| -queuebot:#lubuntu-devel- New binary: lubuntu-installer-prompt [riscv64] (kinetic-proposed/universe) [0.2.0-0ubuntu1] (no packageset) | 00:40 | |
| arraybolt3[m] | Simon Quigley: At the suggestion of vorlon on #ubuntu-devel, I've made a modified version of lubuntu-installer-prompt that installs to and works from /usr/libexec rather than /usr/bin. Is this a change you'd be interested in integrating? If so, I'll submit a PR with the changes. | 00:43 |
|---|---|---|
| arraybolt3[m] | (I also have some packaging changes that go with.) | 00:43 |
| Eickmeyer[m] | Dovetailing on that, does vorlon understand the rationale that went into the decision to use calamares instead of collaborating on ubiquity? Because, to me, it does seem to be a worthwhile discussion. | 00:57 |
| arraybolt3[m] | Eickmeyer: I'm not even sure what the rationale is, I just know that it is, so... | 01:00 |
| arraybolt3[m] | But you went for Calamares, so you probably understand the "why" very well. For me I'd go with Calamares just out of preference since I'm more familiar with it. | 01:00 |
| arraybolt3[m] | I remember something about feature-richness being a problem, that's it. | 01:01 |
| Eickmeyer[m] | I understand the why for Ubuntu Studio. The overarching "why" is probably more multifaceted than that. | 01:01 |
| Eickmeyer[m] | To be honest, Ubiquity is more feature-rich and better integrated in the overall Ubuntu ecosystem. | 01:01 |
| lubot | [telegram] <lynorian> I think I remember stuff of like kde version of ubquity pulling in a ton of stuff making a preetty big iso even bigger | 01:01 |
| Eickmeyer[m] | @lynorian Not really. It's really just a qt frontend, no kde deps at all. The biggest problem I had was that it's hard-coded to Kubuntu and any changes had to go through Foundations, which meant a lot of waiting for weeks for a small change to get through. | 01:03 |
| tsimonq2 | Eickmeyer[m]: Well, one of our long-term goals is to fix that | 01:04 |
| tsimonq2 | Calamares is faster and has much cleaner code | 01:04 |
| tsimonq2 | And very well tested heh | 01:04 |
| Eickmeyer[m] | Simon Quigley: You're not going to get an argument from me, but it seems like we've got a bit of a community rift, which is very problematic from where I sit as you can imagine. | 01:05 |
| tsimonq2 | Respectfully, I... don't care. Technically it's a sound decision | 01:05 |
| tsimonq2 | I don't care if I have to become the damn upstream maintainer myself :P | 01:06 |
| kc2bez[m] | I don't see how it can be a rift when ubiquity itself is planned to be sunset. | 01:07 |
| Eickmeyer[m] | Dan Simmons: Because, from a community standpoint, collaboration should be tried first. So, I can see where vorlon's viewpoint is coming from. | 01:08 |
| Eickmeyer[m] | But, remember, I'm just playing devil's advocate here. It's a worthwhile discussion to have. I don't want to be argumentative for the sake of argument. | 01:10 |
| arraybolt3[m] | I remember trying to configure Ubiquity to try to fix a bug, I couldn't even find the config files, I couldn't understand the logs, it was a mess. | 01:10 |
| Eickmeyer[m] | Ubiquity is what happens when you add-on to something and then add-on to it and add-on to that. It's a mess. I agree. It's like most of America's airport terminals. | 01:11 |
| Eickmeyer[m] | It wasn't exactly built to be scalable, yet, here we are. | 01:12 |
| arraybolt3[m] | Though TBH that's what tons of stuff in Linux is. I mean, that's what APT seems like to me. It's slick to the user and a mind-shattering mess to the developer (at least it was to me :D). | 01:12 |
| Eickmeyer[m] | There's a reason it's being actively replaced, but in the meantime, it has to be maintained. | 01:12 |
| Eickmeyer[m] | In my dayjob, I've had to become fairly intimate with it in order to debug it. | 01:14 |
| arraybolt3[m] | (For the record I like APT and love packaging, but wow was it hard to learn and the docs were a disaster. Thankfully Simon Quigley was willing to basically spoon-feed me through the process until I was able to wrangle it somewhat well.) | 01:14 |
| kc2bez[m] | <Eickmeyer[m]> "Dan Simmons: Because, from a..." <- I can see that point. I think the thing I see is that we are 4 years in with Calamares at this point, switching back to ubiquity seems counter productive. | 01:14 |
| Eickmeyer[m] | kc2bez[m]: I'm not advocating that at all, and neither is he. | 01:15 |
| Eickmeyer[m] | I think he's a little "hurt" for lack of a better term? | 01:15 |
| kc2bez[m] | If we were to try to switch to anything I would think the flutter installer would be the choice. I don't see that as ready yet and flavor modifications seem to be a little further out. | 01:17 |
| kc2bez[m] | And that is only if we chose that. | 01:17 |
| Eickmeyer[m] | No, the flutter installer (code-named Canary, I believe) is far from ready. | 01:17 |
| arraybolt3[m] | kc2bez[m]: There's the problem. If we don't have general-purpose use (like flavor modifications) as part of the goal from the word "Go", I think we're going to get a mess. | 01:17 |
| arraybolt3[m] | Calamares was designed to be distro-agnostic from square 1, and so far it's working fantastic. Ubiquity wasn't so much AFAICT. | 01:18 |
| Eickmeyer[m] | arraybolt3[m]: I have on good word from Ken VanDine that flavor customization is one of the major goals. | 01:18 |
| arraybolt3[m] | OK, then that seems like a good thing. | 01:18 |
| arraybolt3[m] | It would also be nice if we could come up with an installer that was robust. Ubiquity and Calamares both are fragile - mess around too much and the thing crashes or things go wrong. | 01:19 |
| arraybolt3[m] | Hopefully Flutter will be able to withstand a user who keeps going back and then forward again. | 01:20 |
| Eickmeyer[m] | arraybolt3[m]: Ubiquity was, and wasn't. It was designed for Ubuntu and Kubuntu, the only two flavors at the time (both equally and commercially supported by Canonical at the time). | 01:20 |
| Eickmeyer[m] | arraybolt3[m]: Flutter is the language it's being written in. I believe the working name is Canary right now. | 01:21 |
| arraybolt3[m] | Right, forgot what the name was. | 01:21 |
| arraybolt3[m] | And I missed where you said it earlier. | 01:21 |
| arraybolt3[m] | Simon Quigley: Also, while we're on the topic, is that change to /usr/libexec a welcome change? I only have a few more minutes for today, so if I'm going to publish it and submit the PR I should probably know now. | 01:22 |
| arraybolt3[m] | Simon Quigley: OK, I submitted the PR, feel free to reject it if that's not what you want to do. | 01:29 |
| arraybolt3[m] | I'll tackle the packaging change (removing the missing manpage overrides) tomorrow, I'm out of time for now. | 01:30 |
| arraybolt3[m] | Thanks for everything, and I'll probably see you guys tomorrow night! | 01:30 |
| arraybolt3[m] | PR link: https://github.com/lubuntu-team/installer-prompt/pull/2 | 01:30 |
| ubot93 | Pull 2 in lubuntu-team/installer-prompt "Switched to using /usr/libexec rather than /usr/bin" [Open] | 01:30 |
| === Roberalz[m] is now known as Rober[m] | ||
| === Rober[m] is now known as Roberalz[m] | ||
| LeoK[m] | Signing off until AM.. | 18:26 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!