No.388
I was thinking that imageboards are good candidates for
columnstore indices but who wants to use SQL Server for that. Was thinking that each thread deserves it's own database table perhaps, but something like that increases the issues vichan has with table searching.
No.389
hm, i guess that's not needed anymore with the big engines... MySQL excluded of course
https://mariadb.com/kb/en/mariadb-columnstore/ No.481
>>387What userscripts do the people here use?
No.482
>>4814chan X and yentext functionality, so I can
¥ yenwith impunity
No.486
i had fun with my the primeagen mascot going in front of everything on the site using the custom CSS
#mascots {
z-index: 9999;
position: fixed;
}with
https://haiji.kissu.moe/test/src/1712059215978.pngbut its time to retire him
No.487
>>481I made a userscript for this site that loads full threads instead of having pages. Seems to work well
No.488
>>481User scripts are hard to get working on ReactJS sites. I set up an application API so you could send some custom events in to modify the SPA's state, but they're not really fleshed out or tested
No.1200
following documentation to the letter but feature not working. I hate APIs
No.1201
Got myself a token, got myself a URL, put em together and get thrown onto the HTML log-in page!
No.1202
ugh, in the token authentication step you can get a valid looking token by not filling out the URL properly and it never tells you
No.1203
>>387Are client-side rendered webpages the end game? Since having the power/flexibility to render your own contents using simple networking to the server (json) allows you to do anything you want.
No.1204
>>1203I think the blame has been misplaced - the problem is these slow as fuck frameworks. They promise speed but any real projects using them are slowwwwww
No.1205
>>1203For most cases, just caching a page on the server and serving it is not a big deal.
The problem I've found is that each tab on a webpage is actually, for whatever reason, sharing the same thread so if one locks up it can cause the entire site to freeze. Not always though. Haven't looked that hard.
So by putting too much stress on the client you can cause bad site performance. You could throw in service workers and the like, but it will hit similar bottlenecks in rending the page.
No.1206
>>1204>They promise speed but any real projects using them are slowwwwwwI think the real killer is first the webpage must load, then the javascript must load if it's not cached already, then the javascript must connect to an API and load stuff, and frequently it will even have to do yet another connection to grab more after the initial API fetch so suddenly you got 3-4x the latency for seemingly no reason and unlike in the past where you watched a white screen for half a second you now get to watch a bunch of spinners everywhere.
I'm in favor of always generating the full page and delivering it as such, even with millions of concurrent visitors it is not that hard to setup multiple servers to generate the webpage for the users. They did this in the past and only stopped because they believe saving money is more important than building a decent product and because most users use apps they can therefore just ignore the computer part of their website and let it rot on mobile oriented tech debt.
No.1207
>>1206I believe this site is client-side rendered. However, the initial html includes a script element in the head with json data including window.thread_json. Only the thread updates (including auto updates) fetch new json. A simple solution to the latency issue.
No.1208
>>1207It's as you say.
A nodeJS service handles the requests for the site. At one point I experimented with server side rendering and hydration, but it had no benefits for the added complexity. Instead the server prefetches the JSON data it will use to build the site.
The slow part is still the generation and such, but it doesn't have to wait for data at least.
No.1210
>>1208Do you cache the generated json for each post? I would expect so
No.1211
>>1210Well it would have to be regenerated whenever the "cites" json array changes but still.
No.1212
>>1210they're cached the vichan way. As json files stored to the file system
No.1213
>>1207I would also like to say that this made it reasonably easy to load full threads by redirecting /0.json to /full.json >:)
No.1217
>>1213that's a bit of a tricky thing to do I think, but I guess that is a benefit of dyanmically generated websites. Just use the software to render whatever you want
No.1218
TDD is so damn exausting
No.1219
What's your role, TC, and YOE?
No.1220
>>1218>TDD is so damn exaustingits a fun self-gamification but if you're trying to go for like 100% coverage of code from tests it gets annoying
No.1231
>>1220might just be the problem i'm tackling. going at it in the same intensity as you might in mission critical software... so exausting
No.1330
React keys are shiiiiiiiiiiiiiiiiiiit
No.1331
vs code crashing all the time now...
No.1359
[verniy@arch-verniy ~]$ sudo pacman -yS pulseaudio-equalizer pavucontrol
:: Synchronizing package databases...
core is up to date
extra is up to date
multilib is up to date
resolving dependencies...
looking for conflicting packages...
