 Okay, so release testing, most importantly, it is not our formal report because, well, this release is a big one. We have a lot of features and most importantly, deploying it on main took much, much longer than we initially anticipated. So we had to start at least testing the release much later than we initially planned. So we simply are currently in the process of doing the testing. So I don't have the formal release done as of yet. We will be posting it somehow, somewhere to the community, probably within the next week or two. But so far here in the current update. So first of all, we have a large team which is very, very nice considering the scale of this release. We have Catherine Bromhead from Australia. We have Simon Bray, we have Fabio Cambo, Jennifer from PSU, Ahmed and Tyler from Johns Hopkins. So they all have been super productive, very helpful and are doing a lot, which is great. So basically here is what we do in release testing. We have primary, the overall goal is to find everything we might find that a user might stumble into. So we try to follow, to not follow the happy path. So we're trying to explore what we might break intentionally or unintentionally. The three focus areas are historically we focus on the most important stuff which is promoted with the release. It's usually highlighted in the release notes above all the list of all the PRs, usually it's highlights or items to be deprecated, new configuration changes, these kind of things. So we have these high priority items. This is one part. The second part is a long, annoyingly long list of PRs which potentially might warrant manual testing. This usually has been over 200 or 300, but this time we're trying to do a more sensible approach and including only the stuff which has been explicitly tagged as manually testing or UI UX. So we have a list of 141 PRs. And the third focus part is tutorials. So tutorials, we do of course try to discover issues with the tutorials per se, but tutorials primarily are used as a sort of a common use examples of common usage. So it's a scenario-based testing. We take the tutorials, we follow the steps, trying to mimic what the typical user would do with Galaxy. So these are the three areas. And here is just to give you an idea of where we're at currently. So we are currently in the fourth day, although I suppose at least two of us were testing over the weekend, thank you for that. Currently we have, we started with the high priority items and we have covered 10 out of 11, the only thing which has remained untouched as of yet is the storage dashboard. So this is the big PR which made by David. So we'll be testing that one. And after that, all the high priority items will be done. Out of the tutorials, we had 24 tutorials. We are currently on 19, 16 are completely done, three are almost done. So we have almost 100% coverage of the tutorials and high priority items, which is great because these are the things that matter the most. Now out of the 141 PRs, we have covered 37, which is what almost 25%, actually it's more than 35%. We still have a lot left for context. We never get to cover all the PRs. We, I think in the past we've covered, I don't have the numbers. I believe it's not more than 25%, if not less. So we are on track to cover a significant chunk. We are working on that. Now out of the issues we have found so far, most of them are issues with the UI. But again, this does not mean that the problems which might occur with the updates are UI problems because the only reason we are finding mostly UI issues is because that's the thing where I focused on testing because that's what makes sense testing manually. And I suppose that's it for now. There are no major obvious smoking guns or fires. That we have discovered so far. Some issues are, most issues are minor. Some issues are potentially annoyingly non-minor. Like for example, we've discovered that uploading composite data types is broken, which is worth it. If you paste the content by text, which, you know, I don't think it's very common. Not only, but that's what we thought at first. And it's not just that. If you, at least, if I'm not testing it incorrectly, but I tried both and I looked in the code, apparently if you do file uploads, at least with the AFI and valid data types, you get an error. So something is going on there. Other than that, I think there are no massive issues. Oh, and we have, since we have the new history, it is expected that all the tutorials or most of the tutorials will be mostly inaccurate and outdated. So instead of opening a gazillion issues for every little thing, we have a matter issue on GTN where we add comments for all the stuff which we discover for tutorials and most of it applies to all of the tutorials. And as I've already told Helen on the channel, we are explicitly not fixing anything during the release testing period for the simple reason that the moment the team starts fixing stuff, the productivity for discovering new stuff drops dramatically. So we are not fixing that stuff, but as soon as the release testing period is over, which will be today or tomorrow, I'm sure we'll be all jumping on fixing those little things. Yeah, very nice. I mean, do you have a caveat of the number of issues you've opened? I think it's like more than 20 or something. Hard to count because some of the issues are opened explicitly by the team. One of them is a matter issue which contains a lot of stuff. Right now on my list, I have 18 issues. I think it is current, but in addition to that, other people are testing and some issues are opened as a result of discussing stuff between the release team or other channels with the release team. So these 18 are only the issues which I was able to honestly tag as a result of the release team's effort. So 18 plus. All right. I mean, maybe I can add that we're also with every release testing the tools and workflows. I don't know. I mean, if you're a Galaxy veteran, you may remember the situation from a couple of years ago that whenever we would release a new version and somebody tinkered with the features or the performance or anything you could do that is remotely tool related, which essentially is everything. We had like a fraction of tools that