| Time |
Nick |
Message |
| 00:04 |
|
Blockhead256 joined #minetest |
| 00:05 |
Blockhead256 |
bastrabun: Screenshots uploaded to discord will paste a link to IRC that contains a token that will expire in 24h or less. |
| 00:05 |
Blockhead256 |
so whoever reads the conversation any later than that, say on the logs website, will not be able to see the screenshot, since they aren't usng Discord to fetch a new token |
| 00:06 |
Blockhead256 |
much better to use an external image host |
| 00:06 |
MTDiscord |
<wsor4035> alternatively a bot could just curl/wget/etc the image |
| 00:06 |
Blockhead256 |
and put it on the logs website? |
| 00:07 |
MTDiscord |
<bastrabun> Another case where I'd love to have some reasonablly working forum |
| 00:07 |
MTDiscord |
<wsor4035> sure, or paste the new image link, etc |
| 00:08 |
Blockhead256 |
bastrabun: What's unreasonable about the forum? Works fine for me |
| 00:08 |
cheapie |
Down most of the time, slow when it's not |
| 00:09 |
MTDiscord |
<bastrabun> We're talking about the minetest forum, right? |
| 00:09 |
MTDiscord |
<bastrabun> Is there another? |
| 00:09 |
MTDiscord |
<wsor4035> 50 errors my old friend* |
| 00:09 |
Blockhead256 |
I don't think youse have used the Luanti forums recently, it's not as bad as it was. Still kinda slow |
| 00:09 |
MTDiscord |
<wsor4035> well, c55 potato has one less site to run now. soon 2 hopefully |
| 00:10 |
MTDiscord |
<bastrabun> Yes, ofc, I force myself to check at least once in a while and regret immediately afterwards |
| 00:10 |
Blockhead256 |
I suppose the way I use it is still not in touch really... I multithread all the posts I'm opening so by the time I click the last link, it's finished loading the first |
| 00:11 |
MTDiscord |
<wsor4035> same, its the only way i stay somewhat sane |
| 00:12 |
Blockhead256 |
well anyway, "Post your screenshots!" et cetera do remain an option.. you tradeoff speed for not posting to something commercial like Imgur or unaffiliated (and sometimes blocked) like catbox |
| 00:12 |
MTDiscord |
<bastrabun> Yeah, same. I click all of those that are new and I want to see, then do something else for a while, then return to view. |
| 00:15 |
|
freq joined #minetest |
| 00:15 |
freq |
Your Land |
| 00:15 |
|
freq left #minetest |
| 00:15 |
MTDiscord |
<bastrabun> Can't be that hard to set up an image host oneself |
| 00:15 |
Blockhead256 |
pffft |
| 00:15 |
Blockhead256 |
bastrabun: setting up isn't the hard part, it's making sure it doesn't disappear into the void! |
| 00:15 |
stephan48 |
having the bot reprocess(i.e. download and offer) images is a liability, since you can't trust what people will link, you might inadvertently host material you should not. better not :) |
| 00:16 |
stephan48 |
should not=which might require administrative action or could make you liable for distribution |
| 00:17 |
MTDiscord |
<wsor4035> i mean, i could say some vile stuff that might not be allowed in some country, say germany for example. |
| 00:17 |
MTDiscord |
<wsor4035> purely in text |
| 00:18 |
stephan48 |
yes but in this case the bridge op can claim, we only forward we don't log, wsor is now blocked forever. if you store logs somewhere publicly oh well ;) |
| 00:18 |
MTDiscord |
<bastrabun> That's usually the moment in the discussion where I look at how other projects do it. The 0ad people have a fairly modern and fast forum, the beyond all reasons people appear to be on discord mainly ... hm. |
| 00:18 |
MTDiscord |
<wsor4035> (this channel is logged btw) |
| 00:18 |
stephan48 |
i know :P |
| 00:19 |
MTDiscord |
<bastrabun> Doesn't help if the images disappear |
| 00:20 |
Blockhead256 |
