Time |
Nick |
Message |
00:00 |
MTDiscord |
<mark.wiemer> https://github.com/minetest/dev.luanti.org/pull/141 for local automated a11y tests π they're failing and we'll need to change our theme to fix at least one issue, going through the others now. Will add to CI once they're passing, PR has details |
00:06 |
MTDiscord |
<wsor4035> 1. please rebase your pr 2. dont see us changing the theme (at this point) 3. waiting to see actual issues |
00:06 |
MTDiscord |
<mark.wiemer> Oh, did we change stylesheets recently? |
00:06 |
MTDiscord |
<mark.wiemer> https://cdn.discordapp.com/attachments/926231483155378176/1330689927045840988/image.png?ex=678ee511&is=678d9391&hm=46ee5ed6f4c8d9679f7859690d9eeaf5ab3c1a11d7304813b813e1e6595b4e44& |
00:07 |
MTDiscord |
<wsor4035> no? you have prefer light mode aka burning your eyes out |
00:07 |
MTDiscord |
<mark.wiemer> 1. will do 2. yeah hence me calling it out 3. the issues are in the PR, they're verbose, working on making them cleaner Β I think it's because Edge is set up a bit weird, had to install a chromium browser to get a11y insights for web, it doesn't seem to be respecting my default system light/dark setting |
00:09 |
MTDiscord |
<mark.wiemer> This should give a good feel for it |
00:09 |
MTDiscord |
<mark.wiemer> https://cdn.discordapp.com/attachments/926231483155378176/1330690555562295327/image.png?ex=678ee5a7&is=678d9427&hm=da1235a145757747f9748ce197d96cb0dc2fde92ec069848b96caf276e9e0a2f& |
00:09 |
MTDiscord |
<wsor4035> well the whole notice section is getting burned eventually, so who cares |
00:14 |
MTDiscord |
<mark.wiemer> The short answer is that these are easy fixes, I should have some stuff ready in a few hours worth of work, probably midday tmrw. But the sickness is returning so I'm gonna take a nap for now. Feel free to follow the link I sent and work on it if you like, it's fun stuff π I literally teach this at Microsoft so happy to help and/or answer any questions from folks π |
00:23 |
MTDiscord |
<wsor4035> https://ux.stackexchange.com/a/147684 sums it well |
02:23 |
MTDiscord |
<mark.wiemer> Correct, the v2 algorithm isn't perfect, I'm OK pinning the color contrast issue as a "known false positive" for the current theme, there are other failures though that need investigation |
02:24 |
MTDiscord |
<mark.wiemer> Andrew Somers and I have actually talked about the current version of APCA. It's pretty good but still incomplete, so it's not a perfect baseline either. Overall, I think the current color contrast is OK, though there is room for improvement. However, there are other a11y issues to focus on in the meantime |
04:14 |
MTDiscord |
<mark.wiemer> https://github.com/minetest/dev.luanti.org/pull/142: Improve DX, add failing a11y tests. Should make our lives a good amount easier, esp for new contributors or those switching OSes like me, lol. Got sick of installing Hugo-extended-blankity-blank every time I needed it. Figured some other improvements were in order too |
04:17 |
MTDiscord |
<wsor4035> do we really need npm+nodejs hell just for the core tool written in go? |
04:19 |
MTDiscord |
<wsor4035> plus its from some third party |
04:19 |
MTDiscord |
<mark.wiemer> How else do you want us to manage dependencies? Already we have a checker that uses Hugo 0.140 and a builder that uses Hugo 0.137 |
04:19 |
MTDiscord |
<wsor4035> glances at docker/podman |
04:20 |
MTDiscord |
<mark.wiemer> And we are requiring devs to manually install a global version on their machine and remember scripts written in one kinda random file. package.json is pretty recognizable go-to. Not sure what's so hellish about this, basically switching from Hugo as the global tool to Node as the global tool |
04:22 |
MTDiscord |
<wsor4035> see what other people think about this, but i really dont know |
04:22 |
MTDiscord |
<mark.wiemer> What's wrong with docker and podman? I'm not up to speed. If the concern is security, I think we're OK, NPM requires 2FA nowadays and the package gets ~90K installs a week, it's decent enough. The source code is plenty good too, if you want me to just fork it and republish under @luanti/hugo-extended or whatever for security |
