WEBVTT 00:00.000 --> 00:15.000 I can ask how many of you use Jenkins. 00:15.000 --> 00:20.000 I want to show you use Jitabla. 00:20.000 --> 00:21.000 Okay. 00:21.000 --> 00:22.000 Okay. 00:22.000 --> 00:29.000 I don't ask for Jitabla because everything is died now with Jitabla. 00:29.000 --> 00:30.000 Okay. 00:30.000 --> 00:31.000 Okay. 00:31.000 --> 00:32.000 Okay. 00:32.000 --> 00:34.000 Who uses Gerrit? 00:34.000 --> 00:35.000 Yeah. 00:35.000 --> 00:37.000 You use Gerrit. 00:37.000 --> 00:38.000 Yes. 00:38.000 --> 00:39.000 Okay. 00:39.000 --> 00:41.000 Why in the past? 00:41.000 --> 00:43.000 In the past. 00:43.000 --> 00:45.000 Okay. 00:45.000 --> 00:46.000 Yeah. 00:46.000 --> 00:47.000 I can understand. 00:47.000 --> 00:51.000 But let's say different proposed Gerrit and Jitabla. 00:51.000 --> 00:54.000 Different way to regue and work with the code. 00:54.000 --> 01:01.000 I'm not a fan of Jitabla, but just because my model of developing working is Gerrit Orient. 01:01.000 --> 01:02.000 So... 01:14.000 --> 01:15.000 Okay. 01:15.000 --> 01:16.000 Anyway. 01:16.000 --> 01:19.000 The big difference between Jitabla and Jitabla. 01:19.000 --> 01:23.000 And Jenkins is like that you take appeals every day. 01:23.000 --> 01:25.000 And you pay for the pills. 01:25.000 --> 01:27.000 And this is Jitabla. 01:27.000 --> 01:30.000 And you find until the pills kill you. 01:30.000 --> 01:32.000 Or you go to a surgery. 01:32.000 --> 01:33.000 Okay. 01:33.000 --> 01:38.000 And now you can go to a surgery for eight hours. 01:38.000 --> 01:40.000 And after that you don't need to take the pills. 01:40.000 --> 01:42.000 If you survive. 01:42.000 --> 01:43.000 Okay. 01:43.000 --> 01:45.000 So this is a big difference. 01:45.000 --> 01:46.000 Okay. 01:46.000 --> 01:50.000 I try to get the surgery operations to have fun. 01:50.000 --> 01:51.000 You know. 01:51.000 --> 01:53.000 It's much more fun to get the surgery for eight hours. 01:53.000 --> 01:56.000 And then forget about your problem. 01:56.000 --> 01:57.000 Okay. 01:57.000 --> 01:58.000 I'm Michael. 01:58.000 --> 02:00.000 I'm working on a number of solutions. 02:00.000 --> 02:02.000 I'm not really a CICB guy. 02:02.000 --> 02:10.000 So I start 15 years ago to do something CICB just because the company needs some infrastructures. 02:10.000 --> 02:11.000 And yeah. 02:11.000 --> 02:13.000 Now what's coming with AI. 02:13.000 --> 02:15.000 I've been testing this on. 02:15.000 --> 02:18.000 So my main focus was embedded. 02:18.000 --> 02:20.000 It's still embedded. 02:20.000 --> 02:21.000 But in the weekend. 02:21.000 --> 02:23.000 A lot of weekends. 02:23.000 --> 02:26.000 Let's say today's every weekend. 02:26.000 --> 02:28.000 I spend some time in Jenkins. 02:28.000 --> 02:31.000 Now I do some plug-in contributor. 02:31.000 --> 02:34.000 And of course I even contribute for wanting energy. 02:34.000 --> 02:36.000 Plug-in for CV. 02:36.000 --> 02:38.000 Extraction from a doctor for instance. 02:38.000 --> 02:42.000 It's just to play a bit to see how the model was working in the wanting energy. 02:42.000 --> 02:43.000 Yeah. 02:43.000 --> 02:45.000 I find this communion is really fun. 02:45.000 --> 02:47.000 The Jenkins one. 02:47.000 --> 02:51.000 This is from my university and say that basically if you want to be an engineering, 02:51.000 --> 02:54.000 say you need to be prepared to go around the world. 02:54.000 --> 02:57.000 And this of course, those things are improving now. 02:57.000 --> 02:58.000 I go with the bicycle. 02:58.000 --> 03:01.000 So not anymore like this. 03:01.000 --> 03:02.000 So now I move like this. 03:02.000 --> 03:04.000 So it's quite fun. 03:04.000 --> 03:06.000 Because I'm a CICB guy. 03:06.000 --> 03:07.000 It's not needed. 03:07.000 --> 03:08.000 Let's say, put it this way. 03:08.000 --> 03:10.000 It's better to, you know. 03:11.000 --> 03:15.000 Every, you know what is contributing to patients. 03:15.000 --> 03:18.000 Seems that is a kind of topic that everyone knows. 03:18.000 --> 03:22.000 But in my experience, but maybe you're lacking more than me. 03:22.000 --> 03:26.000 The fact that in a lot of companies that, 03:26.000 --> 03:29.000 if they have already a good release process, 03:29.000 --> 03:31.000 is already a good thing. 03:31.000 --> 03:34.000 But you know, not everything is moving so fast. 03:34.000 --> 03:37.000 And you start up and you company that I've seen appearing 03:38.000 --> 03:40.000 Yes, they use extensively. 