you want to learn the rest, it's very easy. 31:47.000 --> 31:48.000 It seems hard. 31:48.000 --> 31:50.000 Those comments are usually self-contained. 31:50.000 --> 31:51.000 You know the target. 31:51.000 --> 31:53.000 We have a lot of thinking fruit. 31:53.000 --> 31:56.000 So don't hesitate if you want to learn the rest for real. 31:56.000 --> 31:58.000 To start those projects is where I learn. 31:58.000 --> 32:01.000 And that's why we have so many contributors. 32:01.000 --> 32:05.000 I think those projects, like we're writing, 32:05.000 --> 32:07.000 we inventing the will, might some crazy. 32:07.000 --> 32:10.000 But I think it will work because there is an appetite from the community. 32:10.000 --> 32:13.000 We saw that the room here is the tool. 32:13.000 --> 32:15.000 But there is a lot of people. 32:15.000 --> 32:18.000 The talk from Miguel before was the same. 32:18.000 --> 32:19.000 A lot of people. 32:19.000 --> 32:22.000 So we see that people are excited by such projects. 32:22.000 --> 32:24.000 But I was seeing earlier. 32:24.000 --> 32:25.000 I'm beginning older. 32:25.000 --> 32:26.000 We all do. 32:26.000 --> 32:29.000 But the new generation is not going to want to do see. 32:29.000 --> 32:36.000 And paving the way to do rest is also a good opportunity for them to be involved in the open source ecosystem. 32:36.000 --> 32:38.000 We have reference implementation. 32:38.000 --> 32:39.000 So it's very easy to test. 32:39.000 --> 32:42.000 You can do different short-fazing as a silver bullet to help you. 32:42.000 --> 32:45.000 And as part of that, you learn how to do fazing. 32:45.000 --> 32:49.000 And in the rest is very easy to do. 32:49.000 --> 32:51.000 And we are also. 32:51.000 --> 32:54.000 I didn't know if I wanted to say that during that presentation. 32:54.000 --> 32:58.000 But I'm going to say it is a lot of those tools that I mentioned earlier. 32:58.000 --> 33:01.000 They are using very old pipeline system. 33:01.000 --> 33:02.000 Some of them. 33:02.000 --> 33:03.000 They don't have CI. 33:03.000 --> 33:06.000 They are still using maintenance to send patches. 33:06.000 --> 33:08.000 I apologize for the news developer here. 33:08.000 --> 33:13.000 I still think that using maintenance to do patch review sucks. 33:13.000 --> 33:19.000 And now with GitHub, we have an amazing opportunity to have a proper pipeline where you 33:19.000 --> 33:20.000 can run CI. 33:20.000 --> 33:22.000 You can run quality tools and so on. 33:22.000 --> 33:27.000 And it is one of the advantage that we have over some of those old projects that I mentioned. 33:27.000 --> 33:30.000 But to be clear, that works for them so I'm happy. 33:30.000 --> 33:34.000 But I think we can do better as a community. 33:34.000 --> 33:37.000 And I'd like to mention that it's not a much easier project. 33:37.000 --> 33:38.000 There is no company behind it. 33:38.000 --> 33:42.000 There is no interest from the architect in that project. 33:42.000 --> 33:45.000 It's really I'm doing that as a passion because I can't do it. 33:45.000 --> 33:47.000 They are in the new extension of all. 33:47.000 --> 33:49.000 So it's really a community effort. 33:49.000 --> 33:51.000 There is no structure behind it. 33:51.000 --> 33:56.000 Now what I'd like to do is to finish that demo. 33:56.000 --> 33:57.000 Sorry. 33:57.000 --> 34:00.000 By making. 34:00.000 --> 34:03.000 So it's a blog post that I wrote to announce that stuff. 