04:23 |
MTDiscord |
<wsor4035> nothing is wrong with docker/podman, i was suggesting them instead of shoving everything under nodejs+npm |
04:23 |
MTDiscord |
<mark.wiemer> Ah, gotcha. I feel like Docker at least would be a bit much, I see it as heavier than Node in this case. Not familiar with podman |
04:23 |
MTDiscord |
<wsor4035> podman is docker but fully open source by redhat. with some variations |
04:26 |
MTDiscord |
<wsor4035> pr: https://github.com/minetest/dev.luanti.org/pull/142 core thing for comment: thoughts on making npm+nodejs central to build/dx/etc people i would like input from: (on the team): @greenxenith @rollerozxa Β (external): @Luanzai Β any maybe lars, but i wouldnt say that they web is an area of expertise i can say for the others, no offense |
04:27 |
MTDiscord |
<wsor4035> maybe im just old man yelling at clouds. see what others think π |
04:28 |
MTDiscord |
<mark.wiemer> I'm still a bit under the weather but will try to monitor the chat as I can. I'm curious for alternatives that would also support accessibility testing, spell-checking, etc. Β Happy to listen to yelling. It means you care π I picked Node because it's what I'm most familiar with, happy to look at other options, just felt natural for a web-dev project as well |
04:29 |
MTDiscord |
<wsor4035> i dont mind using npm for tooling and stuff(value add stuff). just dont want the core to be in npm+nodejs |
04:30 |
MTDiscord |
<wsor4035> when i started this dumpsterfire project, a part of picking hugo was that it wasnt in npm+nodejs ecosystem |
04:31 |
MTDiscord |
<mark.wiemer> I'm curious what's wrong with Node in your view |
04:32 |
MTDiscord |
<wsor4035> dependency hell (whole trees, just adding tons of stuff for the sake of it - see leftpad for example). also churn/turn over. how fast stuff is in and out. |
04:33 |
MTDiscord |
<mark.wiemer> Just making sure we're on the same page: https://en.wikipedia.org/wiki/Npm_left-pad_incident ? |
04:33 |
MTDiscord |
<wsor4035> yeah, just how dumb it was to even have that as a dependency in the first place, rather than the whole it getting removed, etc |
04:34 |
MTDiscord |
<wsor4035> if you want depressing numbers, look at the weekly downloads is-odd and is-even gets |
04:36 |
MTDiscord |
<mark.wiemer> makes sense, the ecosystem definitely suffers from "don't reinvent the wheel, just depend on it" syndrome a good amount. I used Docker a good amount... back in 2019. Would it reasonably solve this problem without introducing unreasonable complexity of its own? |
04:37 |
MTDiscord |
<greenxenith> Maybe I am missing something but it doesnt look like node is actually used here, just npm |
04:37 |
MTDiscord |
<wsor4035> im more so just running docker containers, etc in low numbers. but you could use docker compose to define versions, etc |
04:38 |
MTDiscord |
<greenxenith> Ok, I see the a11y tests, so thats node I guess |
04:38 |
MTDiscord |
<mark.wiemer> https://en.wikipedia.org/wiki/Peacenotwar dang the world is a wild place |
04:39 |
MTDiscord |
<wsor4035> iirc there is a little js that is running in hugo-extended to call hugo |
04:39 |
MTDiscord |
<mark.wiemer> Yeah all the stuff in package.json installs dependencies with npm via Node and then each script is ultimately wrapped by a Node process. It adds a good half-second to execution time but overall allows for consistent environment between devs, which is the main goal. Obv there are other ways of getting consistent env, hapy to listen |
04:40 |
MTDiscord |
<mark.wiemer> happy* |
04:42 |
MTDiscord |
<greenxenith> Its a well-tested build system that we dont have to maintain, and the more components we add the nicer it will be to have quick scripts. I am more against dockerizing a development build system (why add more complexity?). |
04:43 |
MTDiscord |
<mark.wiemer> I guess I don't see docker as a solution. It allows us to create our own very consistent environments, but we'd still need some software package for spell-checking, a11y testing, etc. I'm most familiar with npm packages so I went ahead and used those π€· |
