 Then, welcome to the engineering steering committee. What we usually do is meet once a week, virtually, and there's typically more of us, but the rest try to hide today, I don't know. And just go through our weekly schedule, which is very rigid, so that everybody who's supposed to speak up at some point knows when to wake up and do their part. And when we discuss all the matter that comes up during the week, how to proceed, I think we never need to... We always agree in the end on what we want to do, and whoever wants to do something does something, and that works out all pretty well, and we're all very friendly people. So whenever you have some technical issue around LibreOffice that you think is best solved, not by yourself in your small little chamber, but by reaching out, then do come to these weekly meetings. We're always open to visitors, to guests, and present your issues. But today, we'll do it completely differently and not have any fixed schedule, but awaiting for your input and your questions to hopefully give you some answers. So if any of these experts sitting here might know something that you always wanted to know about spoons or software or whatever. My first question is more like a usual ESC committee session in line with it. There was a recent security problem recently, and it was fixed for the next versions of LibreOffice. Are there plans to create out of schedule fast release? Yes, indeed, there is a plan to create an out-of-schedule release for the 7.5 line and also to shorten the one for the upcoming 7.6 to just one RC candidate. And both will be tagged this week, and we're planning to have a release ready for Tuesday. Great, thanks. Next question. Leo. Anything? Anything. Okay, let's try. Yeah. You'll manage. Well, I'm keeping account of my bank accounts, and it is a rather huge data set, and it is rather private, but it is very slow, because I use a lot of search functions, and I'm now wondering how can I pass on the data without giving free my personal data. So one way or the other, I should try and analyze the data, but still have all the formula there in order to... Maybe my computer is too slow. That could be a reason, but if I look at the system monitor, I see that it's running at 100%, although I have four processors, they keep on jumping to 100%. So in fact, it's a general question, but how to analyze my data, and how to try and reduce the amount of processors' time? Is that the question that is valid for to be asked here? Thank you for the perfect question. That was even two questions in one, so one is about calc performance. I don't think we have any... There's no ICA around. That's the problem, so he might know. Or is anybody... Part of the question is kind of maybe a QA question. Exactly. That's part two. So no help with getting it faster done, but at least with the anonymization. Maybe Cisco... Yeah, so regarding the anonymization of data, that's actually related to the other question as well, because once you can get the document anonymized, then you can create a report and maybe someone from the core developers can take a look at it. And yeah, regarding how to anonymize the data, well, my first approach would be to using calc, try to use replace. You can use regular expressions, so you can replace everything by a single character, like X, and if it works, then everything is fine, it's great. Otherwise, I don't know, I would try to remove parts of the document as long as the problem is still reproducible. So in the end, you just have a minimized document that still triggers the problem, and then when you already have something minimal, then it's much easier to anonymize this document. That would be the two approaches that come to mind, but maybe others have other ideas. Yeah, I mean, especially for calc, it might be because what you need there to reproduce something is the formulas probably and the amount of data, but not the exact data. So if you have some issue in writer, maybe it's more difficult to reduce that, to obscure your love letters and not send them around. But for calc, it sounds like maybe we could try to do an extension that automatically replaces whatever numbers with random numbers and whatever text with the random text or something like that sounds like that could be useful. So if you want to get into extension development, we have a whole website and wiki and whatnot where you can start to implement that. Thanks for doing that. Yeah, random numbers would be great there, because trying to reduce your problem and still see that it runs slow might be, yeah, it would be great for the QA and for the developers if you do that upfront, but it can be very tedious, I guess, so nobody would actually do that probably. Depending on the kind of data and the kind of operation that you are doing, you may evaluate the possibility of using array formulas in array formulas, which is one formula for a big set of data and it may be faster. It's a matter of testing. There's a hidden calc expert here. I've already worked on high performance computing, so we have to tweak the data so that it can be processed faster. Yeah, I have a question. Is it this document, something that used to work properly in the past and is not working as slow or triggering a performance issue nowadays with that recent version, because if it's that the case, maybe, I mean, because we are here in person, maybe we can take a look and don't publish it somewhere, like anywhere in the Paxilla, but at least we could take a look together and maybe see if we can know when it was introduced and from there we can move on. I started with it something like three years ago and most things, you are happy with what you get and then you see new opportunities, so you keep on increasing, adding formulas, adding new functionalities. So I'm now, I think, at version 14, which means every time I improve it, I make a new version to it, but I see it also