03:40.000 --> 03:41.000 Contributions. 03:41.000 --> 03:42.000 Contributions. 03:42.000 --> 03:43.000 Paraguay and so on. 03:43.000 --> 03:45.000 Why we use Garrett because you know, 03:45.000 --> 03:50.000 Jenkins is not really a system that you do contributing. 03:50.000 --> 03:52.000 You interact with the user. 03:52.000 --> 03:53.000 You do. 03:53.000 --> 03:54.000 You know, you cannot, 03:54.000 --> 03:59.000 Jenkins is only a way to our fantastic way to trigger your business. 03:59.000 --> 04:01.000 Or to execute your jobs. 04:01.000 --> 04:04.000 Let's say that Jenkins can execute any jobs. 04:04.000 --> 04:08.000 It can fix any kind of infrastructure problems. 04:08.000 --> 04:12.000 And it does not depend on any locking technologies. 04:12.000 --> 04:15.000 It's Apache licensing. 04:15.000 --> 04:19.000 I think that's under the end now on the continuation of foundation. 04:19.000 --> 04:21.000 So it's quite nice. 04:21.000 --> 04:27.000 But you can figure out an image with your fantasy. 04:27.000 --> 04:30.000 Any kind of things that you want to have in your infrastructure. 04:30.000 --> 04:32.000 And you need to ask if they succeed. 04:33.000 --> 04:34.000 It doesn't exist. 04:34.000 --> 04:36.000 It's also open source developer. 04:36.000 --> 04:37.000 You can just create. 04:37.000 --> 04:40.000 If it's created, you can just improve. 04:40.000 --> 04:44.000 So that is the way why I mean this. 04:44.000 --> 04:46.000 Yeah, the goal is, you know, 04:46.000 --> 04:47.000 is the solution. 04:47.000 --> 04:49.000 We want to define bugs. 04:49.000 --> 04:51.000 We want to our cycle of developers. 04:51.000 --> 04:53.000 We will finance on. 04:53.000 --> 04:54.000 This is the tool that I use. 04:54.000 --> 04:56.000 I can you can see parts on a cube. 04:56.000 --> 04:58.000 But that is open source. 04:58.000 --> 05:01.000 I mean, there is part of a son of a cube that is open source. 05:01.000 --> 05:06.000 All the other things are completely, you know, open source tools. 05:06.000 --> 05:08.000 First of all, the analysis. 05:08.000 --> 05:10.000 First of all, the code checker. 05:10.000 --> 05:11.000 Gerrit. 05:11.000 --> 05:12.000 Explain. 05:12.000 --> 05:13.000 Plugin. 05:13.000 --> 05:14.000 With AI. 05:14.000 --> 05:15.000 Jenkins. 05:15.000 --> 05:16.000 Gerrit. 05:16.000 --> 05:17.000 I plug in. 05:17.000 --> 05:18.000 Sonar cube. 05:18.000 --> 05:19.000 Martin most. 05:19.000 --> 05:20.000 I think so. 05:20.000 --> 05:21.000 It's almost open source. 05:21.000 --> 05:23.000 So yeah, it's fine. 05:23.000 --> 05:25.000 You can have everything that you have. 05:25.000 --> 05:27.000 Consider the other ecosystem. 05:27.000 --> 05:31.000 But what is maybe it's a bit difficult to start with it. 05:31.000 --> 05:33.000 So there is not an easy. 05:33.000 --> 05:38.000 You don't create a repository in GitHub or and you have everything's there. 05:38.000 --> 05:40.000 And you have your GitHub actions. 05:40.000 --> 05:42.000 You need to spend some times. 05:42.000 --> 05:46.000 But when you spend, you have spend some times. 05:46.000 --> 05:47.000 You have Jenkins. 05:47.000 --> 05:50.000 So we have completely control of what you are doing. 05:50.000 --> 05:51.000 You have continually. 05:51.000 --> 05:53.000 You can continue to improve your pipeline. 05:53.000 --> 05:58.000 You are not limited to any things because you have any plugins that you want. 05:58.000 --> 06:01.000 If you don't have you create it, you have administration. 06:01.000 --> 06:03.000 You can maintain your plugin. 06:03.000 --> 06:06.000 You can decide to store your activity in a stream. 06:06.000 --> 06:07.000 Whatever. 06:07.000 --> 06:09.000 You can have your runners inside your company. 06:09.000 --> 06:12.000 You can have runner in the clouds. 06:12.000 --> 06:13.000 You can have runner. 06:13.000 --> 06:15.000 Connect to the SSH. 06:15.000 --> 06:16.000 You can have runner. 06:16.000 --> 06:17.000 Connect to web socket. 06:17.000 --> 06:19.000 You can run your windows. 06:19.000 --> 06:20.000 In Apple. 06:20.000 --> 06:21.000 Yeah. 06:22.000 --> 06:23.000 So what Jenkins offer? 06:23.000 --> 06:24.000 Okay. 06:24.000 --> 06:26.000 For our my type of ecosystem. 06:26.000 --> 06:30.000 We need to a way to trigger job with inger it and Jenkins. 06:30.000 --> 06:32.000 I use personal use jerry trigger. 06:32.000 --> 06:35.000 There is an implementation for jerry forge that you use. 06:35.000 --> 06:37.000 Another paradigm to do trigger. 