 Good, and that's how. Thanks for coming. It's nice to have another annual meeting. I think 2008 was the first one. It's kind of improvised. It's taking place once a year. It's good to see the group this size again. So, Alan already volunteered to take notes. Copy document. Dominic is watching IRC. We have at least three or four, you know, three remote participants. So, let's say hello to Damian, Alex and Fritko. Okay, I think I know most people, but maybe not everyone else knows everyone in the room. So, for those who would like, you can, well, introduce yourself. One sentence, who you are. What's your relation to Pearl, love, hate? I work in a job where I have exposure to Pearl, and I like it. So, also, I have one of the few packages which I occasionally have. Hi, I'm Dom. I bring the look-off to the Pearl Vitepter package with Nico, who's on IRC. I'm Arwen, I'm 18 years old, and I am a Pearl addict. I'm Alex Likert, extra one on IRC, ABE on the New York operations. Holding Pearls instead, I don't know, 20 years. Worked in a company who did a lot of Pearl stuff, and joined the British Pearl group, like, I don't know, I think. Maybe 2009, and that kind of thing. Bought some packages, like, dipsums into the track team, taking them over from inactive upstream, or well, inactive developers, but also doing other stuff inside the region. But with regards to language, love, it's mostly Pearl. I've actually been doing Pearls since certainly more than 20 years. And doing system administration, that's how I got into Pearl. I was a bit disappointed by what Pearl was doing, and I'm quite motivated again since about a couple of years, and made Pearl show how Pearl could actually be a model language again. So I'm motivated, and I joined, but he just missed that call. My name is Lucas, and I'm here. And I contribute to using Pearl models since the beginning of this year, and I joined the first team about two or three months ago. I like Pearl. The heating is busy. That's good. Can I go on? I'm Florian, FSFS on IOC. I work at the Berlin Pile University. I'm a citizen, and I'm doing some power coding. Well, I was a bit inactive over the last year or so with my life, but I'm back now. That's pretty good. Today with Kramner, and these days I mostly am some kind of heat consultant for the Pearl team, and I've also recently started inviting Pearl again because I'm extending that help of a member of Heldrecht. I don't hate it. My name is Salvatore. I've been in the group since a couple of years. At day work, at work. As well as a GPH doing system-able estimation of the day. I first got in from the phone, so it was the language we used to use. I'm on the dance on IOC. Pearl is likely the only program in the number chain group as a citizen. I'm a dance contributor. I just packaged an open PGP applet on top of you a few days ago. I joined the Pearl team some day, and here I am. Second row, if you want to say something, you don't have to approach. I'm just a user. Ronda, I took care of a few Pearl-related packages outside the Pearl team over the time. I was very happy that the Pearl group won, so I could hand them over. In my former workplace, I did program a lot in Pearl, and I still like it a lot. I'm David Herrmann from IOC, and I've been a member of the Pearl team since a couple of years. And uploading both packages is what I do in my free time. I'm kind of weird-looking. Okay, so we have two people here who are in the group. Which is a good sign, sometimes. Okay, as for the agenda, I wrote some things which came to my mind in this document, in no specific order. I've written down a lot of new group sessions, which we had in IOC. Sprints, 325-22, depth camp, status of the team. And, well, the question is if we want to start some specific projects in the next year. Does anyone have other topics for the agenda, which we should mention separately? Does anyone have any strong preference on the order of the topics, or should we just go down as they are now? Okay. So, low-hanging fruit sessions. It wasn't an idea actually coming in three, I think, three years ago. I think because of the good experiences it takes to remember correctly. So, one year ago, we said we would have a monthly meeting on IOC, on the 21st of each month, which is today, by the way. We alternated the times. One month at 17 o'clock UTC, and then the next month at 22 o'clock UTC. Yeah, and if you go down in the copyright document, I made some statistics. Statistics is a bit weird. Just look at the reports and rules and things. So, there was one meeting which we all forgot about. It was in June. And for the other meetings, there were between four and... I have between two and six people. So, for the... Okay. Okay, so that's the amount of participants in the month. Is there someone to leave the line here? There's this screen we're drawing. So, what's maybe interesting is that the meetings at 5 p.m. UTC had an average five or five and a half participants. The meetings at 10 p.m. UTC had an average 1.6 or 2.2. So, that's the amount of people who were there, but didn't say anything. So, participants, which was clearly that... Well, the earlier meetings had more participants, probably because all participants and most participants