when I started with version 2 or version 1, okay, I didn't have that much data, but I didn't have that much functionality either, so it has been an ongoing process and that added to the processing time and to the complexity of the spreadsheet. Okay, thanks again. There's a question up here, it works like a charm. Thank you and it's good to see you all in a row. I have a question regarding official documents that you as a committee are responsible for or policy documents or whatever that can say something in regards to the recommendation of ODFS as the default format for recommendations to users, is that something you can elaborate a bit on? Thank you. Yep, thank you. Is anybody of you eager to answer? Yeah, so I guess this is more of a, maybe even more of a marketing issue than a technical issue, so what we or a political issue, maybe where we are very eager not to get drawn into anything for a reason and which kind of worked well for us for more than 10 years by now. So what we concentrate on is really there's an issue with a code or with a CI pipeline or with anything really technical, not with how does a feature appear to the end user, we're more concerned with how is a feature implemented, does that bring any issues if we want to implement another feature, does that interfere maybe in a way that is non-optimal. But we, I think we never decided on anything in any way visual or apparent to the end user just to the underlying things, even the version numbers we defer them to the marketing team to tell us that the next version will not be 7.8 but 24.2 is the next one. So I'm afraid you better pick, maybe Italo will be the go-to guy for questions like that. Well, thank you. Anything else? Aiko, your chance to get a question in from the other side that you're not sitting down here, so it's your turn now. What? You asked me rather why I'm not sitting over there and it won't be to answer it. No, I don't have a question to the ESC. I just opened my notebook because do we have maybe a jitzy or some streaming? Maybe the audience is giving some input via stream. I don't know. Mike, do we have anything from the outside world? Well, the people here are sitting together anyway and I wonder if we have, for example, a jitzy stream to somewhere in the world and people are listening to us. Do they have questions? Maybe. People can send questions into the chat. I'll take a look now. I will just hold on for a second and wait for the outside world questions. Questions yet? People can ask questions. Come on, people around the world. Where's your questions? Maybe just a simple point to the question regarding recommendations, official documents and the like. Maybe we have the LibreOffice Migration Protocol, more or less. This touches up on open document as a format that represents open standards. So maybe this is a starting point to expand upon. I came to the file form as well. I think one of the things that did come to the steering committee at one point was when the idea of changing the default version, which version of ODF were we going to save by default, that would have come to the steering committee, that the idea was we should now save whatever it was, 1.3. Is everybody okay with that? Does anybody see any downsides? Have a discussion about it? That the oldest supported version of LibreOffice out there would be able to import it in maybe two or three versions before that so we all kind of knew what kind of impact it would have. So that kind of thing does come to the ESC and there is some thought about it but it's about the limit of it that I've seen so far. Obviously if somebody had two competing implementations of something, that's the kind of thing that would be put to the ESC as well. That's a good point. Thanks. New guests, new questions. Yes, hi, it's Stefan from the team. I don't know very much about the ESC but I wanted to know how it's generated the input you discuss inside the ESC. Where does it come from? Is it the personal input you bring in where you are focused on to work, from the companies you are working for? So it's just for me an interesting question where this input comes from. Thank you. Thank you. Anybody wants to pick up spontaneously? I should. It depends. Me, I get a lot of input from users, request to enhancement or issues and I pick those that might be of interest for developers in the hope someone is interested. I believe the ESC notes gets a lot of attraction, people read it or at least skim through it. I see comments on the mailing list later on for let's say some compiler change and whatever. One example that I looked through and example what I do typically, it is a request that we've got this week and it is about mixed addresses in CULC. You can typically say give me some from A1 to B2 or give me the result from all the column A. Mixed means from A1 to B. And this feature exists in Google Sheets and the user asks why not in CULC. So I was wondering is it defined by ODF and it's kind of connected to the question before. I looked through the description and it reads to me as if it is possible. It is not implemented, you cannot do it and when you export the document in Gsheet then you get something replaced. It is A1 to C3 or not. Shall we do an enhancement? I involved our experts Iker and he said no it's not defined. You found another place in the ODF description or the reference where the opposite is stated. It has to be row column column or row column row period. Nothing else. And the point is at this point in discussion where do we have to do the change? Probably the standard needs to be more clear on what is possible. That's currently the outcome but that's what I do. I flag the question during the call. No one can answer of course not but it gets hopefully more attention. Does that typically work out or do you follow up on how many of the devs respond to your tickets? You are often received with deafening silence in the call