04:43 |
MTDiscord |
<greenxenith> Is it sort of weird to have npm/nodejs overhead for a Go system? Yeah. Is there a better set of "pure" scripts that could replace it? I doubt it. Are we likely to already have node set up anyway? Also yes. |
04:44 |
MTDiscord |
<greenxenith> Docker is meant for reproducable production environments, not dev environments. Sure I suppose you could make it fit, but it seems like unnecessary complexity to me. |
04:45 |
MTDiscord |
<wsor4035> docker compose is dev + production.... |
04:45 |
MTDiscord |
<greenxenith> The + production is the key part there |
04:45 |
MTDiscord |
<greenxenith> We dont use Docker for the final build |
04:46 |
MTDiscord |
<wsor4035> i mean, we could :juanchi_face: ||yes, this is stupid i know|| |
04:46 |
MTDiscord |
<greenxenith> Just so I understand this perfectly: What advantages does this Node build system have over the classic Go system? |
04:47 |
MTDiscord |
<wsor4035> anyways, im perhaps being old man yelling at trees. i personally prefer not to need npm to run the project. just have hugo and be off to the races. then you can have npm for value add such as spellchecking, ally, etc |
04:47 |
MTDiscord |
<greenxenith> Is there anything preventing us from allowing both? |
04:47 |
MTDiscord |
<wsor4035> im not mark, but i think the main reason they shoved it in npm was so everyone had the same version |
04:48 |
MTDiscord |
<greenxenith> In fact, what about this system prevents you from running hugo serve/build as normal? |
04:48 |
MTDiscord |
<wsor4035> also aside, im mildly concerned (probably unfounded) about hacing el random middleman between the project and hugo |
04:49 |
MTDiscord |
<wsor4035> nothing. though in the pr it is push hugo through npm as the one supported/offical way |
04:50 |
MTDiscord |
<mark.wiemer> Two main reasons of putting Hugo under npm: 1. consistent versions 2. easy installation / uninstallation (no need to browse to GitHub releases and find whatever version is best for your current system, just npm i |
04:50 |
MTDiscord |
<greenxenith> Hm, is the spell checking and other verifiers running on PR CI? Or is it only run locally? |
04:50 |
MTDiscord |
<mark.wiemer> 3. if spell-checking is under npm and Hugo isn't folks will be confused why they have to install two global tools |
04:51 |
MTDiscord |
<wsor4035> the latter currently, the former once all wiki mess is imported |
04:51 |
MTDiscord |
<mark.wiemer> Yes, once spellcehck issues and others are fixed I plan to add them to PR CI |
04:51 |
MTDiscord |
<wsor4035> <shills other people than me doing imports> |
04:51 |
MTDiscord |
<mark.wiemer> Hey man I'm just here for the long-term DX! Plus I'm sick, you can't shill at me π |
04:52 |
MTDiscord |
<wsor4035> tbh, maybe a compromise is "officially" pushing people the npm route first, but supporting/allowing/whatever just hugo by itself as long as people have the version from package.json |
04:53 |
MTDiscord |
<mark.wiemer> Yeah absolutely folks can still do that. Happy to update the docs accordingly, let me push a draft |
04:53 |
MTDiscord |
<greenxenith> I dont see any harm in including the raw hugo serve instructions as a secondary option, especially since with PR CI on the way |
04:53 |
MTDiscord |
<greenxenith> (PR CI is important as it guarantees checks will be run for contributions, whether they ran them locally or not) |
04:55 |
MTDiscord |
<greenxenith> And once PR CI is implemented, I would suggest the team commit policy should require a PR for anything non-trivial to pass CI, whether we require approvals or not |
04:55 |
MTDiscord |
<mark.wiemer> If you have already cloned you can fetch the submodules as such: Β bash git submodule init git submodule update --remote Β This project uses Hugo to build the site and various Node packages to test it. You can install Hugo locally as a Node package for convenience, or find the relevant hugo-extended build from GitHub releases to install Hugo globally if you prefer. Β To install and run via Node: Β bash npm install # Installs packages |