let me level: I don't trust any community that's mostly on discord |
| 00:20 |
stephan48 |
if you store images/files someone can easily make you guilty of possession of some materials if they want, which often causes quite a heavier fine then someone being a bit dumb and saying stuff they should not in text. |
| 00:20 |
stephan48 |
Blockhead256++ |
| 00:20 |
stephan48 |
:P |
| 00:20 |
stephan48 |
sorry you discord people |
| 00:21 |
stephan48 |
easiest way is really to let people bring their own image host(imgur whatever) and outsource the policing and liability problem beyond "i might need to scrub logs" |
| 00:21 |
stephan48 |
which sucks to some degree |
| 00:22 |
Blockhead256 |
of course, matrix is inadequate for a lot of people. But, that's common in the libreverse (sure that can be a word) |
| 00:22 |
stephan48 |
dooont get me started on matrix please |
| 00:22 |
ireallyhateirc |
linkrot is cool though, people don't need to know what the conversation was about 10 days in the future |
| 00:22 |
MTDiscord |
<wsor4035> to be fair, i do have some concern with all the legal mess happening for the internet that could lead to only large corperate interests on it since they can absorb hits |
| 00:22 |
stephan48 |
discord is bad enough as it is but atleast it does not silently drop group chat messages. |
| 00:23 |
MTDiscord |
<bastrabun> Unfortunately I can't do that. I'm hell-bent on trying to host everything I can myself. Which happens to be how my villain origin story in minetest started |
| 00:23 |
stephan48 |
wsor, sadly thats the issue, people are idiots and while 99% behave, it just needs one bad egg to fuckup the person who spends their hard earned free time providing all the nice infra |
| 00:23 |
MTDiscord |
<wsor4035> hot take to the people on the irc side, discord exists and is popular because irc didnt keep up an actively fought/fights progress |
| 00:23 |
stephan48 |
and burn them out |
| 00:24 |
stephan48 |
how to compete with a multi billion dollar marketing campaign? :P |
| 00:24 |
MTDiscord |
<wsor4035> irc used to be the defacto chat on the internet, now discord, facebook messenger, telegram, etc are in that role |
| 00:24 |
Blockhead256 |
"yeah we know IRC sucks" -- active IRC user |
| 00:24 |
Blockhead256 |
cool so the status quo will prevail I assume |
| 00:25 |
MTDiscord |
<bastrabun> better ethics is not a feature. open stuff can only compete if its better. |
| 00:25 |
stephan48 |
its nice how people compare IRC - a system mainly driven by singular people driving open source projects forwards for decades to something like Discord which are so immensly funded with devs that they can implement almost anything in a few days(if agile would not be a blocker) |
| 00:25 |
MTDiscord |
<wsor4035> obligatory https://xkcd.com/1782/ btw |
| 00:25 |
MTDiscord |
<wsor4035> coughs in ircv3 and freenode/libera refusal to implement it fully |
| 00:26 |
ireallyhateirc |
bastrabun, ethics is a feature, most people simply ignore it though and succumb to megacorporations |
| 00:27 |
ireallyhateirc |
"I have nothing to hide" and other such dystopian mantras |
| 00:27 |
Blockhead256 |
I really didn't want to start another platform debate... |
| 00:27 |
MTDiscord |
<bastrabun> If ethics is a feature for you, then its a feature for you. Only the majority doesn't care. If I could host discord myself, I'd do so. I can't, so I gotta learn hosting matrix |
| 00:27 |
MTDiscord |
<bastrabun> Blockhead: It distracted nicely from the forum debate 😛 |
| 00:27 |
MTDiscord |
<wsor4035> i do need to try revolt or whatever it is at some point. see how it compares to matrix |
| 00:28 |
MTDiscord |