were from Europe. And I also looked, okay, how many people participated, how often, but also what we discussed about the topics. It's quite funny to see that in six of the ten meetings we talked about sprints in some way, like organizing an apartment in Barcelona. There was work on RC-PACs. There was various kind of work on different QA. Questions, transition, general packaging in one meeting was only on the quotation mark. Socializing was also a very nice meeting. Yeah, so, the question is if other people who participated or could not participate have any impressions of these sessions, which are worth noting down, and then if we continue and how we continue. Any good ideas? Yeah, and I think the statistics show that we should continue them at always the same time, I guess. I wonder if we should take something like 18 UTC instead of 17, but mostly because it would be a little bit closer to the previous late date. But that's just an idea or not. So for most of the middle European living people, there are people that would be 8 p.m. instead of 7 p.m. But then again, not sure if that's maybe 7 p.m. is fine and 8 p.m. would be a dinner time for some of us, so just an idea. But I think we should focus on always the same time. That's also easier to get into some ICAR file or similar. It matters on which day it was, so we decide every day on 21 to be honest. But it didn't check the week days. I actually like that kind of way because at least locally we really have to take care if we add some other regular events locally and not take that weekend where this event is or this event. So I think most of the local events are based on weekdays. So it's like Monday is Nino's Meetup in Middletooth, Tuesday is Debian Meetup, Wednesday is Cairo's Computer Club, Meetup Thursday and Friday is Alternating. It looks Meetup so, well, yeah. So you're saying a bit of that variation is good? Yeah, variation is good. So everyone has a chance even if he's booked out every Thursday or so. So I think what we have heard so far is that people want to continue the meetings. I also agree that going back to one time instead of Alternating might make things easier. I mean, I also can think of going back a bit. I guess the question is what I've done and Nico is saying, because they are one out of the east of us. I don't know how this for you, because for example, what would be a good time for you? I think this is a great time for me. So then Express is? Yeah. What for which time? Nico says that Tuesday night. So David? I feel like I shouldn't vote because I just didn't really find time in the last year. Yeah. So you should decide without any extra constraints. So Axel proposed 18 UTC? Mm-hmm. Okay, cool. Speaking about ICAD files, there is an ICAD file which has the dates so far. Yeah, but there are still enough calendar systems where you're planning both then. I know you have similar constraints or heavy constraints. So I didn't manage to put it into either a 900 or YOLA properly, so I'd be happy about one single time. Okay. Okay, yeah, one single time level update. I think we'll set the impression that it might help if we're going about this a little bit more seriously in terms of writing invitation a few days before and then also maybe not really like setting an agenda but making suggestions about things that we should work on. Okay, this time let's focus on bugs that we haven't looked at before and this time let's look at this or that and that better. Rather than meeting up there, see who's there and whether we're going to work on the norm. So maybe this would like invite others who would want to learn about things or also things that some things are important and should be pushed. You know, I think we have this push on hardening plans, adding hardening plans to all of any packages. Not work quite well here. So maybe that's an idea. That works if someone takes care of putting together such a thing more if people remember to say update between the two or many two sessions and then before the session one would look if there are public questions that people have to address. I think a reminder mail is necessary last year in three and for me try to send a reminder we didn't always manage. If you put this also down at the note that it doesn't have to be me if someone else volunteers to set up a contract. Exactly. It's one point. That's the point. Okay, who sets up a contract? Can we do that? As Tim was an idiot a week before and this is also the invitation to think about the agenda and people then can add ideas for the agenda on the Wiki. On the Wiki page there will be a belting flash and an hf which is quite empty currently because we only have the dates of the last year but we can use this for preparing the meeting. Okay, good. Next topic, sprints. How did you like Barcelona? It was very nice, very nice. Sprint was good too. So I think we spent, well we spent a good two full-length day sessions working on stuff which was in the report and it felt very productive. Darla Breeze, Sider's Nico. I don't know how it compares with DebtCamp because I've never been to DebtCamp. Was it more productive? I think most of us at DebtCamp were not only doing the pearl sprint. So the pearl sprint here was like