04:55 |
MTDiscord |
needed to build and test the site npm start # Builds and serves the site on a localhost port for manual testing Β To run globally: Β bash hugo server # This is the command internal to the `npm start` command |
04:55 |
MTDiscord |
<wsor4035> i mean, thats already in practice/applied/common sense. i guess it wouldnt hurt to document it tho |
04:56 |
MTDiscord |
<greenxenith> Did we ever get around to documenting dev policies? |
04:56 |
MTDiscord |
<mark.wiemer> Should definitely require PR. That's something that can just be enforced in GitHub itself. It's easy to "set auto-merge" right when you open a PR and have it do its thing without further attention π |
04:56 |
MTDiscord |
<wsor4035> https://dev.luanti.org/about-this-site/rules/ |
04:56 |
MTDiscord |
<greenxenith> Mind, I said for non-trivial things |
04:56 |
MTDiscord |
<greenxenith> Im not sure I want to make a branch for 2-char fixes |
04:57 |
MTDiscord |
<mark.wiemer> Gotcha, fair enough, no complaints from me (I'll let you enforce PRs once someone accidentally pushes a breaking change π ) |
04:57 |
MTDiscord |
<wsor4035> i mean, could probably document something along the lines of im lead, greenxenith is backup, docs team owns the repo or some crap, but π€· |
04:58 |
MTDiscord |
<greenxenith> Unless theres some --force-push-skip-the-pr-i-know-what-im-doing flag or something |
04:58 |
MTDiscord |
<wsor4035> iirc you can require ci pass, and then if your an admin github will let your force it past with a button iirc |
04:58 |
MTDiscord |
<mark.wiemer> You can allow certain contributors to bypass branch rules but I'm not sure if there's a fancy flag for it |
04:58 |
MTDiscord |
<mark.wiemer> That is, I'm not quite sure what you're asking :/ |
04:59 |
MTDiscord |
<greenxenith> Actually, is there a way for git to automatically create a branch for commits |
04:59 |
MTDiscord |
<wsor4035> you lost me now |
04:59 |
MTDiscord |
<mark.wiemer> Anyway, are we OK with the changes I proposed above? OK with moving hugo into an npm package for CI? |
04:59 |
MTDiscord |
<greenxenith> If PRs are required, all changes have to be in a PR (hence, branch) |
04:59 |
MTDiscord |
<wsor4035> ah |
05:00 |
|
MTDiscord joined #minetest-docs |
05:00 |
MTDiscord |
<mark.wiemer> Yes, if you protect main branch and make an edit via GitHub, when you try to commit, GH will automatically say "can't commit, click here to commit to a new branch named mark-wiemer-patch-1 and open a PR" |
05:00 |
MTDiscord |
<wsor4035> if your centralizing on npm, why not just jam that into npm scripts as well? |
05:00 |
MTDiscord |
<greenxenith> I said git, not GitHub |
05:00 |
MTDiscord |
<mark.wiemer> Jam what? |
05:00 |
MTDiscord |
<wsor4035> sorry, git submodules context |
05:00 |
MTDiscord |
<mark.wiemer> ah, sry, misread. You could try a git alias but I haven't explored it |
05:00 |
MTDiscord |
<mark.wiemer> Maybe for a future PR? submodules are wonky even for experienced folks like me π |
05:00 |
MTDiscord |
<greenxenith> Anyway, this is off in the weeds |
05:01 |
MTDiscord |
<greenxenith> Re submodules: Is there a disadvantage to including our themes directly in-repo |
05:01 |
MTDiscord |
<mark.wiemer> I'll push these changes for review for now, lmk your initial thoughts |
05:02 |
MTDiscord |
<wsor4035> perfer it to be mroe path a, path b type options, also say see package.json for require hugo version |
05:02 |
MTDiscord |
<wsor4035> so its more explict |
05:03 |
MTDiscord |
<wsor4035> in hugo, you usually either use hugo modules or submodules. we have dependabot on the repo to make prs for us when there is updates |
05:03 |
MTDiscord |
<wsor4035> https://gohugo.io/hugo-modules/ |
05:05 |
MTDiscord |
<greenxenith> You know, the npm script could be configured to install submodules when missing |
05:05 |
MTDiscord |
<wsor4035> your late π https://discord.com/channels/369122544273588224/926231483155378176/1330763822977388616 |
05:05 |
MTDiscord |
