Yukkuri

クッキー☆ & 淫夢 & P-MODEL

Public Key

npub1j05dh9hemkteevv6env2ssv6zrlhj9s904zzknsmylqm2p03w4qsxlw8rl

NIP-05 Address

[email protected]

Profile Code

nprofile1qqsf86xmjmuam9uukxdvek9ggxdpplmezczh63ptfcdj0sd4qhch2sgpz4mhxue69uhhyetvv9ujumt0wd68ytnsw43qvz4fad

Publishing to

relay.mostr.pub

Author Public Key

npub1j05dh9hemkteevv6env2ssv6zrlhj9s904zzknsmylqm2p03w4qsxlw8rl

Show more details

Last Notes

2024-04-26 17:50:11

- reply

@npub1ajw…aj8d Thanks, I'll keep it in mind, but CPU consumption isn't an issue for Eientei yet, nor is database size (for now). Also for MRF maybe it's better to ignore an activity rather than reject it, to avoid some implementations getting stuck in redelivery loop.

2024-04-17 21:09:49

- reply

@npub1n78…87ev
https://eientei.org/media/8c270d32448bdee669760ecba7e1faa618fcd237f514851a29056cc03e9705b9.jpg?name=__iwakura_lain_and_bike_chan_serial_experiments_lain_drawn_by_abe_yoshitoshi__0e7367c13b51817f22986a9e41aeb1c6.jpg

2024-04-15 18:21:15

- reply

@npub10l8…vrsg It's a tomb, not a crypt. And naturally, yes.

2024-04-15 18:10:24

My tomb construction is finally complete. Now all that left is to fill it with trinkets.

2024-04-04 20:17:39

Youkoso Japari Park e Fandub!!!
#クッキー #ネッコー #クッキー本編リンク #ネッコー本編リンク #例のアレ
https://www.nicovideo.jp/watch/sm43551605
https://eientei.org/media/361724c62eb8b461fb71062fe40dc9991200cd6eaffff86d1bd90157eb85d514.mp4?name=sm43551605-max-video-h264-720p--audio-aac-128kbps.mp4

2024-03-24 11:07:35

- reply

@npub1urs…dc0y To be effective prevention against finding a configuration that is easy to solve, it must conceal the puzzle pieces until a time passes. Loading with limited bandwidth offers exactly that, forcing attempting user to wait until at least 50-90% of image loads and bandwidth can control the time required for that to happen. It can be simulated with animation as well, but there is no point in implementing inferior solution to achieve the same effect.

2024-03-24 06:20:35

- reply

@npub1urs…dc0y That is what I am talking about, it is relatively easy to control the fully loaded gif/apng playback, since all information is transmitted to the user agent at once. With bandwidth limitation you get roughly same effect as with delayed animation, but with advantage of information itself being fed to the user agent piecemeal over considerable time.

2024-03-23 22:53:23

- reply

@npub1urs…dc0y If you want waste automation's time as much as human's, limiting bandwidth to 1-10 kb/s on captcha image serving endpoint might do better job.

2024-03-22 13:55:37

- reply

@npub1n7z…t9xj @npub15sv…puhr The only true kolchan is the news resource :chiyochichi:

2024-03-18 13:44:48

- reply

@npub1urs…dc0y Mental capacity required for basic spatial reasoning isn't very high to be honest, and while CAPTNHA was proven effective against french, it may do little against coomers. Nonetheless, you could try running 0.1.4 and increase difficulty with -rotations-minboxes 10 -rotations-maxboxes 10 -rotations-dof 3 -rotations-variation 10 -rotations-sdf-distance 22

2024-03-14 14:25:49

- reply

Lol, apparently some google-indexed french forum found it and at least some of them had difficulty with the concept of 3D rotation.
Accidental battle-test was a success.
https://avenoel.org/topic/1464385-1-le-futur-des-captcha-les-captnha

2024-03-14 07:48:18

- reply

@npub1u83…r2s5 @npub1ecj…fchy @npub1g45…tdw8 @npub1php…6l98 @npub1zr9…uu4j @npub140w…su0k There also must be ways of approaching this goal, and thus metric on how close you got, even if it is given subjectively by other player. Otherwise you can again have the image editor with the goal of defeating demon king, so I don't think you can workaround including progression, and a way of gauging it. And since you could also have progression without agency, both agency and progression are required indeed.