<wsor4035> matrix is just discord at home tbh |
| 00:28 |
stephan48 |
welllll |
| 00:28 |
stephan48 |
how about tmux to run weechat and irssi - the latter being a kind of bouncer for weechat? |
| 00:28 |
stephan48 |
i even just switched to weechat for weechat android! |
| 00:28 |
stephan48 |
(althrough i jumped from screen to tmux 5 years ago) |
| 00:28 |
stephan48 |
would discord provide a unified server view across all servers i would probably look at it more favoreably but this server switching is what kills me UX whise. |
| 00:28 |
stephan48 |
(not that matrix did this in any kind or shape better) |
| 00:28 |
|
stephan48 was kicked by ShadowBot: Message flood detected. Use a pastebin like pastebin.mozilla.org. |
| 00:28 |
ireallyhateirc |
lol |
| 00:28 |
|
stephan48 joined #minetest |
| 00:29 |
stephan48 |
wellllllll thank you ShadowBot |
| 00:29 |
Blockhead256 |
bahaha, your fingers are just too fast |
| 00:29 |
stephan48 |
@bastrabun, have fun and enjoy the pain when matrix.org silently rejects messages from your users :P |
| 00:30 |
Blockhead256 |
I'm lazy, I had a bouncer/logger once but it went down after a few weeks from a dc. I just have the IRC log website on my firefox new tab page |
| 00:30 |
MTDiscord |
<bastrabun> You're not telling me matrix isn't all paradise, rainbow and unicorns? Noooo! |
| 00:30 |
Blockhead256 |
> (not that matrix did this in any kind or shape better) |
| 00:30 |
Blockhead256 |
the (hypothetically) neat part with matrix is you have a choice of clients |
| 00:31 |
stephan48 |
yes so i found out |
| 00:31 |
stephan48 |
all suck |
| 00:31 |
stephan48 |
yes i should not complain when i am not willing to fix, but i want a messenger dammit. |
| 00:31 |
stephan48 |
not another life long project |
| 00:32 |
MTDiscord |
<bastrabun> I got an idea! We could make a protocol that unites them all! (waits for the approrpiate xkcd) |
| 00:32 |
stephan48 |
Standards! |
| 00:32 |
Blockhead256 |
anyway played any good Luanti games or mods lately? |
| 00:33 |
MTDiscord |
<wsor4035> https://xkcd.com/927/ |
| 00:33 |
stephan48 |
872? 972! |
| 00:33 |
MTDiscord |
<bastrabun> thankyou |
| 00:33 |
stephan48 |
welll off by one and slightly different order |
| 00:34 |
stephan48 |
my xkcd numbers got a bit rusty |
| 00:34 |
MTDiscord |
<wsor4035> blockhead: i3, xdecor, morelights, dreambuilder hotbar - probably some of most most prefered and well designed/built/whatever mods |
| 00:34 |
MTDiscord |
<bastrabun> waits for the big endian VS little endian joke |
| 00:35 |
MTDiscord |
<wsor4035> from a user point of view |
| 00:35 |
stephan48 |
we recently had that issue in #icecast, with embarresment we found we never compiled for a big endian system. stuff went foobar. |
| 00:35 |
stephan48 |
debian's packaging out of all the things did catch that |
| 00:36 |
Blockhead256 |
what's the deal with long hotbars? I wouldn't ever use one longer than 12, and even then, that's past where my left hand is ever going. |
| 00:36 |
Blockhead256 |
mouse scrolling hotbar selection.... ehhhh.... |
| 00:38 |
MTDiscord |
<bastrabun> We're using a 16 hotbar by default |
| 00:39 |
Blockhead256 |
you must either have a keyboard with 16 keys on the number row, use the F-keys or spare keys on the main letter blocks, .... or for some reason you're scrolling your mousewheel around a lot |
| 01:26 |
[ |
what's wrong with mouse scrolling hotbar selection? |
| 01:26 |
[ |
it's fine with a real scrollwheel (but is harder to use with a touchpad or trackpoint) |
| 01:27 |
Blockhead256 |
it's really slow compared to hitting numbers, especially numbers 1-6 |
| 01:27 |
Blockhead256 |