<greenxenith> Ah, that context |
05:05 |
MTDiscord |
<greenxenith> Understood |
05:07 |
MTDiscord |
<greenxenith> Why are we using hugo-extended? |
05:07 |
MTDiscord |
<wsor4035> its the defacto one, has all the css, etc garbage in it |
05:07 |
MTDiscord |
<wsor4035> iirc the theme requires it |
05:07 |
MTDiscord |
<mark.wiemer> The theme does require it |
05:07 |
MTDiscord |
<greenxenith> Mark made it sound like its some sort of hassle to install it manually ... I just run the site with whatever hugo version go isntalled |
05:08 |
MTDiscord |
<wsor4035> then you have extended iirc |
05:08 |
MTDiscord |
<wsor4035> unless your not on windows/macos/linux |
05:08 |
MTDiscord |
<mark.wiemer> How did you install via Go? I did not see that in the docs. They've been updated in the past few weeks, looks like, but still a bit of a pain compared to npm i: https://gohugo.io/installation/ |
05:09 |
MTDiscord |
<mark.wiemer> Xenith living the Haiku life lol |
05:09 |
MTDiscord |
<wsor4035> hugo extended allows image conversion + sass/scss |
05:09 |
MTDiscord |
<wsor4035> side note, scss/sass really should die |
05:09 |
MTDiscord |
<greenxenith> https://cdn.discordapp.com/attachments/926231483155378176/1330766169875152937/image.png?ex=678f2c13&is=678dda93&hm=91fbbc2f40246c3f240f6a4db18218546646daa56994d2ed092391a457a47a79& |
05:10 |
MTDiscord |
<mark.wiemer> Yeah, honestly my problem is unique, in the past 3 weeks I've switched from Windows to NixOS to Linux Mint, all very very different installation processes. Both for Hugo and for everything else. In any case, I think it's important to ensure devs have consistent versions, sooooo |
05:10 |
MTDiscord |
<mark.wiemer> https://github.com/minetest/dev.luanti.org/pull/142 ready for review π |
05:10 |
MTDiscord |
<greenxenith> Mint and NixOS both have extended hugo in their package managers |
05:11 |
MTDiscord |
<wsor4035> ...yeah, unless your on some weird platform or botched a manual install you will have extended |
05:12 |
MTDiscord |
<wsor4035> your ci seems to have issues mark |
05:12 |
MTDiscord |
<mark.wiemer> Hey man I tried my best π it was a hectic holiday season for Mr Mark |
05:12 |
MTDiscord |
<greenxenith> Do you want "perfectionist GreenXenith" or "just get it done GreenXenith"? |
05:13 |
MTDiscord |
<wsor4035> https://cdn.discordapp.com/attachments/926231483155378176/1330767115463954565/image.png?ex=678f2cf4&is=678ddb74&hm=32291f773e2e7cd240dbcffd875d052e34ed4d5e8723e1fac9692548879a0d8d& |
05:13 |
MTDiscord |
<wsor4035> https://github.com/minetest/dev.luanti.org/actions/runs/12861434459/job/35854830225 |
05:13 |
MTDiscord |
<wsor4035> https://tenor.com/view/why-not-both-why-not-take-both-gif-11478682 |
05:13 |
MTDiscord |
<wsor4035> :juanchi_face: |
05:14 |
MTDiscord |
<greenxenith> That would be "looks fine, merge it ... but if you want some feedback: ..." |
05:14 |
MTDiscord |
<mark.wiemer> I want pragmatic GreenXenith who is open to splitting future work into future PRs because I'm sick and scared of feature creep π |
05:14 |
MTDiscord |
<greenxenith> We have feature creep? |
05:14 |
MTDiscord |
<mark.wiemer> Pushed fixes to CI, thanks |
05:15 |
MTDiscord |
<wsor4035> yeah, on that, really shouldnt have pushed build and ally in one pr, but idrc |
05:15 |
MTDiscord |
<mark.wiemer> I'm worried perfectionist GreenXenith might ask for more features in this PR |
05:15 |
MTDiscord |
<greenxenith> That wouldnt be perfect |
05:15 |
MTDiscord |
<mark.wiemer> Yeah sorry @wsor Β I know it's already bloated |
05:15 |
MTDiscord |
<greenxenith> That would be inefficient |
05:15 |
MTDiscord |
<mark.wiemer> hahahahaha you got me |
05:15 |
MTDiscord |
<mark.wiemer> you do you I'll be here to listen π |
05:15 |
MTDiscord |
<greenxenith> Perfectionist GreenXenith wants half those sections reworded |
05:15 |
MTDiscord |
<mark.wiemer> Feel free to leave some comments and suggestions π |
05:16 |
MTDiscord |
<wsor4035> idk, the one bit of feature creep i would suggest is jamming git submodule init && git submodule update in install, and git submodule update on start or something like that |