06:37.000 --> 06:38.000 It's okay. 06:38.000 --> 06:39.000 It's fine. 06:39.000 --> 06:45.000 The problem that jerry trigger that I hope that I'm going to solve this year is to create a way to use 06:45.000 --> 06:48.000 HTTPS for triggering event. 06:48.000 --> 06:54.000 Much more fluidly and not all the SSH connections on it to get rid of this. 06:54.000 --> 06:55.000 But yeah. 06:55.000 --> 06:56.000 It can trigger. 06:56.000 --> 06:57.000 Okay. 06:57.000 --> 07:00.000 When I trigger in GitHub, you have the concept. 07:00.000 --> 07:02.000 The wrong concept of pull request. 07:02.000 --> 07:06.000 The worst things that the people create is the pull request. 07:06.000 --> 07:09.000 Some people that create 20 patches. 07:09.000 --> 07:14.000 And people that are only the pull request as a huge giant piece of code. 07:14.000 --> 07:18.000 And that's not understand what is in each part of the code. 07:18.000 --> 07:20.000 What is this patch is needed. 07:20.000 --> 07:22.000 It's clean up. 07:22.000 --> 07:23.000 Yeah. 07:23.000 --> 07:24.000 Always get a long things. 07:24.000 --> 07:25.000 They just said change change. 07:25.000 --> 07:26.000 And this one day. 07:26.000 --> 07:28.000 Just add a new commit on top of it. 07:28.000 --> 07:29.000 Okay. 07:29.000 --> 07:30.000 It's not really in jerry. 07:30.000 --> 07:31.000 It's committed oriented. 07:31.000 --> 07:33.000 It means that we are going to review each commit. 07:33.000 --> 07:36.000 We can trigger builds that. 07:36.000 --> 07:39.000 Very much each commit on your pull request. 07:39.000 --> 07:40.000 Call it pull request. 07:41.000 --> 07:44.000 And because you get it, we have the concept of topic. 07:44.000 --> 07:49.000 That is a concrete number of commits that implement a functionalities. 07:49.000 --> 07:53.000 That makes to have the same functionalities in CDI as a pull request. 07:53.000 --> 07:56.000 So you can build the entire topics as once. 07:56.000 --> 07:57.000 Verification build. 07:57.000 --> 07:59.000 Or you can build single things. 07:59.000 --> 08:02.000 So we're really quite interesting and powerful. 08:02.000 --> 08:04.000 And there is a lot of reason that we need. 08:04.000 --> 08:08.000 You know, our field to build each commit because we want. 08:08.000 --> 08:10.000 Our code to be be acceptable. 08:10.000 --> 08:12.000 And we want our code to be well-documented. 08:12.000 --> 08:14.000 So get it help. 08:14.000 --> 08:16.000 Improve this or enforce this to developers. 08:16.000 --> 08:18.000 Then okay, artifact manager in Jenkins. 08:18.000 --> 08:20.000 You can use any artifact manager. 08:20.000 --> 08:22.000 You can connect to S3. 08:22.000 --> 08:24.000 Personal use S3. 08:24.000 --> 08:25.000 Then we have local resources. 08:25.000 --> 08:27.000 You know, sometimes you need to contents between jobs. 08:27.000 --> 08:29.000 So you have limited hardware resources. 08:29.000 --> 08:30.000 You want to. 08:30.000 --> 08:32.000 Then we have what do you have. 08:32.000 --> 08:33.000 The other tools. 08:33.000 --> 08:36.000 One energy that permit me to have any kind of analysis. 08:36.000 --> 08:38.000 This is static analysis collection of problems. 08:38.000 --> 08:40.000 Put on the thread page. 08:40.000 --> 08:41.000 And so on. 08:41.000 --> 08:44.000 Now we have recently worked with the guys that from China. 08:44.000 --> 08:46.000 They explain a role plugin. 08:46.000 --> 08:47.000 They were quite nice. 08:47.000 --> 08:50.000 Was introduced to use AI to interpret it with result. 08:50.000 --> 08:51.000 It's for free. 08:51.000 --> 08:52.000 I mean for free. 08:52.000 --> 08:53.000 If you take the Chinese. 08:53.000 --> 08:56.000 Chinese AI is basically for free. 08:56.000 --> 09:00.000 If you want to spend more money, you can take the other AI that you post. 09:00.000 --> 09:01.000 Very expensive. 09:01.000 --> 09:02.000 Renamex. 09:02.000 --> 09:03.000 Because it's a log build. 09:03.000 --> 09:04.000 I try to figure out. 09:04.000 --> 09:06.000 How does Chinese AI works? 09:06.000 --> 09:08.000 And there was pretty good. 09:08.000 --> 09:11.000 Yeah. 09:11.000 --> 09:13.000 That's why we're adopting. 09:13.000 --> 09:14.000 Yeah. 09:14.000 --> 09:16.000 Let's try to improve things. 09:16.000 --> 09:19.000 Okay. You know, the things I try to just go fast here. 09:19.000 --> 09:23.000 Because I want to give some example how the things I use it. 09:23.000 --> 09:29.000 So code review for us is basically is a basic steps for improve code quality. 