2024-03-14 07:34:43

- reply

@npub1u83…r2s5 @npub1ecj…fchy @npub1g45…tdw8 @npub1php…6l98 @npub1zr9…uu4j @npub140w…su0k I don't agree, having agency alone is not enough. You can have lots of agency, but no progression, e.g. image editor or a dress-up simulator. A judgement on how successful player choices were, either imposed automatically, by game mechanic itself, or by other players is also a requirement, otherwise it is just a creative tool that only becomes to resemble a game once you start showing what you've created to other people.

2024-03-14 07:18:49

- reply

@npub1u83…r2s5 @npub1ecj…fchy @npub1g45…tdw8 @npub1php…6l98 @npub1zr9…uu4j @npub140w…su0k Probably what defines games the most is the sense of progression. Various genres approach it differently, some rely on solely player getting good, others also on some stat growth, but all would still give clear gauge of how successful player is. By that metric, what progression most VNs could offer is usually limited to plot progression, as getting better at reading is not as easy to gauge, and there is very little player agency.

2024-03-13 08:16:40

- reply

@npub1ecj…fchy There is a bunch of whitepapers on dealing with salt-and-pepper noise up to 0.9 using just general algorithmic approaches, and you can probably go past that with specifically trained network, so until someone actually bothers with implementing at least the 3d reconstruction, any noise might be just excessive, and past that moment - won't stop anyone dedicated to image/shape recognition. I'm thinking on adding some form of Raven's progressive matrices as another module, at least it would be entirely different domain for someone to tackle.

2024-03-13 06:51:30

- reply

@npub1ecj…fchy Sorry, my use of "uniform" here might have been misleading, I only referred to noise saturation approaching entire image, with 50% flip probability. With noise, you might also want to effectively disable fake anti-aliasing by setting both aliasing thresholds to the same value, so there would be no detectable half-tones in rendering, while increasing wireframe thickness (or using noise values less than 0.5) so it still be recognizable. Here is an example of -rotations-sdf-alias-start 0.005 -rotations-sdf-alias-stop 0.005 -rotations-wire 0.03 -rotations-noise 0.5
https://eientei.org/media/fc3f459698750f30e9413858623f860921bd76d9e64c3cdfa31ea45228968044.png?name=image.png

2024-03-13 06:00:35

- reply

@npub1ecj…fchy Added -rotations-noise parameter with values from 0.0 (no noise, default) to 1.0 (uniform noise). Example with -rotations-noise 0.5 attached. But opencv and specifically trained denoise networks might be better at filtering even extreme cases than humans are, hence not on by default, probably would only help warding off the lazy.
https://eientei.org/media/a797c2ce0f06e81ed2b0ca8cf8d9ce92ec27b7a5c0d4633bc32a011cf145ec86.png?name=image.png

2024-03-11 05:52:21

- reply

@npub1w4j…cead Huh, I have astigmatism. Also in one of the latest updates reduced default wireframe thickness, it improves things a bit, but otherwise captcha being taxing on attention is by design.
https://eientei.org/media/8275ce38a5a734296b7d9a3f90eb8b5d591e0d5a06e8aa40300c8cae0e0e9dd0.png?name=09dd2b3a5feb2c625858038d0db58f79.png

2024-03-11 05:33:11

- reply

@npub1w4j…cead If you mean album orientation of the picture frame, it's meant as drop-in replacement for pleroma[-fe] captcha, which design does only allow vertical image growth. If you mean rendering figures as wireframe instead of shaded cubes, on higher box counts it is no longer possible to have orientation showing every one of them, requiring the wireframe.

2024-03-10 12:45:29

Seeing an image of 2B disappointed by UN boat clicking simulator, I realized what would be useful is not another CAPTCHA, but a CAPTNHA, a test to tell apart niggers rather than computers, regardless of their hardware, so here is a first iteration approach to the concept, requiring user to be capable of at least some spatial reasoning: https://gitlab.eientei.org/eientei/captnha
Compatible with kocaptcha API, so it can be a drop-in replacement for pleroma.
https://eientei.org/media/43744d24558dce7558d8ea498e692c131d5e3f4ba7e985d4884dff6874f92807.png?name=image.png

2024-03-07 14:32:40

- reply

@npub1vwr…fnzd @npub15c7…suze Armageddon was best so far.
https://eientei.org/media/81a4200ac0410604dac631d988a6fe4a6570c8da51a6afffb37f66c5f1701cad.png?name=zekfix.png