34:03.000 --> 34:04.000 I wrote it a few months ago. 34:04.000 --> 34:07.000 But now I think it's time to make it public. 34:07.000 --> 34:09.000 So make. 34:09.000 --> 34:12.000 Project. 34:12.000 --> 34:13.000 Click. 34:13.000 --> 34:17.000 So here it is. 34:17.000 --> 34:20.000 So it's probably going to be on account news in a few hours. 34:20.000 --> 34:22.000 And last time it was number one. 34:22.000 --> 34:24.000 It's probably going to be number one again. 34:24.000 --> 34:26.000 I know that I'm showing off by saying that. 34:26.000 --> 34:30.000 But it shows the interest of the community when we do it. 34:30.000 --> 34:34.000 It's a blog post. 34:34.000 --> 34:36.000 I don't know how quickly it is going. 34:36.000 --> 34:37.000 Oh, no blog post. 34:37.000 --> 34:39.000 So I guess there are some cases somewhere. 34:39.000 --> 34:42.000 So we have to wait for the cases to be refreshed. 34:42.000 --> 34:43.000 One more time. 34:43.000 --> 34:46.000 It's a risk of the demo effect. 34:46.000 --> 34:47.000 Anyway, thanks for your time. 34:47.000 --> 34:49.000 Don't hesitate if you have any questions. 34:49.000 --> 34:51.000 It's very easy to contribute. 34:51.000 --> 34:54.000 There are many people who are helping. 34:54.000 --> 34:56.000 We are using a simple PR model. 34:56.000 --> 34:57.000 We have plenty of tests. 34:57.000 --> 34:59.000 Don't be afraid to make mistakes. 35:00.000 --> 35:01.000 Thank you. 35:07.000 --> 35:09.000 Many thanks for the inspiring talk. 35:09.000 --> 35:10.000 Perfect. 35:10.000 --> 35:11.000 Great talk. 35:11.000 --> 35:12.000 Thank you. 35:12.000 --> 35:13.000 Great pleasure. 35:13.000 --> 35:15.000 Question to the audience. 35:15.000 --> 35:21.000 If you meet to leave early, please leave to those for doors. 35:21.000 --> 35:28.000 Not these ones because we'd like to have the Q&A to be understandable. 35:28.000 --> 35:31.000 Who's the first who's got any questions? 35:31.000 --> 35:32.000 Okay. 35:32.000 --> 35:33.000 Over there. 35:33.000 --> 35:36.000 I'm coming. 35:49.000 --> 35:55.000 One of the things that Gnucho co-utives has. 35:55.000 --> 36:00.000 Do you hear me, Sylvestre? 36:00.000 --> 36:01.000 Really? 36:01.000 --> 36:02.000 Okay. 36:02.000 --> 36:09.000 One of the things that you have with Gnucho co-utives is extreme 36:09.000 --> 36:10.000 Portability. 36:10.000 --> 36:13.000 It works on many, many platforms. 36:13.000 --> 36:19.000 As far as I understand, you leverage the REST C portability to be able to use 36:19.000 --> 36:21.000 REST co-utives everywhere. 36:21.000 --> 36:24.000 Are you tracking other REST competitors? 36:24.000 --> 36:27.000 As well as to extend your coverage? 36:27.000 --> 36:30.000 We will when they are ready for now. 36:30.000 --> 36:33.000 So Nikola is asking. 36:33.000 --> 36:35.000 Can you hear me? 36:35.000 --> 36:36.000 Yeah. 36:36.000 --> 36:39.000 Nikola is asking the question about keeping track of the other 36:39.000 --> 36:40.000 Compiler for us. 36:40.000 --> 36:43.000 As far as I know, they are not there yet. 36:43.000 --> 36:44.000 But we will. 36:44.000 --> 36:48.000 Like it's just adding a few lines in a thermal file. 36:48.000 --> 36:49.000 So yeah. 36:49.000 --> 36:52.000 We will test everywhere that we can. 37:01.000 --> 37:04.000 Any plans to do a shell like zush? 37:04.000 --> 37:05.000 Sorry? 37:05.000 --> 37:07.000 Any plans to do a shell? 37:07.000 --> 37:08.000 No. 37:08.000 --> 37:11.000 I want to replace the existing in C, not to do a shell. 37:11.000 --> 37:13.000 There are also people doing. 