:: pulseaudio and pipewire-pulse are in conflict. Remove pipewire-pulse? [y/N] Y
error: failed to prepare transaction (could not satisfy dependencies)
:: removing pipewire-pulse breaks dependency 'pipewire-pulse' required by pulse-native-provider
I hate arch and the people who recomended it
No.1360
>>1359Just use Mint or Debian or something like that. It's not cool, but it works.
No.1361
>>1218just do TDD they way I do it, make the tests after you've finished coding the feature and don't do TDD
No.1374
>>1359I didn't migrate to Pipewire yet but going back to PulseAudio seems a bit more complicated. By the way, I don't know which DE you run but some of them like GNOME seem to require pipewire even if you use PulseAudio.
>>1360It all depends on what you wanna do with your OS. In some cases, choosing rolling release distributions over stable ones with old packages is better.
No.1408
Die pacman
No.1409
the pacman, the
No.1410
vecna lives
No.1413
Been trying to get audio to output Firefox to HDMI and line-out. Timing is a mess though and I'm not sure how to do it in pulse audio. The suggestion I found was quite depressing.
I need to figure out how to cause a delay between the two sources so they sink
>it works but the timing is a mess. i understand now why no professional wanted to write software for linux. the complete audio architecture is a bit of a mess. pulseaudio being more network oriented and jack single audio device. it is hard to create tight sync between multiple audio devices. The synchronisation issues are simply not easy to solve.
No.1414
'pavucontrol' might be the toolkit I need to do it, but it really seems like audio dev on Linux is a bit of a dead horse
No.1435
I'm not replying to anyone btw. I just like the image
No.1442
amazing. You have to compile everything for gento
No.1443
>>1442You don't. They started offering binaries for most packages. They've always offered them for the stuff that takes longer than 3 seconds to compile (browsers, Rust and a few others).
The end game: You install Gentoo on all your machines. Then you use your home server to compile packages for the rest of them. It churns away unnoticed and provides updates for everything from your workstation to your puny 15 year old thinkpad.
Gentoo is the best desktop OS because you can configure it any way you want and it stays out of the way after the initial 2-3 day manual install. You can have anything from a autism WM where you run everything from command line to a full blown KDE bloated desktop abomination with stupid useless screen effects. Need a patch? Drop it into a directory and let portage handle it. Don't like pipewire or systemd? No problem, set a USE flag or two and everything else just werks. No hacky work arounds. No backports.
Anyone that complains about compile times or it being hard hasn't tried it. Gentoo is the great filter. It's really too bad that RedHat has taken it over and is slowly killing it from the inside by bribing the council and people that seized control of the repos about a decade ago. Thankfully, Funtoo is a thing and gets better with each passing year and Gentoo can still be unfucked for the moment. But I expect to see it go systemd only at some point in the next few years. It's already effectively systemd by default. They just invoke everything from OpenRC instead of systemd itself.
I personally use s6 init and have removed all traces of potteringware from my config. It took a couple of days to set all that up like 4 years ago. But since then it just works. I run nightly updates with portage from the ~amd64 repos. Unlike Arch systems I've maintained portage doesn't break itself without my direct input. I probably have to manually do something for updates like twice a year.
No.1453
¥haven't played video games in months
¥no real desire to
¥play mahjsoul finally
¥get a double kan in a row and nothing came of it
¥do terrible in bronze
¥close the game
¥open world of warships with my old clan
¥terrible games there too
¥leave and just go back to browsing
i should just work on drawing
No.1454
you have to get angry to like multiplayer games which have luck... like warships and jong
No.1461
¥ open up text document
¥ closes text document
man it really be like that in this bitch of a kali yuga.
No.1462
Lately my mouse has been freaking out a bit and being inaccurate when making precise movements. Usually this means there's a cat hair in the little optical enclosure thing, but it LOOKS clear. I guess I need a better flashlight or something...
No.1463
>>1462I bought myself a can of compressed air to blow out any dust in my mouse wheel. Helped a bunch
No.1464
>>1463Ohh, compressed air. Yeah, that could work. Guess I need to buy one...
No.1465
There's this relatively new Game Maker competitor created by RPGMaker which seems kind of promising. More games have been popping up by Japanese people using it
https://pixel-game-maker-mv.itch.io/It's still a "No Programming Required" type product, but it seems more flexible in what it allows.