so I'm a bit sorry for your attacks. Unfortunately I do not get more time from volunteers. No. Okay, sad part. Coming back to that question of where do we get the input from of course everybody who regularly sits in on the IEC have their own set of questions that they bring but another recurring pattern is that maybe so we usually meet on Thursday afternoon European time and what often happens is that during the week on IRC developer channel there is something coming up and some interesting question that somebody has who is not usually going to the IEC and then somebody steps in and says yeah that's something that we should actually discuss there and then takes on the job of bringing that in and having a quick discussion and a consensus then on whatever topic arises there. It's often about another thing that we try to encourage people if they have any incompatible changes so that they would break anything for the user or for the external developers then that they should bring that into the IEC to discuss it there whether that's something that we can change because the chances of breaking people's external code is low or whether that's something where somebody who's an expert in that field says oh no everybody's using that hands off please and then there's usually somebody who is kind of proxying and for whoever has that issue and discuss it in the IEC calls. I just want to add something to the question of Stefan we have this weekly meeting of the IEC it's rather classical project management meeting where each of the individuals presents the results of what they are doing we have information brought by release engineering documentation quality assurance, crash testing and all this information is put in common on this meeting and comments are welcome if by any chance someone wants to make a comment on the topics that are being reported so it's quite a very tight management of the project and very professional. And also to mention that the IEC meetings are not a close group so everyone is free to attend and bring up their own topics ideally when the minutes are created so people can have an idea and prepare whatever and yeah but to make it true it's not about your personal issue that needs fixings and not for this general stuff so this is pre-filtered by your IECs and QA more or less but yeah you're free to attend so nobody will kick you out and you're welcome to give your own input. I suppose just to fill in as well and one of the questions was the input from companies I think there's a regular slot about what people are working on and I guess one of the main purposes of the IEC is to just flag in advance that you are working on something that's going to be potentially disruptive to make sure people are aware of it so that you don't wake up some morning and everything has changed without there being some forewarning so you had a regular feedback and there was going to be changes to the the gradient work that we see now the multi-step gradient there's about four or five talks on gradients in this conference here so that was flagged well before it landed that there was going to be changes to gradients and to get some idea of what kind of changes there would be in case anybody had any objections or maybe more partly in case there were two or three different people working on gradients at the same time and treat it from ways to make sure that that didn't happen it's just to let other groups know that you're working on something so if you're working on something and you're asking letting people know in advance that there's going to be a lot of translation changes required for this I say if it's going to be a feature that lands just before the usual translation freeze let people know that there's going to be a huge number of translations and maybe you say okay that's too many translations to land just at the end why don't you put translations in now even though you don't have the feature available then translate it and then you can get the feature in at the very end and the translations are available and that kind of coordination is helpful to you see helps do that kind of coordination Another interesting piece might also be that yes we always have an agenda we create up from Miklos who's I don't know why he's not around this year he's always the one or thankfully the one who prepares that every week and typically sends it out on Wednesdays to the developer list and the QA list I think at least and if you have any topics then it's great if you just send a reply to that maybe so that we have an idea that there's more coming and we can allocate some time or prepare even for there will be some more questions coming in Okay, thank you Sorry? We're nearly at the end I think we have a full hour We tried hard it didn't work out Are we at the end of the full hour already? Wonderful So any more questions? Any last questions? We're running out of time So another thing we usually have in these meetings in the regular schedule is that we have this what's cooking section which is maybe the most interesting part also for the outsiders because that's where if somebody is working on something shiny, big, new, interesting potentially disruptive or just large enough to let everybody else know that there's something coming then they inform the rest of the meeting about that in this what's cooking section One thing we obviously missed for example is what Marina presented earlier today on this outreach project where they are going to replace the windows installer construction of what then ends up as the installer which is I think a very great project because the code we have there is very, very old written by somebody he's no longer with the project so we can confess that he had no idea about what wrote a million line of code maybe and we're stuck with that now and we will be very, very, very happy if we eventually get rid of that and I think this is a project that