it would be better to have some system like factorio for easily swapping hotbars around than to make one huge one |
| 01:28 |
Blockhead256 |
my personal preference: numbers 1-6, then Shift+1-6 for the next 6 slots, then Ctrl+<Number> to switch hotbars |
| 01:29 |
Blockhead256 |
But I suppose at least wide hotbars work right now, against my hypothetical. Luanti's control scheme really could use some work... |
| 01:30 |
Blockhead256 |
Minetest Game's inventory is only 32 slots, so you'd only need 2 hotbars as well |
| 01:32 |
MTDiscord |
<greenxenith> I made a mod for myself to help when building that has an infinite inventory (expands when adding items) and a 16-wide hotbar with items (buttons) at each end to scroll through the inventory. Really handy, even if its not perfect yet. |
| 01:32 |
Blockhead256 |
intriguing. Definitely not the approach I'd use, but interesting |
| 01:33 |
Blockhead256 |
as a vim user I really think more programs need to embrace modality/prefix type shortcuts. Like if you could hit tilde or something and within 5 keystrokes get almost any registered item |
| 01:35 |
Blockhead256 |
first keypress would open the selection UI, then each subsequent one navigates the UI with all the letters and numerals exposed as categories, subcategories, until you get to an actual item |
| 01:36 |
Blockhead256 |
the navigation key sequences could be automatically arranged, hopefully with a deterministic algorithm |
| 01:36 |
Blockhead256 |
if any of you have used vimperator or tridactyl, you'll have an idea of what I mean |
| 01:37 |
Blockhead256 |
unfortunately, I don't think formspec can response to single keypresses, even with some hack like focusing a textbox, I don't think the client can be made to force-submit the form |
| 02:09 |
MTDiscord |
<warr1024> I had a similar idea for a bigger hotbar, in the editor for Klots. I made the hotbar 10 wide, since that's how many hotkeys I have for the slots ... and then I added a "tap aux" input to swap the hotbar with the row directly under it, so you can "bank swap" and effectively have a hotbar that's 20 wide, with up to 2 keystrokes to reach any slot. |
| 02:15 |
Blockhead256 |
fun fact: Luanti actually has 32 keybindings for hotbar slots |
| 02:20 |
MTDiscord |
<warr1024> less fun fact: I only know 10 of them |
| 02:38 |
Blockhead256 |
well, the other 22 aren't bound by default. First you'd have to find 22 free keys to use! |
| 02:45 |
|
Can0xfBows_ joined #minetest |
| 03:01 |
|
SFENCE_arch joined #minetest |
| 03:10 |
MTDiscord |
<mark.wiemer> random aside, someone mentioned GitHub being slow recently, I think they bungled the rollout of Issues 2.0 and I've been seeing issues ever since. Today I got my first unexpected error! |
| 03:10 |
MTDiscord |
<mark.wiemer> https://cdn.discordapp.com/attachments/749727888659447960/1333635322105565185/image.png?ex=67999c2e&is=67984aae&hm=7d6e3abd35ae75cbeee66d1ed6557613ae287cc6be9ba33a69514fd48bee4279& |
| 03:11 |
|
Lunatrius joined #minetest |
| 03:13 |
MTDiscord |
<mark.wiemer> https://github.blog/changelog/2025-01-13-evolving-github-issues-public-preview/ it is opt-out if it becomes a significant concern |
| 03:17 |
|
amfl2 joined #minetest |
| 05:00 |
|
MTDiscord joined #minetest |
| 05:04 |
|
book` joined #minetest |
| 05:08 |
|
calculon joined #minetest |
| 05:23 |
|
est31 joined #minetest |
| 05:38 |
|
Trifton_ joined #minetest |
| 05:52 |
|
Trifton_ joined #minetest |
| 05:57 |
|
est31 joined #minetest |
| 06:11 |
|
illwieckz joined #minetest |
| 06:28 |
|
Trifton2 joined #minetest |
| 07:02 |
|
SFENCE_arch joined #minetest |
| 07:40 |
|
diemartin joined #minetest |
| 08:15 |
|
amfl2 joined #minetest |
| 08:52 |
|
tarsovbak joined #minetest |