No.1529
Is there a file explorer that uses a tagging system? For example: There's a directory for all your anime songs with each anime having its own directory and what not, but creating a OPED directory for each anime, then having to go through all those anime to create a playlist of only OPEDs is a pain. So you tag files as "OPED" and when you want just the OPEDs you can bypass the directory barrier and show all the files tagged as OPEDs inside a certain parent directory.
Directory view:
AniSongs:
anime1:
OPED:
op.mp3
ed.mp3
song1.mp3
song2.mp3
anime2:
OPED:
op.mp3
ed.mp3
song1.mp3
song2.mp3
anime3:
OPED:
op.mp3
ed.mp3
song1.mp3
song2.mp3
Filter by OPED tag:
AniSongs:
anime1/OPED/op.mp3
anime1/OPED/ed.mp3
anime2/OPED/op.mp3
anime2/OPED/ed.mp3
anime3/OPED/op.mp3
anime3/OPED/ed.mp3
No.1532
fan that came with my case started rattling so I took it out. Apparently the ballbearings aren't infinite
No.1533
Is it possible to convert PSG subtitles into SRT?
No.1545
To spend 4 years making an RPGMaker game, and 3 years making a prior game in the same system. If this is what it takes to make a legendary indie game and they're not even tackling programming related problems...
Not to forget that whoever probably spent prior years learning to draw or coordinate a team for a product..
No.1564
>>1533>>1550It's bad and requires local software. PGS subtitles are image files. Meaning they have to be OCRed which is never perfect process and will require a lot of manual editing on your part to the resulting SRT (text) files.
What show are you trying to convert? You might be better off searching for existing SRT conversion by someone else. I can help you (maybe do it for you) if you really need it for something hard to find though.
PGS -> text file is why the most popular release for Moon Phase on nyaa right now has tons of errors in it. I was correcting a lot of them a few months ago. But then I got busy on other things.
Bit off topic but: I'm working on a new subtitle format that doesn't kill toasters while allowing advanced typesetting. Well it isn't my format (it's W3C standard) but hardly anyone uses it right now even though it has been around for years. I'm basically working on Aegisub replacement.
No.1566
>>1564the kaiji subs from last friday were PSG and handbrake couldn't burn them in so I figured something was odd about them.
I imagined that it was something about OCR
No.1567
>>1566PGS is the subtitle format native to DVD. They had to make them image based because most DVD players lacked the ability/power to decode advanced subtitle effects in real time and they wanted more advanced subtitles than a format like SRT (SubRip) could offer.
If you see PGS in an .mkv file it came from a DVD 9 times out of 10. That said it should be possible to "burn them in" (hardsub them) provided you have the right software. I'm sure ffmpeg could do it. Although I wouldn't go through the trouble.
Can you source .ass/.srt from somewhere else for this show? If the timing is a bit off that's an easy fix.
No.1568
i'm sure I can, but I was wondering if there were a way around it. For some reason handbrake should be able to do it, but it can't. Maybe it requires text for something, but it's supposed to not require it
No.1572
>>1571>allit only takes one episode
No.1575
>>1574Seems like it worked. You just have to make sure it's not on the 'foreign audio scan' since that never works. I have no idea what that even entails because in reality it means "select this if you don't want subs".
But, it's interesting to see "Burn in" greyed out on this default option titled Official Subs. There's definitely something unusual about these subs when compared to most of the ones I've seen in Handbrake.
No.1576
>>1575>>1575This is why I don't like frontends and I'm working on one. In general I try to stick with working with codecs directly or through ffmpeg at the CLI. Causes less issues once you know your way around.
Happy you got it to work.
No.1578
>In Handbrake, the setting was of limited use, as it relied on the subtitle being correctly tagged. You'll now see Foreign audio scan in subtitle settings, where Handbrake looks at the lines in the subtitle track to work out if it's likely to be a forced / foreign audio subtitle track, but it can and does get it wrong.
Strange
No.1579
>>1578Oh, huh. Yeah, I've never had it actually do anything. And weird that the 'forced only' thing was what broke it. Huh. Yeah, I try not to mess with these settings as long as what I have works.
No.1580
>>1579It may be (without looking closer myself) that there are two subtitles. One is dialog+signs while the other is signs only. Assuming the DVD had a dub of any kind the "forced" subtitles would be the ones that play along with dubbed audio. Kind of like the "forced" subtitles in a movie where certain characters are speaking another language and subtitles are needed for the English speakers to understand what's been said. Think of movies like Dances with Wolves and things of that nature.