09:29.000 --> 09:32.000 I know that now people tends to generate code by AI. 09:32.000 --> 09:33.000 Let the code review by AI. 09:33.000 --> 09:37.000 Let the AI inject the fixed inside the code. 09:37.000 --> 09:40.000 So the role of the developer decides to mix. 09:40.000 --> 09:45.000 Start to be not, but I want my in the company that people progress and 09:45.000 --> 09:48.000 have an overview of the code should be done. 09:48.000 --> 09:54.000 So what we're using AI is how the AI can improve the code review faces. 09:54.000 --> 09:57.000 How the AI can down a bit. 09:57.000 --> 10:02.000 The promise, the entry promise that a design can have when you push some code. 10:02.000 --> 10:09.000 So, but the AI code review is, we start it to understand what can be the 10:09.000 --> 10:10.000 I code review. 10:10.000 --> 10:12.000 We implement a plugin. 10:12.000 --> 10:15.000 We fork a plugin from a community, right? 10:15.000 --> 10:17.000 Another Chinese guy that started. 10:17.000 --> 10:21.000 And then we just recently changed in review. 10:21.000 --> 10:22.000 I plugin is an aim. 10:22.000 --> 10:24.000 It's not the official one in Garrett. 10:24.000 --> 10:26.000 Because we have a large discussion. 10:26.000 --> 10:28.000 And forget about it. 10:28.000 --> 10:35.000 So they just have, thank you missing some very old version that we provide. 10:35.000 --> 10:41.000 So now we integrate the length change, like Java for change to support multiple 10:41.000 --> 10:42.000 Becend. 10:42.000 --> 10:50.000 The same concept I implement in explainer plugin to allow us to use any AI that we want. 10:50.000 --> 10:52.000 So the concept is very basic. 10:52.000 --> 10:56.000 The comments that you can inject inside the, you know, the Garrett interfaces. 10:56.000 --> 11:02.000 You can point it to the I to ask to review some part of code or to re-review the code. 11:02.000 --> 11:06.000 The review of the code can start, let's say, automatic. 11:06.000 --> 11:12.000 Every time that you push a change or you can be the CI that push a curve 11:12.000 --> 11:17.000 Request on a post over the Garrett to say, okay, now review because the test will pass. 11:17.000 --> 11:30.000 So the problem of the I, of course, is the I is like, it's like, you know, someone that in your apartment, you don't like him. 11:30.000 --> 11:34.000 So it's someone that you find any kind of error in what you are doing. 11:34.000 --> 11:40.000 I was cute at, you don't wish to wash, you know, the washing machine, you forget the door open. 11:40.000 --> 11:41.000 This is the I. 11:41.000 --> 11:43.000 It's always fine something. 11:43.000 --> 11:46.000 I don't understand why it's possible, but okay. 11:46.000 --> 11:57.000 So what we try to do in the company said, to try to make make sense of, let's say, say to the I, okay, this is something that we want to do. 11:57.000 --> 12:05.000 And try to accept what is really needed because we never make, seems that it's difficult to make a repeat. 12:05.000 --> 12:11.000 And but because our I is able to vote, is able to provide some things. 12:11.000 --> 12:14.000 And so it's very annoying. 12:14.000 --> 12:23.000 And I don't try to make the I able to vote minus two that the Garrett means you can not be able to push. 12:23.000 --> 12:32.000 But the, the fact that the I can, cannot all review the code, that is one thing. 12:32.000 --> 12:37.000 But in my opinion, if we review even your commit message over the code. 12:37.000 --> 12:41.000 So it's what you describe, it's corresponding what you wrote. 12:41.000 --> 12:46.000 Because it's not for me to review like I was in a company corporate. 12:46.000 --> 12:50.000 And one guy sent me a batches of 2000 lights of code. 12:50.000 --> 12:53.000 I said, smell something strange. 12:53.000 --> 12:54.000 I do different. 12:54.000 --> 12:57.000 We used W and it was 50 lights. 12:57.000 --> 13:00.000 You would just miss up all the identical everywhere. 13:00.000 --> 13:04.000 So I said, okay, just let it push where the four lights so code are should be there. 13:04.000 --> 13:09.000 So it's very important that the, the commit message, the what you present is consistent with the commit message. 13:09.000 --> 13:15.000 And this is a can be done easily by the I because the I found that your commit message not descriptive. 