| 09:16 |
|
PoochInquisitor joined #minetest |
| 10:15 |
|
gregon joined #minetest |
| 10:41 |
|
Talkless joined #minetest |
| 11:06 |
|
Kimapr_ joined #minetest |
| 11:23 |
|
jaca122 joined #minetest |
| 11:38 |
|
SpaceManiac joined #minetest |
| 12:08 |
|
BuckarooBanzai joined #minetest |
| 12:25 |
|
kimapr joined #minetest |
| 12:46 |
|
Thermoriax joined #minetest |
| 12:47 |
|
Kimapr_ joined #minetest |
| 13:19 |
|
ireallyhateirc joined #minetest |
| 13:20 |
|
PoochInquisitor joined #minetest |
| 13:20 |
|
SwissalpS joined #minetest |
| 13:32 |
|
kimapr joined #minetest |
| 14:25 |
|
SFENCE joined #minetest |
| 14:39 |
|
SFENCE_arch joined #minetest |
| 14:41 |
|
SFENCE joined #minetest |
| 14:46 |
|
Desour joined #minetest |
| 15:19 |
|
emorrp1-irc joined #minetest |
| 15:28 |
|
Talkless joined #minetest |
| 15:31 |
|
fling_ joined #minetest |
| 15:44 |
emorrp1-irc |
heya folks, fwiw I've finished my review of the mods packaged in debian and 17 of them needed changing https://lists.debian.org/debian-devel-games/2025/01/msg00006.html |
| 15:44 |
MTDiscord |
<warr1024> The fact that debian is packaging mods at all is a problem... |
| 15:45 |
MTDiscord |
<warr1024> It might have made sense 5 years ago when Luanti had no built-in content distribution, but that's like nearly 2 whole debian releases ago... |
| 15:45 |
emorrp1-irc |
Warr1024: obviously we're gonna have to agree to disagree there, see the quote in the email for an example where contentdb isn't sufficient |
| 15:46 |
MTDiscord |
<warr1024> Lol, no thanks, you can agree to disagree if you want but I'm not interested. |
| 15:47 |
emorrp1-irc |
👍️ |
| 15:47 |
MTDiscord |
<warr1024> That argument about not being able to access the internet breaks the Debian package manager just the same as the Luanti one |
| 15:48 |
MTDiscord |
<warr1024> In any event, if Debian has decided to take on package management of a software ecosystem in a way that's at odds with that ecosystem's standard distribution practices, then it's also responsible for however that breaks... |
| 16:04 |
MTDiscord |
<warr1024> If external package managers really want to package Luanti content, it would probably actually be a good idea for them to source them from ContentDB to begin with. As long as they don't just leave versions frozen perpetually and actually periodically check for updates, it should be fine. But there are also a LOT of packages in CDB and it could easily be overwhelming... |
| 16:04 |
|
TheCoffeMaker joined #minetest |
| 16:05 |
MTDiscord |
<warr1024> External package managers will have to work to avoid becoming museums of historically interesting mods rather than actually useful-today packages. |
| 16:20 |
MTDiscord |
<bastrabun> Could we make deb packages from CDB releases automaatically? |
| 16:21 |
MTDiscord |
<luatic> I really don't think that is a job for the Luanti community. Frankly I don't know if it's worthwhile at all. |
| 16:24 |
MTDiscord |
<luatic> If the Debian community thinks there is tremendous value in having mods be Debian packages, they're free to throw their time at it, but it's definitely not something that "has" to be done and probably creates more problems than it solves. |
| 16:24 |
|
SFENCE joined #minetest |
| 16:29 |
|
Lunatrius joined #minetest |
| 16:30 |
emorrp1-irc |
luatic: agreed |
| 16:33 |
emorrp1-irc |
it gets more interesting when we have enough to make a larger game like voxelibre Just Work with an apt install, like erle wants to do |
| 16:34 |
emorrp1-irc |
bastrabun: probably, or at least a faster way to import mods to start with, most of the debian packaging can be templated (we've discussed a cdb2deb tool) |