05:16 |
MTDiscord |
<mark.wiemer> Your suggestion has been heard. I'd rather not π |
05:16 |
MTDiscord |
<greenxenith> Meh, that can be added to a later PR |
05:16 |
MTDiscord |
<mark.wiemer> Not in this PR, at least π |
05:16 |
MTDiscord |
<wsor4035> also the docs should mention that you need git |
05:16 |
MTDiscord |
<mark.wiemer> I feel like git is a given, the source is hosted on GitHub, after all |
05:16 |
MTDiscord |
<greenxenith> Is there a way I can put comments on the whole file that a PR changes instead of the diff |
05:17 |
MTDiscord |
<mark.wiemer> Not that I know of, you can highlight sections or just click the first line I think |
05:17 |
MTDiscord |
<greenxenith> :( |
05:17 |
MTDiscord |
<mark.wiemer> Something like this not sufficient? Hover to reveal the button |
05:17 |
MTDiscord |
<mark.wiemer> https://cdn.discordapp.com/attachments/926231483155378176/1330768261754654720/image.png?ex=678f2e06&is=678ddc86&hm=d5e991163642f660a6df41bdf4f38ae8e27b98590524118a960057fe6e4b6c4a& |
05:17 |
MTDiscord |
<greenxenith> It might be faster to just alter wording in a second PR |
05:18 |
MTDiscord |
<mark.wiemer> your call |
05:18 |
MTDiscord |
<greenxenith> The problem is github breaks up your diffs. I want to comment on an entire section of the markup |
05:18 |
MTDiscord |
<mark.wiemer> added a comment highlighting a range |
05:19 |
MTDiscord |
<mark.wiemer> ah, yeah you can do that, just expand until the ranges merge: |
05:19 |
MTDiscord |
<mark.wiemer> https://cdn.discordapp.com/attachments/926231483155378176/1330768530282516490/image.png?ex=678f2e46&is=678ddcc6&hm=582f2f41c61ad8f8fcc1a4b498a722600ded06276443fe21e346993efe130818& |
05:19 |
MTDiscord |
<greenxenith> Aha |
05:19 |
MTDiscord |
<wsor4035> so the real question is if roller is introducing php into the project next week or not as a require component(since using npm/nodejs, could use php wasm) |
05:19 |
MTDiscord |
<mark.wiemer> Then click and drag to comment on a range instead of a single line: |
05:19 |
MTDiscord |
<mark.wiemer> https://cdn.discordapp.com/attachments/926231483155378176/1330768702022353010/image.png?ex=678f2e6f&is=678ddcef&hm=cd5738d6958104cfae88d92f0ff8cef2db6f46b1023c1e33af253ec7b3b298b8& |
05:20 |
MTDiscord |
<mark.wiemer> @wsor not sure if joking :/ |
05:20 |
MTDiscord |
<mark.wiemer> @greenxenith you know about the suggestion button? |
05:20 |
MTDiscord |
<wsor4035> i mean, idrc. be fine by me |
05:20 |
MTDiscord |
<greenxenith> ? |
05:20 |
MTDiscord |
<mark.wiemer> I guess we'll cross that bridge when we get there π |
05:21 |
MTDiscord |
<mark.wiemer> https://cdn.discordapp.com/attachments/926231483155378176/1330769109079687269/image.png?ex=678f2ed0&is=678ddd50&hm=740c0ebc971084190c470dbf640f72d3de61f48b76528ed647b27f3e63130a7f& |
05:21 |
MTDiscord |
<greenxenith> Oh, nice |
05:21 |
MTDiscord |
<wsor4035> anyways, shall see what roller says tommorrow about the question. probably after that ill review it |
05:21 |
MTDiscord |
<greenxenith> Thats dangerous, I might just rewrite the whole file |
05:21 |
MTDiscord |
<mark.wiemer> GH is really good π I'm no expert but they've imported a lot of features from Microsoft's ADO π |
05:22 |
MTDiscord |
<wsor4035> unless its new/fresh microsoft, generally that would be terrifying |
05:23 |
MTDiscord |
<mark.wiemer> haha ADO definitely has its issues, but stuff like multiline comments and the suggestion button were there long before they were added to GH, IIRC. That said, GH definitely wins because it's just lighter, IMO |
05:23 |
MTDiscord |
<wsor4035> >azure so new unless they built on top of some old ms dumpsterfire |
05:24 |
MTDiscord |
<mark.wiemer> Azure DevOps is the most recent name of what used to be "Visual Studio Online", to get a feel for how old it is. Believe it or not, we use it a good amount at my job at Microsoft π |
05:24 |
MTDiscord |
<greenxenith> Is it out of scope to suggest alterations to the instructions that you didnt change |
05:24 |
MTDiscord |