five days or even seven days. And I think the topics were much quarter and rather dynamically decided on the first day while we had quite a clear plan what we want to do in Barcelona. I like Barcelona a lot too, it was too short I think. But then again well. The sprint here was more... So I guess the question is doing... Do we want to have more sprints in the future? We won't always have such a conveniently located DebtCamp so I suppose for some people going to a regional European sprint will always be easier than going to wherever DebtCamp is this year. I wonder if for example usually pearl events would be a nice surrounding for a sprint. So like jump pearl workshop, Swiss pearl workshop, yet another pearl conference, wherever. Just as an idea. That would kind of make it an extended trip for people. Either that or trying to do work during the conference which probably wouldn't work. I mean the advantage of the seven people in Barcelona was that there were seven of us in one room. Not distracted. That was what I think different from here. So I've been sort of running a sprint during DebtCamp and I can certainly agree that there's lots of distractions. I mean the people were only available during DebtCamp so it wasn't really a choice for me. I think just in terms of me keeping the ball rolling and actual work done nearly as effective as either a separate sprint. On the other hand it's great to have a sequence of meetings just to keep talking over issues. So for that kind of thing, if you had a sort of policy that's less hacking and more thinking and decision making then I think that could work pretty well during DebtCamp. Just because you have time off to sort of subliminally process things and when you come back. So Alex says that conferences and sprints are very different goals and I guess maybe I could add that it's worth keeping. Okay, so a sprint during the next year is something we would like to plan, that's it. Any rough ideas about locations or about autumn or spring or tournaments? I mean Mayish time seems to work quite well. I mean that would work quite well for me. But sometime not too close to either DebtCamp I suppose. I mean maybe we should ask people fairly soon what their general thoughts are and try to collect some sort of consensus when people have a chance to look at their kind of own punt I guess we're talking about somewhere in Europe. Do we have any bits for the location? By the numbers it should be insert. No womb which we can probably use for that may cost like one of the Swiss firms per day but the KS Computer Club and the Linux User Group once share a room which they usually also went to other communities if there's a slot free. It would be a suitable location. We'd need someone with a key because the outside door is usually locked pretty forward but there's no place to sleep there locally so we'd need some other place to sleep. We're going to provide a venue in Oxford at the UK and see the rooms. Alex points out that Mayish is a good time because it's around 10pm for releases. That works well actually. We've got a lot of good O5-22 work done. It's true, no? Yeah? Are you sure you're not going to submit it? No, I'm just just going to put it down. Okay, so if you want to work on this on a mailing list then someone should kick this off who wants to write the starting name on the mailing list. Should I take that? I can take that. Perfect, thank you. Okay, so that's going to do the starting name. Can you hold on just for a minute? I'm trying to reconnect to Patti and let's give him a try. There's something in left-clicking. You're out. No, we'll flip through. Yes. Copy into the web browser. We open the file and keep the old one open. Let's see. Oh, that's right, that's the first one. So should we carry on? Yes, thank you. So I just wrote this transition tracking bug to the release team which summarizes the date of O5-22 which can be summarized as mostly there are a lot of trivial bugs in architecture all packages that can be fixed, probably many of them by team members. It wasn't strictly related to O5-22 but the removal of some obviously make-maker implications caused quite a few build failures. We need help with modpel because there's no particular movement from upstream to fix that and that's likely to block the transition altogether. So if anyone feels like taking a stab at it, the bug number is, I think, on that list. It didn't look like a particularly complicated problem to fix for somebody new enough about the terms which I don't sadly. Otherwise, I think we're just waiting for the GCC transition to finish. So yeah, in terms of the team... Oh yeah, there is another Govee document which is helping to track work on these issues which is in the Teams Perl folder. So that's a good place to make down if you're working on something. I think back to the modpel stuff, I'm not 100% sure but I think there was already some future regression that manifested for packages that were behind the torch and it was actually at the end it was laid by some other solutions which was actually not working or something like that. So removal may be a possible option which you also said there but if it may be first to consider I don't know, just thinking about