| 16:34 |
emorrp1-irc |
but it still needs to go through human review to e.g. flag any hidden copyright issues (though I'm told contentdb is good at this too) |
| 16:35 |
emorrp1-irc |
and make them work with each other, rather than everyone forking for minor changes |
| 16:36 |
pgimeno |
IMO game packages make much more sense than mod packages |
| 16:38 |
emorrp1-irc |
one tends to need the other first... :) |
| 16:41 |
MTDiscord |
<warr1024> I believe it should be possible to "pin" a debian package to a specific release from CDB. CDB authors can delete old releases, but I don't think that it's possible for them to modify or substitute them or anything. That means that people can manually review packages and then pin that version until another manual review. The prospect of CDB removing old releases is nothing you wouldn't experience from other package managers that can have |
| 16:41 |
MTDiscord |
similar issues (or even get 451'd even if they don't expire versions) |
| 16:45 |
MTDiscord |
<wsor4035> > contentdb isn't sufficient i fail to see this |
| 16:46 |
MTDiscord |
<wsor4035> if your talking about the prision thing, then just bundle a zip of all cdb downloads when you send it in or something |
| 16:47 |
MTDiscord |
<wsor4035> > The fact that debian is packaging mods at all is a problem... but yeah, by far agree with this |
| 16:49 |
emorrp1-irc |
well it makes us happy :) |
| 16:51 |
MTDiscord |
<wsor4035> the reality is Debian fails to keep anything up to date, and ships out of date stuff all the time, this will be an issue if users install an out of date package from Debian, and then a mod from cdb that depends on that mod with a more modern version. the user will then be required to clear out the Debian version to get the right version. vs if they install from cdb, they get both of the correct versions (yes, there is the whole versioning |
| 16:51 |
MTDiscord |
of mods debate which relates to the github issue, etc) |
| 16:51 |
sfan5 |
the main menu could probably handle it better if you have a mod installed but it's non-updateable |
| 16:52 |
|
SFENCE joined #minetest |
| 16:52 |
MTDiscord |
<wsor4035> but then again, doesnt change much in just recommending to completely use cdb/use ppa or flatpack for an up to date version of luanti |
| 16:55 |
MTDiscord |
<wsor4035> re licensing, if Debian is shiping and ancient enough version of minetest game, the water texture is copyrighted. might have been sound licensing issues in mtg? not sure if that was mtg or a mod soup |
| 16:56 |
sfan5 |
<https://github.com/luanti-org/minetest_game/commit/bbc70c95ef0d211962060b842b8c614a7633633e> for reference |
| 17:05 |
pgimeno |
"<emorrp1-irc> one tends to need the other first... :)" I think the problem is the lack of game-local mods. Different mods with the same name in different games shouldn't be an issue (in particular, different versions of the same mod). |
| 17:07 |
ireallyhateirc |
isn't there any package-manager-agnostic software packing format what would allow for easy imports? |
| 17:15 |
MTDiscord |
<warr1024> "package-manager-agnostic software packing format" ... sounds like tarballs or zip files. |
| 17:16 |
emorrp1-irc |
... which are available from contentdb |
| 17:16 |
MTDiscord |
<warr1024> yep |
| 17:16 |
MTDiscord |
<warr1024> I'm actually kind of curious about what kind of use-cases luanti has in prisons. Like, is it just entertainment to help pass the time, or are there like educational uses or something? |
| 17:17 |
MTDiscord |
<warr1024> Is it just a matter of making free software available in prisons on the principle that free software should be available everywhere, or are there like prison-specific needs that it serves? |