would break, right? And in return that meant we had a fraction of workflows that break just by upgrading that Galaxy. So I think for the last two or three years we've run through all the tests of the IUC. So that's the largest repository of tools. And we run these tests before we update main, right? So main often runs a version that isn't released yet. And main has a lot of users. So we don't want to break a lot because then Gem and Nate will be turning in circles. So I mean, that's something that's not exactly release testing because that's always the release management team that's doing that. But we can be fairly confident that, you know once we ship Galaxy at least most tools, most workflows are gonna continue to work. And for the last two releases we're also testing all the IWC workflows. So I mean, I wanna highlight like, you know it used to be a thing that once we went to a new release we were scrambling to fix stuff and now we don't have to because we catch these things early. I think this is always worth highlighting that's a success story. Yeah, I think we can, I mean if I know further questions we could move over to the second part which is I just wanted to give you like a super brief overview of the new features in 20205 and the things that I've changed. So let me share my screen. Oh, I was supposed to say how the participants can share. Well, Mary's is swapping over. I would just, if anyone else uses the new release on main and you have any sort of little tweaks or things like that, please submit issues. Don't hesitate to do that. It's never too late, so. All right, so got it. I hope everyone can see my screen. Make it nice and big. So yeah, I mean, John already mentioned that we have a list of pull requests. So if you wanna see the big sexy things and we haven't produced the release notes yet you can check the milestone and you can check the highlight label. So the highlight label ones. These are cool things or finicky things or things that change something for the user that is maybe not super obvious. It's always gonna be one of these things. So yeah, with a new history come a lot of new features and especially in this release we have like big tweaks towards doing things in bulk. So I think one of the like serious limitations of the old history was that you couldn't do things in batch really beyond the 500 items that the history would load. So if you had more than 500 items in your history you had to find workarounds that were never really great. So now what you can do is you have this little icon here you can click that and you go into the sort of multi-select mode. You can do select all which behaves like the Gmail search. So if your search has found these things you can apply them, you can select them all and then do things with them. We had this in the old history except the old history it would only apply to what you saw on screen what you currently had there. And you can do all these things. So you can change the DB key, you can add tags, you can build lists, you can delete things. And when salary is enabled and we had it briefly enabled on main but it was writing to temporary directory shouldn't which we fixed but we haven't re-enabled it but once salary is enabled you can also change the data type. And you can change it for everything in your history you can change it for datasets, dataset collections. So I think that's a big deal. I think Sam already the last time showed this cool new search panel and this search panel is integrated with the selection. So if you search things here you can apply things to that search. So let's say we wanna get some bound datasets. That's unfortunate because there are clearly bound datasets in here. All right, found the bug. Yeah, let's do something else that say mapping. You have the same thing. I mean, now I've just one dataset here. Yeah, you can clear your search, go back to the things and notice how everything is like really instantaneous things are coming from the cache. Yeah, so this is the bug selector, this is the search. We have a little icon here that we added based on user feedback. So whenever Galaxy refreshes, we mentioned this here. So you know that, you know you don't need to reload the entire page things are getting updated. Yeah, then we have the storage dashboard. So that's something that we took the lead on and this is like really just the first phase. So right now, it presents a little bit like if you've used your Google account to clean up things. I mean, it looks kind of similar. So if you click on this little icon here you get to the dashboard. I got 250 gigs of total disk quota. I can refresh the disk quota here which previously you had to know that you need to log out and it would automatically recalculate the disk quota. So that's a big step up already. But the really cool stuff is that you can free a disk. So there are different modules right now. You can only delete, I mean, permanently delete datasets that are already deleted. I mean, that's a common issue that's really valuable but we're going to add additional things like deleting your largest datasets or just finding your largest datasets and deciding whether or not you want to delete them. So now I don't have any deleted datasets. So let's select some maybe under one. We can say delete. If this is the icon here that we're doing things we're going to trust that's going to work. Yep, there we go. So we can do that same thing again. And this time you see there are datasets. I mean, this is all very small but yeah. These are the datasets I can delete. And there we go. Yeah, I mean, all of this can take advantage of the salary task queue. So for this small thing went really fast but in general all these things if the salary task queue is enabled it would be much faster. Yeah, we have the new sketchbooks Sam added this. So if you've seen the sketchbook before I mean it works the same way. You just click here. And this is like much newer, much nicer technology, much snappier. You can minimize it, open another one, maximize, you can switch them around. So that's really very cool. From the history, we have this context menu. Yeah, it mentions how many histories we have. What else do we have? I mean, not on the UI side. Do you want to