kind of went under our radar for the most part of the last three months I think the time this outreach project ran and that's something we should pick up on and see how to move that forward for example by so that's now an optional way of building it so we have the old way still because the new way is not working probably or fully functional function complete yet so we have a switch whether to use the old or the new and we should set up one of the CI boxes probably to always test the new way so that it doesn't break again and then see how to move forward with that so what I wanted to come back to is our what's cooking section of anybody there's a question what am I talking talking I'm coming to the question thank you I don't know if it's an ESC question but in the ESC minutes there is always those reports of failing test those are failed a lot and I was wondering is there a way to come back to go from to find all the instances where one test failed like for example I now have this accessibility that sometimes fails enough to appear here but is there a way for me to look at all the failings so I could see why they are failing yes but it's a little convoluted so you can search and get it for the failure category and then you see all the changes that failed with this error and then you have to take a good guess which one is the first one or which one is the most interesting one and as you mentioned with most of these failures those are not persistent failures so they happen in the decision and this is what makes it hard to check out the importance of the specific one because one failure can happen in just one change that has 10 different versions of a patch until the problem was fixed so those kind of reasons are also not reflected in this simple stats those are just blindly counting all the failures of the last week and not doing any further analysis on how widespread those were in terms of different people affected and also there can be the case that a series of patches is based on top of each other so if one of the first ones has an actual error that causes a test to fail all the other ones that have this specific commit as a parent will also fail even if those changes themselves would cause the failure so it's can be used as a trend and indicator for your own but if you see this error then you can assume it is a flaky test and you can resume the build and expect the next build to pass so this is the primary use of the cellist so to speak but yeah you can look in Garrett itself and see if it fails unfortunately it's not possible in Jenkins on the system there you cannot search for it but you can do the search using Garrett and go to the old build logs from there so the question was you can look for the failure test as it appears in the ESC method logs so it's not limited to one specific test but rather to make file target that runs the tests but yeah those terms that appear in the ESC minutes you can search for in Garrett and that works because Jenkins writes that that command into Garrett and we search through the commands okay yeah even something I learned a subsequent question is for how long the like the Jenkins results are kept for like if I'm looking for a chain a build log from last year will I find it okay yeah I repeated for the recording so the failing logs are kept for at least one year I think we don't even limit it currently but yeah at least one year for the failing logs but the successful logs get truncated and the build results themselves still available so you can see that the platform succeeded in building but the the log will be empty for those but the failing logs are kept in full thanks next question anybody you have a question that's great yeah I'll ask the question just so we get it recorded what the answer is there's a little recurring section where we have kind of highlighted bugs from QA which are most in need of looking at how do you pick the most how do you pick the bugs that you list there what's the criteria for appearing in that list you mean the highest okay so basically I check every week when a bug in Baxila has been changed the priority to highest then if that's done I just bring it to the ASC is it that the question why is changed to highest yeah and the same for there's another part after that one which is high severity bug of the weeks so it's the same logic so when a bug during the week from previous ASC meeting the current ASC meeting is changed to high then it's added to the minutes even more just wanted to add that this is a good way to also double check that those issues are really very important or high severity bugs because sometimes we have I mean even if changing the priority it's only limited to a group of people not like every user can do that but sometimes you see issues that has has been increased and the priority but then you see that well it's not realistic so then you just adjust it so it's just a way to for me to double check that those are real high issues or high risk issues okay so yeah our real world meetings Thursdays they also nominally always run for one hour and what sometimes happens is that Haiko comes up with such a long list of UX issues what we take up the full hour and everybody's fall asleep through half the hour but what often also happens is that Haiko has a wonderfully prepared three item lists of the most pressing UX issues and Miklos always steps in and says I know the fix for already and we're done with that part very quickly and like a whoosh go through the rest of the things as well and then there's no new pressing back of the week because we have introduced no new bugs and we're done in like 15 minutes and everybody closes their computers and goes back to whatever else they were happily doing and if there's no more questions today then I suggest we also close our computers and go back to coffee and whatever the sunshine happy hacking everybody thank you for joining us