| 17:18 |
MTDiscord |
<wsor4035> cdb doesnt do tarballs, just zips iirc? |
| 17:20 |
|
SFENCE joined #minetest |
| 17:20 |
MTDiscord |
<warr1024> Yeah, though there's not a ton of difference between them from luanti's perspective. We don't rely on permissions stuff like the execute bit that zip doesn't have, and while the "solid" compression of tarballs is nice (and we don't take advantage of the individual file access that zips offer) I don't think the difference between them is actually important in most cases. |
| 17:20 |
emorrp1-irc |
I'm not entirely sure, I assumed from the comment that it was entertainment, minetest isn't the only feature https://github.com/cyberitsolutions/bootstrap2020/blob/main/NEWS.rst#changes-in-soes-september-2023-since-august-2023 |
| 17:21 |
MTDiscord |
<warr1024> I have suggested that CDB offer solid archive format options (would require client support) to make it possible to shrink downloads, though it's not a high priority. You can actually hack it with the zip format by just putting a no-compression zip inside a high-compression zip. |
| 17:21 |
MTDiscord |
<wsor4035> luanti now has ztsd, i wonder if cdb where to ztsd compress packages if it would yield disk savings |
| 17:21 |
MTDiscord |
<wsor4035> zips on the fly for older clients |
| 17:22 |
emorrp1-irc |
ireallyhateirc: appstream would be good for all the metadata, IIRC it's the same as used by flatpak, but other sources can use the same data too |
| 17:22 |
MTDiscord |
<warr1024> as far as I know, the luanti ecosystem understands zlib (and thus could do .z and .gz), zstd, and zip. What it lacks is tar. You can do .zip.zstd, and get decent compression that way, it's just ... awkward. I think zip directory metadata may be at the end so you can't easily "stream" decompression, and you have to be careful to disable compression on "inner" archives to avoid defeating the outer compressor. |
| 17:23 |
emorrp1-irc |
https://github.com/luanti-org/luanti/blob/master/misc/net.minetest.minetest.metainfo.xml if you're unfamiliar |
| 17:23 |
MTDiscord |
<warr1024> The tar format has all this metadata that's not relevant to luanti packages, but then again, zip has its fair share of that too. |
| 17:24 |
MTDiscord |
<wsor4035> can xml just die already? |
| 17:25 |
MTDiscord |
<wsor4035> warr: yeah, was mostly just thinking out loud since cdb is getting big, save ruben some disk if possible |
| 17:25 |
emorrp1-irc |
wsor4035: there's a YAML variant |
| 17:26 |
emorrp1-irc |
but the appstream folks prefer xml |
| 17:26 |
MTDiscord |
<wsor4035> ill see you in norway :p |
| 17:26 |
MTDiscord |
<wsor4035> running go version 1.20 |
| 17:27 |
MTDiscord |
<wsor4035> to bad there isnt json at least |
| 17:27 |
MTDiscord |
<warr1024> https://github.com/luanti-org/contentdb/issues/496 |
| 17:28 |
MTDiscord |
<wsor4035> ah |
| 17:28 |
MTDiscord |
<wsor4035> anyone make the relevant engine issue or naw? |
| 17:28 |
|
gregon joined #minetest |
| 17:28 |
MTDiscord |
<warr1024> good question. I don't remember doing so myself, but even that, I'm unsure of. |
| 17:29 |
MTDiscord |
<wsor4035> if i recall maybe ill dig around tonight and see, else 🤷 |
| 17:29 |
MTDiscord |
<wsor4035> dont off the top of my head remember seeing one in the engine tracker |
| 17:29 |
MTDiscord |
<warr1024> Personally I'm pretty torn between .zip.gz and .tar.gz; I think each of them has different challenges. .zip.gz is easier to accidentally misuse and get bad results. .tar.gz would be harder to get merged due to the potential need for more dependencies. |
| 17:30 |
|
fling_ joined #minetest |
| 17:31 |