2024-02-29 19:46:58

- reply

@npub1ajw…aj8d @npub1f8l…7frq @npub140w…su0k With one simple trick you can ward off 90% of copy protection and anticheat systems. A client-side validation crutch developer fears self-signed kernel modules.
https://eientei.org/media/67432f37fe1a40c0bb2f1697946f124ace8c0575d122b4038e682b38b4a5a53a.png?name=image.png

2024-02-14 19:52:30

- reply

@npub1xcq…h0nf Must be nice.
https://eientei.org/media/a9a9ee0defd6d543f7a5a488c3cd8e5f4e273b378ac9d7e4987d00623fa3690f.jpg?name=__cirno_wriggle_nightbug_and_the_man_touhou_and_1_more_drawn_by_grummsed__d245d29cd0982f2b0210dabddaf1b37f.jpg

2024-02-14 10:24:57

- reply

@npub1sdr…zct8 @npub1xan…jlvy
https://gitlab.eientei.org/eientei/pleroma-fe/-/commit/711e17391a97a7fa165cf21babf53ec2b4e3a584
https://gitlab.eientei.org/eientei/pleroma-fe/-/commit/1cfb1fd11025878ea5f51471cdb6db0dad5c54ab

2024-02-13 17:50:20

- reply

@npub1dph…t7jj Annoying, breaks into private property to talk nonsense, sounds pretty fae to me.
https://eientei.org/media/adc09811502c4aa243eaadab59b7c6d7157a888bba353ca46390a94bb9ab19da.png?name=image.png

2024-02-13 17:46:05

- reply

@npub1dph…t7jj
https://eientei.org/media/7c3875d02818f2a5a36abf438240413449fdda6ef12effec73bdb78383f6f4c3.png?name=image.png

2024-01-28 20:29:43

- reply

@npub1wg0…t0yf BSD is nice so long you have all your hardware supported, which is rarely the case for desktop/laptop. Any ports-based system should be decent enough with varying nuances. BSD variants usually purist in using makefiles/shellscripts to implement it, linux ones tend to rely on python for some reason.
Going full source-build can have it's advantages, e.g. if you're on non-intel architecture -- you won't get optimized for in binary builds, but here it's you who is setting the build flags. Or if you want to exclude some feature from entire system, but expect hand-patching whatever doesn't support excluding it [any more] yet you still want to have. Or if you want to down/upgrade some widely used library system-wide. Or if you write own stuff that you want to statically link with every dependency, but expect that not all libraries work well in that use-case, especially bloated desktop ones can be a pain to statically link. Thats about it though.
To customize/patch some specific thing, you could recompile just individual package(s), and debug symbols for binary distributions are in uniform repositories now, thanks to debuginfod, so you wouldn't miss attaching gdb to anything that moves even on binary distro.

2024-01-26 13:40:01

- reply

@npub1n78…87ev But if you wish to do the arithmetic with awk as well, then certainly.

2024-01-26 13:37:08

- reply

@npub1n78…87ev
> That awk fails to match xrandr's annoying output, I'd use 'x| +' for the field separator
If you run it in isolation, yes, awk would include remainder of the string into $2, however set -- always use whitespace separators, so extra $3 and above should be of no consequence.

2024-01-26 06:15:29

- reply

@npub1n78…87ev
set -- $(xrandr --current | awk -F x '/*/{print $1, $2}'); PAD=$((($1-$2*4/3)/2));
Unless shell doesn't support arithmetic expansion.

2024-01-12 23:02:44

- reply

@npub1xan…jlvy
https://eientei.org/media/9878bb063e21e8f6d791cae3eeec533a66074de978a278f7798f53834ebff8da.jpeg?name=_______-________-Hearts-of-Iron-____-crossover-6454951.jpeg

2024-01-06 14:13:43

- reply

@npub1n78…87ev @npub1c4s…aqv4 It isn't so much question of good or bad, ceramic/ceramogranite is always hard and therefore brittle. Tiles have near no elasticity, hence only way for kinetic energy to dissipate is to introduce a fracture. ~5kg dropped from 1.5-2 meters with spherical point of contact may crack or chip away virtually any tile with standard for 60x60cm ~1cm thickness if underlying surface is also hard and rigid. Spares certainly do help, but if you ever replaced an isolated tile, you'd know the fun of clearing out remains of old tile and then trying to level/align fresh adhesive with already set from adjacent ones. Laminated wooden tiles are much more serviceable, and come in heavy duty certified variants too.
> please wear something + you are from Russia
I'll leave that to burger golems.
https://eientei.org/media/826581d70302180725a7b42f22ba8c2fff0508060768469ef4403ce08106de89.jpg?name=ezc3tik1pyw91.jpg