13:15.000 --> 13:20.000 Your commit message, this, some part that you need to add and so on. 13:20.000 --> 13:24.000 And we started to, you know, to continue this plugin. 13:24.000 --> 13:27.000 We implement a lot of things. 13:27.000 --> 13:32.000 So basically what we try to do, it's to create certain interactions. 13:32.000 --> 13:34.000 Even we create a way to devugate. 13:34.000 --> 13:37.000 So we have we can dynamically change the prompt. 13:37.000 --> 13:43.000 We can dynamically change some as some parameters of the I because it's, we need to experiment. 13:43.000 --> 13:44.000 It's in the beginning. 13:44.000 --> 13:48.000 We are not able to figure out what, what will be needed. 13:48.000 --> 13:53.000 So at now, we are still, you know, improving. 13:53.000 --> 13:59.000 Recently, we, we even create a way to give a structure to the I. 13:59.000 --> 14:03.000 So to say that you can commit in your project like AI instruction dot mark down. 14:03.000 --> 14:06.000 So it's will analysis, it will add to the prompt. 14:06.000 --> 14:08.000 Each project they can have is on prompt. 14:08.000 --> 14:13.000 So we try to figure out what we can, you know, now everything is open source. 14:13.000 --> 14:15.000 Of course, it's in generate up. 14:15.000 --> 14:17.000 And then it's mirrored in GitHub. 14:17.000 --> 14:21.000 So every people that love GitHub, they have there. 14:21.000 --> 14:23.000 But you know, the main project is in generate up. 14:23.000 --> 14:26.000 That is a way to you get it. 14:26.000 --> 14:30.000 Then push back to to GitHub. 14:30.000 --> 14:31.000 Yeah. 14:31.000 --> 14:35.000 So I put some trivial example for some, sometimes you go, but, 14:35.000 --> 14:40.000 basically, what is doing is that pointing out to the context and review it. 14:41.000 --> 14:45.000 Now, the, the plugin is several way towards it. 14:45.000 --> 14:48.000 It can works only on the context of the patch. 14:48.000 --> 14:50.000 Or it can work in upload mode. 14:50.000 --> 14:54.000 It means that it will pull things that is not known. 14:54.000 --> 15:01.000 So if you have some see file that you have some function that it doesn't find in the context, 15:01.000 --> 15:04.000 but you won't understand why this was changed. 15:04.000 --> 15:09.000 It tried to pull the, the, the files that is contains the information that 15:09.000 --> 15:12.000 is to continue to review. 15:12.000 --> 15:13.000 It's not easy. 15:13.000 --> 15:16.000 We have still problem to connect it to the factory for the eyes, 15:16.000 --> 15:19.000 not trigger after testing after build. 15:19.000 --> 15:20.000 It's still over. 15:20.000 --> 15:23.000 Oh, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, 15:23.000 --> 15:28.000 it's, it's, and we've, I will be able to, to interact with the, 15:28.000 --> 15:29.000 with the eye. 15:29.000 --> 15:33.000 So we can ask question and we have interactive comment. 15:33.000 --> 15:37.000 This is some, some drawback about because Gerrit, 15:37.000 --> 15:42.000 we save each messages that is even reply as getting information. 15:42.000 --> 15:46.000 So this can scale a lot of the memory data, the digital data, 15:46.000 --> 15:48.000 the history of the Gerrit to maintain the threads. 15:48.000 --> 15:51.000 But it's the same of the code review. 15:51.000 --> 15:55.000 So the line comments also probably we have a lot of comments, 15:55.000 --> 15:57.000 but this will ask to review. 15:57.000 --> 16:01.000 And you can ask to review discard all the old comments that he was done before. 16:01.000 --> 16:02.000 So yeah. 16:03.000 --> 16:07.000 So the plug-in rom-up is very simple. 16:07.000 --> 16:12.000 So we want to basically give the possibility to understand what was the, 16:12.000 --> 16:14.000 the results of the test of the build. 16:14.000 --> 16:17.000 So he has more information back to review. 16:17.000 --> 16:21.000 And we want to integrate it on the Gerrit API, 16:21.000 --> 16:26.000 API because they do a lot to improve in 3.13. 16:26.000 --> 16:29.000 And we want that our comment can be a suggested change in the code. 16:29.000 --> 16:31.000 So apply these changes. 16:31.000 --> 16:34.000 So the user can directly apply this suggested code. 16:34.000 --> 16:37.000 But you know, it's, it's open. 16:37.000 --> 16:42.000 So we hope that something will, will happen. 16:42.000 --> 16:43.000 Okay. 16:43.000 --> 16:44.000 Okay. 16:44.000 --> 16:45.000 The testing. 16:45.000 --> 16:46.000 Yeah. 16:46.000 --> 16:47.000 The testing. 