MTDiscord |
<warr1024> I'm uncertain whether there's a "streaming" workflow for .zip.gz or whether it requires the thing to be "spooled" somewhere. It might not be an issue though if the objective is JUST to unpack it onto disk, i.e. we expect the necessary storage to be available already. |
| 17:31 |
|
grouinos joined #minetest |
| 17:31 |
|
est31 joined #minetest |
| 17:31 |
MTDiscord |
<wsor4035> the web prefers Brotli, with fallback to gzip. chrome added support for ztsd and some sites/cdns (including cloudflare?) are starting to prefer ztsd |
| 17:32 |
MTDiscord |
<wsor4035> > One of the most exciting recent developments in web performance is Zstandard (zstd) — a new compression algorithm that we have found compresses data 42% faster than Brotli while maintaining almost the same compression levels. Not only that, but Zstandard reduces file sizes by 11.3% compared to GZIP, all while maintaining comparable speeds. https://blog.cloudflare.com/new-standards/ |
| 17:33 |
|
Can0xfBows joined #minetest |
| 17:33 |
MTDiscord |
<warr1024> The difference between compression algorithms isn't as big of a deal as the difference between solid and non-solid archives. The fact that it's possible to order files and take advantage of repetition between them is what nets you most of the savings from the proposal. |
| 17:34 |
MTDiscord |
<wsor4035> yeah, different content types |
| 17:38 |
ireallyhateirc |
Guix has an importer for CDB packages but the set of packages is small and outdated. Such importers still need plenty of manual cleanup. |
| 17:39 |
ireallyhateirc |
I think about just sending them a PR that would remove the outdated packages |
| 17:45 |
MTDiscord |
<rollerozxa> vo |
| 17:52 |
MTDiscord |
<warr1024> I'd think that if your objective is to bundle software to be deployed in an environment that won't have unfettered access to the internet for updates, just making a "deb soup" wouldn't be sufficient anyway; you'd probably want something more like a disk image where you can bundle configuration as well ... and in that situation, including packages installed via non-system package managers would make more sense rather than trying to convert |
| 17:52 |
MTDiscord |
them to system packages. |
| 17:53 |
MTDiscord |
<warr1024> I mean, Debian kinda "needs" flatpaks to be useful as a general-purpose "desktop" system already 😅 |
| 17:54 |
ireallyhateirc |
but at least bugs can be exploited in a stable and reproducible manner |
| 17:59 |
|
Verticen joined #minetest |
| 18:00 |
MTDiscord |
<warr1024> At the cost of not having access to all the hottest NEW bugs, yeah. |
| 18:27 |
|
fluxionary joined #minetest |
| 18:53 |
|
diemartin joined #minetest |
| 19:01 |
|
Krock_ joined #minetest |
| 19:02 |
|
sofar_ joined #minetest |
| 19:03 |
|
khimaros joined #minetest |
| 19:07 |
|
greeter joined #minetest |
| 19:16 |
|
Verticen joined #minetest |
| 19:19 |
|
SFENCE joined #minetest |
| 20:03 |
|
SFENCE joined #minetest |
| 20:26 |
|
sinvet joined #minetest |
| 20:34 |
|
dabbill joined #minetest |
| 20:42 |
|
jistr_ joined #minetest |
| 20:42 |
|
Meli joined #minetest |
| 21:53 |
|
YuGiOhJCJ joined #minetest |
| 22:01 |
ireallyhateirc |
I brought back to life this thing https://github.com/luanti-org/luanti/pull/15730 now looking for some brave devs to review :) |
| 22:02 |
ireallyhateirc |
it appears that the original author did a really good job to make this work, then disappeared and left some trivial issues unfixed |
| 22:03 |
|
Kimapr joined #minetest |
| 22:42 |
|
Sharpman joined #minetest |
| 23:00 |
|
est31 joined #minetest |
| 23:33 |
|
panwolfram joined #minetest |
| 23:35 |
|
Sharpman joined #minetest |