it. I would be very nervous about the new modpel from Debian at this point. I mean it's got 10,000 installations according to the podcom. It's slightly above the average. I mean it may not be being deployed for new mods these days but yeah, I don't know, it'll be bad for Debian. I don't know if anyone else disagrees. I don't know if they'll care about it. People will be able to. And there are many reverse dependencies I haven't checked how many but... I think there was one on the thing. There was this Libcoron old package which we probably need to decide whether we're going to just remove it or work to import the artificial change which probably makes it work again with the 4552. Is there some helpful upstream like with EGNPC in the past? Yes, there is a Github repo somewhere and I forget who did it now. And I think it says this is probably working but not quite released. Okay, so if someone wants to look at that that would be a good thing. It all works out to find it but we don't have to have long discussions. Cool. Okay. So we... we enter once, two thirds of the process. Oh, and there are 14 minutes left. Yeah, that camp we already mentioned some impressions from that camp and the sprint stream camp. So for those of you who haven't been here we met in the morning at 1.30 p.m. after breakfast discussed what everyone had done and what they plan to do and to stay and then everyone went ahead and worked a lot together so that was like that camp and the coordination part. One interesting topic is a topic which we talked about last year in Barcelona. It's very fruity. It's called that cherry. Yeah, that cherry is a... I just covered that camp that basically makes it possible to not use kilts for managing patches. We're still doing 3.0 kilts packages and there's a WPatchestive actually with kilts patches but basically that directory is managed by DeptCheri and as a package I can... I'm already forgetting but I don't have to use kilts for like when upstream does an update and then only takes half of my patch or changes things in the vicinity and then the patch doesn't apply to me anymore but I don't have to use kilts anymore but I can use git mechanisms to update the patch. It's a lot more smooth and comfortable. I was... Yeah, during that time I found out about DeptCheri and did a bit of polishing on a few of the scripts around it so that they actually moved beautiful from what they were looking before and also I uploaded two packages where I transitioned to DeptCheri which are HDNetMasonPerl and NetOpenSSHPerl and I'm planning to do a few more if I find some that are relatively frequently updated so that we can actually get some experience using it and see what the edges might be and how it can be used. So can I get those names again HDNetMasonPerl and... NetOpenSSH You should recognize them by the repeat of source files which I left there I'm trying to explain a few bits but the idea is that actually if you don't want to know about DeptCheri you can actually work with kilts patches still it just requires that the next person when I'm next touching the package to see that you did that and reintegrate those patches into DeptCheri So one more question occurred to me which is after watching this year's de-git talk is are your experimental packages in a shape where uploading the de-git will work? I'm not sure if you would want to but I think they have this git work tree matches exactly the source package unpack which I think is all you can So if somebody was really fired up about that then in the next two days I'd be willing to sit down talk about it over the next year we can investigate this as well so this might or might not be a benefit but it's an interesting tool that would be nice to at least optionally be able to use I guess it's way too far to take any decisions or to make any huge revolution or changes in the workflow but it's good that you work with it and now we have David, Nico and you three persons who know something about it So yeah, continue experimenting with it and report to the next sprint which is probably the way to go Okay, anything else from DeptCamp? Yeah, just a second So one thing I planned to do here and did was to move all the LinkedIn checks we had in the package pearl tools package inside the package pearl LinkedIn profile which makes sense for non-team-maintained packages too I moved them over to LinkedIn proper they were included as of engine 2.536 we also optimized a few things with regards to performance one of these optimizations went amok and currently has false positives on all pearl packages but Niels plans to do another upload 2.537 with that one fixed and get like for five days now so that should not bother us too long anymore Yeah, those tests are already removed from the package pearl tools package so there may be a short time where we don't have the proper LinkedIn checks currently but that should be over quite soon we renamed many of those texts to be more suitable for use outside the pearl community so they got some pearl dash added in front of so and well yeah so if we have overwrites in our packages we might need to rename them too actually LinkedIn has now support for renaming texts but that was introduced after I imported the