16:47.000 --> 16:50.000 That will show something. 16:50.000 --> 16:54.000 So we want to basically about a way to test our code. 16:54.000 --> 16:58.000 So it's nice to collect the things of a testing. 16:58.000 --> 17:00.000 But to what we, what I understand. 17:00.000 --> 17:04.000 I'm coming from the age where people write the code. 17:04.000 --> 17:06.000 And that's forget about testing. 17:06.000 --> 17:07.000 The code is working. 17:07.000 --> 17:08.000 It's perfect. 17:08.000 --> 17:09.000 It's written. 17:09.000 --> 17:10.000 Okay. 17:10.000 --> 17:13.000 But now it's every people that joined the company. 17:13.000 --> 17:15.000 Science the last six to seven years. 17:15.000 --> 17:19.000 They were ready at the concept that each part of the code should be tested. 17:19.000 --> 17:23.000 So we need to design the code to be testable. 17:23.000 --> 17:25.000 And it's, it's not only this, 17:26.000 --> 17:31.000 it's not the way to, you said, if we have this, if you don't test the code. 17:31.000 --> 17:32.000 You know, we can buy this. 17:32.000 --> 17:34.000 You need to design your code to be testable. 17:34.000 --> 17:39.000 So what this, I can help is to collect all this information through the testing 17:39.000 --> 17:42.000 and create the, the tools, 17:42.000 --> 17:44.000 can generate the output as there was, 17:44.000 --> 17:46.000 with describing the traditional code. 17:46.000 --> 17:49.000 And we get this output and interpret it. 17:49.000 --> 17:51.000 It is output through Jenkins. 17:52.000 --> 18:01.000 And let's see how this can be done in Jenkins. 18:01.000 --> 18:05.000 So how can Jenkins can help to run a testing? 18:05.000 --> 18:09.000 I wish you basically how Jenkins can collect results coming from 18:09.000 --> 18:11.000 JU, it's a unit testing. 18:11.000 --> 18:16.000 Or how can Jenkins can provide and collect information from 18:16.000 --> 18:17.000 tools like Lab Grid? 18:17.000 --> 18:20.000 I don't know how much of you use Lab Grid in other integration 18:20.000 --> 18:22.000 or for continuous integration, but okay. 18:22.000 --> 18:27.000 It's just a way of wrapping of resources that you connect to your hardware 18:27.000 --> 18:28.000 and we see later. 18:28.000 --> 18:33.000 And you can create easy pipeline using Python so to test your hardware. 18:33.000 --> 18:36.000 So that is the concept. 18:36.000 --> 18:40.000 So what I mentioned, so we need to decide the code in order to be 18:40.000 --> 18:41.000 testable. 18:41.000 --> 18:42.000 So it's not important that, 18:42.000 --> 18:47.000 recently we done some IoT design firmware with, 18:47.000 --> 18:50.000 in the, in the Netherlands and the office in Netherlands. 18:50.000 --> 18:55.000 And the lead decided the, the IoT in an order way 18:55.000 --> 18:59.000 that we can generate the test for this IoT 18:59.000 --> 19:03.000 and forever like around 79 or 80% of coverage 19:03.000 --> 19:04.000 of the code. 19:04.000 --> 19:07.000 And create all the abstraction that allow a lot to test 19:07.000 --> 19:09.000 all the part of the, of the IoT. 19:09.000 --> 19:11.000 So if we designed the code in the process, 19:11.000 --> 19:16.000 the code in the proper way, then we can come out 19:16.000 --> 19:20.000 with trained statistics and this can be happened 19:20.000 --> 19:22.000 when you generate the releases. 19:22.000 --> 19:25.000 You can have it if you have like verification build, 19:25.000 --> 19:28.000 you can create gate, you can, if you have, 19:28.000 --> 19:30.000 you know, everyone know that if you have unit test, 19:30.000 --> 19:32.000 you can generate code coverage because, you know, 19:32.000 --> 19:35.000 without unit test, you go coverage when anything. 19:35.000 --> 19:39.000 Then you can run the mem checking, use the value grid, 19:39.000 --> 19:43.000 and then you can, everything that you can check 19:43.000 --> 19:45.000 at the time the CI is better to implement. 19:45.000 --> 19:48.000 We even check documentation warnings, 19:48.000 --> 19:50.000 semic file warnings. 19:50.000 --> 19:53.000 So everything that can, let's say, 19:53.000 --> 19:57.000 reduce the time to review code that is not satisfied. 19:57.000 --> 20:00.000 Our standard should be integrated in the pipeline. 20:00.000 --> 20:07.000 So, so we want only to review buildable patch. 