37:13.000 --> 37:17.000 There is no standard with a shell, 37:17.000 --> 37:20.000 except using what has been done and what has been done. 37:20.000 --> 37:23.000 The folks with a liquidity have been doing amazing work. 37:23.000 --> 37:25.000 So there is no need in this place. 37:25.000 --> 37:45.000 Hi. 37:45.000 --> 37:48.000 So first of all, thanks for the talk. 37:48.000 --> 37:57.000 And what is your stance on the other utils that have that are a part of the rewrite 37:57.000 --> 38:02.000 it in rust, kind of trend? 38:02.000 --> 38:10.000 What is your stance on the other utils that have that are a part of the rewrite it in rust, 38:10.000 --> 38:19.000 kind of trend, like ripgrap, dust, bat, and all of that already exist. 38:19.000 --> 38:22.000 And are kind of alternatives to those? 38:22.000 --> 38:23.000 Yeah. 38:23.000 --> 38:25.000 I think they are providing a lot of value. 38:25.000 --> 38:29.000 Like ripgrap or FD fine and so on. 38:29.000 --> 38:33.000 However, the issue is that they are not built in the placement. 38:33.000 --> 38:38.000 So it's not very easy to adjust your practice. 38:38.000 --> 38:40.000 So it's a very good project. 38:40.000 --> 38:45.000 And the advantage that they have over us is that they don't have to maintain the argument, 38:45.000 --> 38:48.000 legacy, and all the issues and the ordering. 38:48.000 --> 38:50.000 Those are the self-imaging project. 38:50.000 --> 38:53.000 I use the readgrap everyday. 38:53.000 --> 38:54.000 Thank you. 38:58.000 --> 38:59.000 Okay. 38:59.000 --> 39:02.000 Other questions? 39:03.000 --> 39:07.000 I can. 39:07.000 --> 39:12.000 Oh, over there. 39:21.000 --> 39:24.000 Yeah, thanks for the interesting talk. 39:24.000 --> 39:28.000 And you also mentioned that you have this already. 39:28.000 --> 39:33.000 The rust core utils in package for most Linux distributions. 39:33.000 --> 39:36.000 And I have myself worked for Debian and there. 39:36.000 --> 39:44.000 It's really, really hard to package rust packages because of these strict dependency 39:44.000 --> 39:45.000 management. 39:45.000 --> 39:52.000 So you have to often patch each and every dependency to have it with a less strict version 39:52.000 --> 39:53.000 specification. 39:53.000 --> 39:58.000 What is your experience there and how do you want to work around that in the future, 39:58.000 --> 40:03.000 given that probably only more and more crates are added to the library? 40:03.000 --> 40:08.000 So if I understood correctly, you are asking about packaging when it is hard to package rust 40:08.000 --> 40:09.000 software. 40:09.000 --> 40:10.000 Yeah, it's a dependency. 40:10.000 --> 40:12.000 Yeah, it's a big deal. 40:12.000 --> 40:15.000 It's because of the ecosystem in the rust. 40:15.000 --> 40:19.000 It's quite hard to package, but I think it's going to stabilize. 40:19.000 --> 40:23.000 And that more and more some crates have been doing a lot for years ago. 40:23.000 --> 40:25.000 And now they are starting to be stable. 40:25.000 --> 40:30.000 And upstream is making is more careful about the API and API changes. 40:30.000 --> 40:35.000 So I think we are going to grind the right direction at the next time. 40:35.000 --> 40:46.000 Yeah, I agree that it is a pain to implement in rust packages at that time. 40:46.000 --> 40:48.000 Thank you. 40:48.000 --> 40:50.000 Apparently no more questions. 40:50.000 --> 40:57.000 So many thanks once more and please a applause. 40:57.000 --> 40:58.000 Thank you.