<mark.wiemer> I'm happy to listen, I may say "out-of-scope", but if they're relevant and easy I'm good at clicking "accept suggestion" π |
05:24 |
MTDiscord |
<greenxenith> Awesome |
05:28 |
MTDiscord |
<greenxenith> Are comments in inline bash acceptable practice |
05:30 |
MTDiscord |
<greenxenith> .... GitHub wont let me add comments to expanded sections |
05:30 |
MTDiscord |
<greenxenith> Cant be bothered now |
05:31 |
MTDiscord |
<mark.wiemer> hm it is a bit finnicky but it should work, otherwise you're welcome to just say smth like "line 10-12 should change to this ..." |
05:32 |
MTDiscord |
<greenxenith> The option to add a comment does not exist for those sections, and you cannot drag into them for multi-lin either |
05:32 |
MTDiscord |
<greenxenith> If I actually care enough later Ill just make a PR to alter wording |
05:32 |
MTDiscord |
<greenxenith> I have more important things to do right now |
05:33 |
MTDiscord |
<mark.wiemer> yeah you're right I can't get it to work, my bad. Oh well :/ |
05:43 |
MTDiscord |
<mark.wiemer> @greenxenith should I stay up for your comments or will it be a while? I'll be happy to check tmrw morn of course |
05:44 |
MTDiscord |
<greenxenith> I am not going to bother with comments, Ill save it for a PR if I decide it matters enough |
05:44 |
MTDiscord |
<mark.wiemer> OK, so are we good to merge? |
05:44 |
MTDiscord |
<greenxenith> Was this about your PR or something else |
05:45 |
MTDiscord |
<mark.wiemer> I think @wsor was talking about my PR, the question being "is Node OK for this?" basically |
05:45 |
MTDiscord |
<mark.wiemer> no rush, gonna sign off for the night. Thanks for reviewing π |
08:08 |
MTDiscord |
<rollerozxa> as long as I can completely ignore it and still just build with hugo |
10:30 |
MTDiscord |
<mark.wiemer> You can completely ignore it and still just build with Hugo π |
14:22 |
|
Desour joined #minetest-docs |
16:31 |
MTDiscord |
<wsor4035> re #engine discussion, might not be a bad idea for us to have a "features" page https://www.ogre3d.org/about/features |
17:12 |
MTDiscord |
<mark.wiemer> @wsor , good to merge #142? |
17:28 |
MTDiscord |
<wsor4035> what witchcraft is the first line? no formatting for it. was looking to remove the whitespace at the start |
17:28 |
MTDiscord |
<wsor4035> https://cdn.discordapp.com/attachments/926231483155378176/1330952149525332008/image.png?ex=678fd948&is=678e87c8&hm=40a8b0dba7148f7f7c0273a9532a3affe71da3c16924bd1122bbc1c7e851ed6c& |
17:28 |
MTDiscord |
<wsor4035> https://cdn.discordapp.com/attachments/926231483155378176/1330952149852749937/image.png?ex=678fd948&is=678e87c8&hm=3362078755612344e3ee730dc55249a0967d496c9bed4d2b5d3265da3e561629& |
17:28 |
MTDiscord |
<wsor4035> havent looked over it yet tbh |
17:28 |
MTDiscord |
<wsor4035> ill take a look now |
17:30 |
MTDiscord |
<wsor4035> . btw, colon syntax in scripts, is that some sort of magic thing? like if you do just npm test it runs all three scripts, and then you can run them individually with the full path? havent seen that before |
17:30 |
MTDiscord |
<rollerozxa> indented markdown codeblocks? :thinksmirk: |
17:31 |
MTDiscord |
<wsor4035> is that a thing? if so weird? |
17:31 |
MTDiscord |
<rollerozxa> pretty sure that's even the original way you did codeblocks until GFM fenced codeblocks took the world by storm |
17:31 |
MTDiscord |
<wsor4035> ah.... thats cursed |
17:32 |
MTDiscord |
<wsor4035> i see why gfm wins yet again |
17:32 |
MTDiscord |
<rollerozxa> lol yeah |
17:32 |
MTDiscord |
<wsor4035> anyways, just edited it to a normal code block to remove the white space |
17:32 |
MTDiscord |
<wsor4035> dunno even why i was slightly being ocd about it |
17:33 |
MTDiscord |
<wsor4035> . not a blocker mark, just curious |
17:34 |
MTDiscord |
<wsor4035> roller: did you want to have a look before i merged 142? |
17:34 |
MTDiscord |
<wsor4035> otherwise merging in 5 |
17:36 |
MTDiscord |
<rollerozxa> I would like for a link to the hugo installation page to be kept for people who choose to install it manually without using NPM |
17:37 |
MTDiscord |