20:08.000 --> 20:11.000 So CI can run a test several big configurations 20:11.000 --> 20:15.000 and run, because we have those, those important things 20:15.000 --> 20:19.000 that we discuss it, that we want to build a busy capability. 20:19.000 --> 20:22.000 So we don't like GitHub for that one, for instance. 20:22.000 --> 20:24.000 It's one of the problem. 20:24.000 --> 20:28.000 Then we want, if we want really to work as a GitHub 20:28.000 --> 20:31.000 or a GitHub, we want to use topic build. 20:31.000 --> 20:34.000 But the topic is more sophisticated than per request. 20:34.000 --> 20:36.000 Topic is across multiple bits. 20:36.000 --> 20:40.000 So a topic can be a change set that is in three of four 20:40.000 --> 20:41.000 key repositories. 20:41.000 --> 20:44.000 So when you verify an operating system, 20:44.000 --> 20:47.000 you want to apply all the topic across two 20:47.000 --> 20:49.000 repositories and buildable as a bundle. 20:49.000 --> 20:52.000 So a topic is a concept that is across multiple 20:52.000 --> 20:54.000 git and this guarantee is possible, 20:54.000 --> 20:56.000 because you can create connection between the poor 20:56.000 --> 20:59.000 request or between things in the different repositories. 20:59.000 --> 21:02.000 This is the reason why I was created by Google for Android. 21:03.000 --> 21:07.000 This is an 있다 PROoped type build at trigger. 21:07.000 --> 21:10.000 So you have the patch set that is tricky. 21:10.000 --> 21:14.000 Get it in the instance and the run the job and execute the test. 21:14.000 --> 21:15.000 So nothing more. 21:15.000 --> 21:20.000 And then this is a cities, this is a new stage view. 21:20.000 --> 21:23.000 This is a new쟁ist stage view, a pipeline graph view 21:23.000 --> 21:26.000 that is better now that show the steps. 21:26.000 --> 21:28.000 You can sew the green and the steps in this 21:28.000 --> 21:30.000 are what is going on. 21:30.000 --> 21:34.360 going down, and this was the result up to Gerrit. 21:34.360 --> 21:36.960 So Gerrit is constantly connecting what happened here. 21:36.960 --> 21:41.680 You will have a notification back to your job. 21:41.680 --> 21:44.760 Yeah, this is what happened in a standard Jenkins 21:44.760 --> 21:47.960 unit test pipelines that there is some 21:47.960 --> 21:49.960 then in this slide, there are some weak 21:49.960 --> 21:52.080 that remind how to create them. 21:52.080 --> 21:55.440 But the basic what you are trying to do is that you 21:55.440 --> 21:57.600 want at some point to perform the test, 21:57.600 --> 21:59.640 and then you have a parallel blocking Jenkins 21:59.640 --> 22:01.840 that all the tests can be run in parallel, 22:01.840 --> 22:06.360 and execute all the testing features. 22:06.360 --> 22:09.600 Yeah, there are some examples out to use a code checker 22:09.600 --> 22:11.360 or one again to cheat in these slides, 22:11.360 --> 22:17.880 but I prefer you to see what to talk more about now. 22:17.880 --> 22:19.680 What we have testing, we are unit testing, 22:19.680 --> 22:21.320 we cannot complete things. 22:21.320 --> 22:23.280 Now, we are not only testing films, 22:23.280 --> 22:27.880 but we increase all the power power elements together 22:27.880 --> 22:30.720 to cover all the things like other testing. 22:30.720 --> 22:34.280 OK, now, the rest of the year, we started to have projects 22:34.280 --> 22:37.880 that allow us to test hardware, so we are up to her lump grid, 22:37.880 --> 22:40.960 that is basically, I frame us that a lot of 22:40.960 --> 22:44.880 alokiter resources of colored places to hardware, 22:44.880 --> 22:46.680 and then you can interact with user other 22:46.680 --> 22:49.960 to replace it, so basically what the CI is doing is 22:49.960 --> 22:52.880 ask for resources, if the resource is busy, 22:52.880 --> 22:55.480 by other CI, there are other job that are running, 22:55.480 --> 22:59.040 so the CI is waiting for the resource to be free, 22:59.040 --> 23:01.640 and then get the resource automatically free. 23:01.640 --> 23:04.240 So basically, there are lots of run, 23:04.240 --> 23:07.240 multiple testing, both even the resource are partially used 23:07.240 --> 23:11.600 by other jobs that are running, but when you have get the resources, 23:11.600 --> 23:15.480 then you are located until the job is finished. 