text so that's not yet in there but I may be able to fill that in after I think it's not worth it not worth it It's only now in terms of text and I think there are hardly any overwrites for this experiment well I know I added some but I also know how to fix them and LinkedIn will warn about overwrites not hitting anything anyway Okay, cool I also run duck over all our team-managing packages but I actually haven't found time to analyze the log so the log is online and people they'll be on feel free if you find something in there to fix it if I have time the fine time I will probably work on these and well yeah no problem Okay, it's the last sentence I said that I would write a report for that camp and I'm still willing to do this but after I have a moment I have slept for three hours and then volunteer to translate to translate together No, that's uncounted Yeah, yeah You are our Justin Bieber Maybe Okay, we have six more minutes The question of how active is the group how many members do we have whatever comes as every year since I forgot how I counted the commits last year and now put this into a script which is in our script repository Oh, that doesn't look so good Still too small So the main answer is there were 58 people who made at least one commit between ten minutes ago and a year and ten minutes ago There were 57 when I checked two hours before so that's interesting I think last year we still had 77 Yeah, I'm not sure what to make And this is also sorted by who made how many commits I mean it's not about the high score but the interesting thing is we have I think 11 people who made at least 100 commits and then three who made at least 50 or between 1500 commits and well then the long thing somewhere starts So the question is how would happen with this is there anything to worry if there isn't a worry what are we going to do about it I don't know So I'm curious what fraction of the uploads are Greg or these days are you still uploading most packages or other people I think we have some statistics and we still wrote that I checked the commits that's there but he only lists the top ten of all time and there are only two guys left in the team for last year that's I think Gregor and I forgot and Damien or Dominic probably See, but this is about up a lot of packages per team So for 2015 well that's only half a year Gregor sends out but I put the link in the domain just a second so we can add team status The question was is that the person in the uploader's field or the person signing the changes part Yes, so I'm interested in the two uploads as from the app scripts I was looking at the people in Damien Chainsaw which is also interesting but more or less covered under the gift commits information I think there's no doubt that if Gregor was supposed to disappear then we would have a problem So I consider stealing Gregor's guitar just to make sure that he didn't develop any other hearties Should I put that in as a resolution? Anyway, let me put this in a positive way then say thanks Gregor again Thank you round and I would like to thank Lucas especially for the little two packages uploaded in the last week You really hit the high score for the packages per week and looking at your statistics it's already on top three and with only being with us for a year or so less than a year So welcome, okay Good, we have one minute left I guess if something to be said about this membership thing then we have to take it to the mailing list We talked about documentation for a new company Is there something for the remaining minute from this list of potential projects at the bottom where someone says yeah, we should tackle this and I will kick it off at the mailing list What's the status with the signature verification stuff? I think that would be very good to have I'm not sure if Seaman actually makes those signatures available So this relates to the discussion, the talk which is a talk that the R5 will perhaps soon support verification ship and verify detached signatures of cables C-PAN style signatures have a signature file inside a signed signature file crashes inside the disk I think that the signatures may also be separated may also be distributed separately from that But that's not really the use-gun kind of signature No, I don't really know if there is any way that we can make that I guess that would probably be more package-pulled tools workflow, if indeed the agreement is needed Actually, very funny to the panel I think the question is are there actually the streams signing their packages? Yeah I think they have to sign from so probably, don't they? Because I I'm not supposed to authenticate it It's just from all these seven passwords In fact, it's less than that Okay, so this looks like an interesting topic where one minute of a time Does someone want to make this I'd like to talk with DKG about it because I think he has the best insight on that topic anyway so I'd really like to see maybe support for that in use-gun Too far on it from what I see Niko says, auto-package test coverage has improved tremendously Thank you, everyone Adam Zem says, thank you for being such a great team Yeah, okay Stop here before we get kicked out So yeah, thanks to this great team See you