<rollerozxa> actually wait I just saw that is still there, just below the other stuff |
17:37 |
MTDiscord |
<rollerozxa> yeah it's fine, merge it |
17:37 |
MTDiscord |
<wsor4035> i was about to say lol |
17:44 |
MTDiscord |
<wsor4035> https://github.com/minetest/dev.luanti.org/issues/145 |
17:44 |
MTDiscord |
<wsor4035> pokes roller with 126 (since we probably dont want me doing it) |
17:46 |
MTDiscord |
<rollerozxa> I haven't touched the android version in a while, so grorp would probably be a better candidate for it |
17:46 |
MTDiscord |
<wsor4035> anyways, hopefully today/tonight i will have more time + motivation to do more wiki imports |
18:02 |
|
Desour joined #minetest-docs |
19:46 |
MTDiscord |
<mark.wiemer> @wsor what are the remaining wiki imports, for reference? |
19:46 |
MTDiscord |
<wsor4035> its one of the pinned issues |
19:47 |
MTDiscord |
<wsor4035> https://github.com/minetest/dev.luanti.org/issues/15 |
20:02 |
MTDiscord |
<mark.wiemer> lmk if you have quick docs for how to migrate, common pitfalls, etc. Happy to take a look at this, I thought it was done but clearly I was wrong! |
20:09 |
MTDiscord |
<wsor4035> url to markdown site, touch up output minimally, migrate images to asset folder, make sure you have the full version else roller will fight you |
20:10 |
MTDiscord |
<wsor4035> then annoy roller to "redirect" from the current wiki |
20:10 |
MTDiscord |
<mark.wiemer> full version? Like all the contents of the page? |
20:10 |
MTDiscord |
<wsor4035> in context of the previous, images |
20:10 |
MTDiscord |
<mark.wiemer> "url to markdown site" like there's a site where we put in the URL and it outputs markdown for the site? Or something else? |
20:11 |
MTDiscord |
<wsor4035> you cant just download the one in the page, thats small scale, you have to click on it and download the full version |
20:11 |
MTDiscord |
<wsor4035> media wiki is dumb |
20:11 |
MTDiscord |
<wsor4035> https://urltomarkdown.com/ |
20:11 |
MTDiscord |
<wsor4035> you could use pandoc or whatever to convert straight from the wikitext, but that takes longer and is more annoying |
20:12 |
MTDiscord |
<wsor4035> also the wiki takes 8 years to load, so the less clicks you have to do, the better |
20:13 |
MTDiscord |
<mark.wiemer> 1. Use https://urltomarkdown.com/ for convenience 1. touch up output minimally 1. migrate full images to asset folder (manually downloaded, not just from the site) 1. open and merge PR 1. ask @rollerozxa to add redirect from MediaWiki to dev.luanti.org Β sounds good π |
20:14 |
MTDiscord |
<wsor4035> i doubt you have an account on the wiki, but if you do, make sure your logged out, else every page will take 16 years to load instead of 8 |
20:15 |
MTDiscord |
<wsor4035> also have to have the file in our format, and add an alias for the old url |
20:15 |
MTDiscord |
<mark.wiemer> Gotcha, I need the account to download the images. Is it possible to create a new account still? |
20:15 |
MTDiscord |
<mark.wiemer> oh wait I don't need an acct, misread, my bad |
20:16 |
MTDiscord |
<rollerozxa> you don't need an account, you probably won't be able to get one in time anyway |
20:16 |
MTDiscord |
<mark.wiemer> what is "our format"? Hugo-flavored markdown? |
20:16 |
MTDiscord |
<rollerozxa> URL format is lowercase kebab case |
20:16 |
MTDiscord |
<rollerozxa> and then add an alias for the old mediawiki format URL |
20:16 |
MTDiscord |
<wsor4035> https://github.com/minetest/dev.luanti.org/blob/master/content/calinous-law.md?plain=1#L3-L4 |
20:46 |
MTDiscord |
<mark.wiemer> was having some issues with images, is this correct? Β 1. migrate full images to static/images folder (manually downloaded, not just from the site) |
20:49 |
MTDiscord |
<mark.wiemer> is optipng or similar required at this point, or will that be a separate step? |
20:55 |
MTDiscord |
<mark.wiemer> Test PR π https://github.com/minetest/dev.luanti.org/pull/146 |
21:19 |
MTDiscord |
<wsor4035> merging in 5 |
21:25 |
MTDiscord |
<wsor4035> cc @rollerozxa π |
21:27 |
MTDiscord |
<wsor4035> probably webp at some point |
21:44 |
MTDiscord |
<mark.wiemer> thanks π |