23:15.480 --> 23:19.400 Lovely, there's an architecture that is basically, 23:19.400 --> 23:23.800 it's let you to basically have several components, 23:23.800 --> 23:27.080 what is important is to allocate, as I said, displays, 23:27.080 --> 23:32.440 so you will have the client to disconnect to the lab grid. 23:32.440 --> 23:34.360 A lab grid is not only used by the CI, 23:34.360 --> 23:36.440 you can use it by any user to assess the resource 23:36.440 --> 23:39.600 of what it devices, so we use not only to interact with CI, 23:39.600 --> 23:43.440 but we use even to interact with the hardware. 23:43.440 --> 23:45.720 Then we have the support, the coordinator, the clients, 23:45.720 --> 23:49.480 very easy, the exporter is the one that has the resources, 23:49.480 --> 23:52.720 connected to coordinate is one, all the exporter, 23:52.720 --> 23:56.200 give to the coordinator what they have in them, 23:56.200 --> 23:58.360 connect to it, and the client is you, 23:58.360 --> 24:03.480 the task that the resources to execute any test. 24:03.480 --> 24:06.320 OK, the key feature is basically the support for the hardware, 24:06.320 --> 24:09.080 so you have serial, SSH, power manager, 24:09.080 --> 24:11.040 social location, so it's quite easy, 24:11.040 --> 24:13.960 because they have a lot of driver information tests. 24:13.960 --> 24:18.000 And there are, they're not a nice things that I don't find any 24:18.000 --> 24:21.840 a lot of testing, but I will make my things public sooner, 24:21.840 --> 24:26.600 so that someone can start from there, I think. 24:26.600 --> 24:29.960 Yeah, the lab grid configuration is done in a YAMFI 24:29.960 --> 24:31.760 that you have in the nodes. 24:31.760 --> 24:34.320 So let's see, because we have not a lot of times 24:34.320 --> 24:38.320 this is a test example, lab grid, but I want to talk straight 24:38.320 --> 24:42.480 forward on, OK, this is the benefits, 24:42.480 --> 24:44.840 but I want to show some pipeline there. 24:44.840 --> 24:47.680 So this one, for instance, this one is a kind of pipeline 24:47.680 --> 24:50.520 that is part of the test, multiple configuration, 24:50.520 --> 24:53.240 go to lab grid, execute all the tests. 24:53.240 --> 24:56.120 It's not, this is the due to preparations, 24:56.120 --> 24:59.320 then it is the running test for different architectures, 24:59.320 --> 25:02.160 and then this can run basic hit-down, yes. 25:02.160 --> 25:05.480 OK, that is fail, because the YFI is unstable, 25:05.480 --> 25:08.120 the YFI test, so it's a one in that. 25:08.120 --> 25:14.320 It's not really fail, but one gate was not passing during this pipeline. 25:14.320 --> 25:18.720 But you can find it basically there, the pipeline. 25:18.720 --> 25:20.480 OK, well, one of the things 25:20.480 --> 25:24.240 that you know, all the things that I want to have in an infrastructure 25:24.240 --> 25:27.640 is understand why my build was failing. 25:27.640 --> 25:31.040 So now I add up what is in the G, as I said, multiple things. 25:31.040 --> 25:34.440 Now, even the languages was added yesterday by the content of today's 25:34.440 --> 25:38.360 goal, and this is very powerful, because we know 25:38.360 --> 25:43.040 when a build failure is happening, and we have 25:43.040 --> 25:47.400 a dedicated explainer or a part where we can see where 25:47.440 --> 25:52.640 the arrow comes from, and you can find the plugin. 25:52.640 --> 25:54.360 There is even screenshot, how should we? 25:54.360 --> 25:57.560 Now I don't know if I have time to show you. 25:57.560 --> 26:02.160 And yes, basically my roadmap is to integrate in pipeline 26:02.160 --> 26:05.240 like I feel, and improve the nodes. 26:05.240 --> 26:08.080 So what I'm doing with the I is that I need to find the steps 26:08.080 --> 26:11.560 that you fail, but if you have parallel step, like 10 steps, 26:11.560 --> 26:15.800 then you fail three things together, and the I should not 26:15.800 --> 26:16.920 only take the first one. 26:16.920 --> 26:19.880 My goal, the free, collect all the results, and give to the 26:19.880 --> 26:22.400 OK, it failed, the unit test for this reason. 26:22.400 --> 26:25.400 If fail, you don't know, the same check for this reason, 26:25.400 --> 26:26.840 it failed, you know, for this reason. 26:26.840 --> 26:28.760 So the idea is now to navigate. 26:28.760 --> 26:32.360 If you want to improve the committees this one, it's very easy. 26:32.360 --> 26:33.200 I go the three. 26:33.200 --> 26:35.400 I go back to the first node, the failure, then there's 26:35.400 --> 26:37.360 sent to the I. 26:37.360 --> 26:39.520 And then there is a lot of useful link. 26:39.520 --> 26:43.340 OK, sorry.