get into deferred data or any of that? Yeah, sure. I mean, that's a good idea. While we're waiting here. Actually, I was just updating Felix on the side. That may not have been a good idea. Let's switch to the test side. Let's see. We have deferred data now. I would just have to find some data set that is on a remote location. So what deferred data is, is you can basically take a URL or anything in Galaxy's file sources and you can create an upload that doesn't actually end up on Galaxy's disk object store. It's immediately ready and it will only be pulled in once you run a job on it. So I'm just gonna pick here file from Galaxy's repo and this option is available here under the settings. So what you do is you select deferred data set resolution. So this data set will not actually enter Galaxy's object store. So, see, I mean, it says now this data set is remote. Actually, let's do that again. Let's say, so one thing is that you might wanna set the data type because otherwise Galaxy can't really look at it now because it's remote. So it will not infer what type of data set this is. But then once you run a job on it, the data set will be pulled in. I forgot to click the defer button. Let's do it the third time. And you know what? Probably forgot to set the data type. Whatever, we can also do it in auto. Server has also reported at this point. Okay, so there we go. So I ran this data set on the deferred data set because I didn't set the data type. It just took the default data type if we don't know it, which is data. But you can always change that afterwards. I think Marius may have frozen. Well, not anymore. Yep, there you go. I think he's frozen. Marius, I think you have frozen and or muted. Doesn't seem like muted, right? So it should be frozen and it's not changing much the screen, so. Sorry about that. Wi-Fi is a bit unstable sometimes. Yeah, it says I'm sharing the screen but you can't see it, right? Because I can see it. We can't. Oh, you can't see my screen? I'm frozen. Let's, now we can't. Okay, that makes sense. Yeah, I can share the screen right now. Because, ah, there we go. Okay, so for whatever reason, the setting metadata ended up with an error state. But if we click on the I, I can, we see it actually work. So we've materialized the data set. In the meantime, main has rebooted. So we can also look at it here. So let's take that URL, do the same thing again. Paste data, say defer data set resolution. I mean, this doesn't have any effect. And this time, let's say it's a txt file. The last five lines of the file. So yeah, I mean, you see this is the expected thing. The data set is remote. If you click on the info button, you can also see that the data set is remote and deferred. So the data is not actually in Galaxy but here we've cut the last five lines and those are the last five lines of our citation file. Yeah. Let's say that's a galaxy and doesn't use your photo. Exactly. So the big thing is here, if you, yeah, let's pick something big from the thousand genomes. I don't know what's big here. I mean, surely that's pretty big. Now we can also say defer data set resolution. I think that was a BAM file. So now we go and I mean, copying this into Galaxy's object. So we typically take a good while while this should be near instantaneous. So Marius, does the file become local after you ran the tail command? After I ran what? After you ran the tail command. Just on the worker node and depending on where you're in worker node is, but I mean, yeah, it doesn't enter Galaxy's object store at all. Yeah, that's the deferred data. It's there. So we could run, what can you run? SamTillsView, it's a little slow here. So, yeah. I mean, there's a lot of other really cool things in there. So the rules, I mean, okay. So SamTillsView is here. We can select this file even though it's remote. And now we could say, we could just filter that down. Yeah, let's sub-simple. So we could say we just take, I don't know, a very small fraction. And so that will only run on the worker node. I guess I forgot to click something, but okay. Is it this? Looks like it. Okay. Yeah, we have a couple of other things. All right, I think I'm gonna stop here. Yeah, and as you see, we're still working on the release. And some of these things will be presented in GCC, right? Oh, yeah. The features separately or how do you plan to organize that? So, I mean, we have a lot of different talks on this. So there will be talks on the message queue, so the salary tasks that I just briefly mentioned. The PIs will do a talk on, that will involve remote data and the new history. We'll have talks on the API. I mean, another thing I didn't even mention is that also the legacy API routes are now visible in slash API slash docs. There are individual talks about the new history in a broader form and about the storage dashboard too. There were a lot of changes to the model and to the migration system. It's not quite user facing, so there are no features which would be obviously different on the UI or in how user operates, but there would be differences in how Galaxy is administered and well, how Galaxy works in its guts. So there will be talks on that too. Okay, nice. Thank you. Are there questions for Mario or for John or for the others or any comments that you would like to share? Nothing else? John, do you have a quote from a movie or something? I'm sure you do. I do not. Anything else to discuss from both your side? John or Mario's? Anything else that you would like? No, I mean, please do use Galaxy currently. Let us know what seems weird. Just by doing like this impromptu live demo, we've seen a few things that didn't quite work. So yeah, that's always more work. And if you come across something, please open an issue and don't worry about the issue being in depth with screenshots and whatnot. We encourage our release testing team to be very explicit about what exactly goes wrong, but if you can pinpoint something, just give us a general idea what's happening and we'll take care of it. I mean, it should be at least, you know, what have you done and where, you know, what's going wrong? Yeah, but otherwise it doesn't take a lot to report bugs. Yeah, and if you want the extended version, then go to DCC. Exactly.