2024-01-06 02:04:53

- reply

@npub1n78…87ev @npub1c4s…aqv4 Until you drop something heavy on one and it cracks, then you have to hope matching ones are still produced and color don't vary between series too much. Also other than for office buildings, a floor heating system may be required, otherwise tiles can get really cold to be used without footwear; and those are real pain to maintain and repair.

2024-01-04 08:55:59

Increasing postgres statistic level turned out to be way more useful than anticipated; in case of a new user, with page size exceeding number of activities matching user timeline, planner was estimating over 200-300k rows and had to do filtering on id desc nulls last index instead of using bitmap scan on recipients gin index when filtering for recipients intersection, causing query to stall for several minutes.
alter table activities alter column recipients set statistics 1000;
analyze activities;
(a 10 times increase from base 100)
reduced estimates to ~1-3k rows and query time to 20-50ms with just bitmap scans of respective indices.
Another option might've been to force planner to scan ordering index after applying bitmap scan with putting everything but order/limit into subquery, but for it to work, subquery must include all Ecto preloads, which isn't something Ecto currently supports.
ORMs are a lie, as usual.

2024-01-02 23:14:49

- reply

@npub1wg0…t0yf @npub1tfm…38x3 @npub1rl6…xjcn Oh, now @npub1t8l…pgly rushed uploading those; https://eientei.org/emoji/user/mint/#

2024-01-02 23:12:08

- reply

@npub1wg0…t0yf @npub1tfm…38x3 Slander! All our emojis are here: https://eientei.org/emoji/all/# ; at least since indiscriminate emojisteal from every federated post was replaced with @npub1rl6…xjcn

2023-12-27 21:52:04

*soon*
https://eientei.org/media/4091706471265b950465292635c11d99037b0c79458437e4d81d06f2c3c88c3f.png?name=image.png

2023-12-27 10:23:49

- reply

@npub1xan…jlvy @npub1wg0…t0yf "IT" might be rather broad. Junior/middle QA and analyst positions are often dominated by women, they are very apt at tedium of executing manual scenarios and mediating multiple parties in endless calls, so long there is something or someone telling what to do, but lack passion required to proactively seek improvements. If you'd see a woman with own initiative to break status quo and established consensus for the sake of betterment (e.g. refactoring, redesign of architecture, adopting something), thats a sure sign of either a man behind that woman, or a man in woman's clothes.

2023-12-25 17:29:38

- reply

@npub18dh…dq5s @npub1ajw…aj8d Looking at their repo, at least migration-wise it should still be possible to detransition. Most migrations are not destructive, some even providing the down/rollback option. You could start with mix ecto.rollback --to 20220108213213 (last upstream migration). Some migrations do not define down option, but those are just adding tables/columns, you would be able to drop them manually, just go through files between
https://akkoma.dev/AkkomaGang/akkoma/src/commit/6a19fbac2f0d9c05bc25159d6dc19b89d19f2241/priv/repo/migrations
and
https://akkoma.dev/AkkomaGang/akkoma/src/branch/develop/priv/repo/migrations (or corresponding to the akkoma version that is installed)
Note that you do need to drop at least announcements table, as pleroma migrations would try to add it themselves, it would be safer to revert every change.
And then migrate everything up from upstream, it may even work.
Testing afterwards if everything works would be tedious, but doable. No need to take cosplay to 51% extreme just yet.

2023-12-25 16:58:22

- reply

@npub18dh…dq5s @npub1ajw…aj8d If it is trivial as adding/renaming the field, you just do the logical reverse, drop/rename it back. Only possible with non-destructive migrations, switching to another hashing algorithm would likely require to explicitly support both algorithms on application level, or notify every user to set working emails and then do mass password reset after returning back to upstream. Assuming there isn't any more backwards-incompatible changes, so make a backup and experiment on local test instance maybe.

2023-12-25 16:50:25

- reply

@npub18dh…dq5s @npub1ajw…aj8d Eientei did actually migrate to akkoma and back shortly after release to claim the title of most racist and transphobic akkoma instance. At least at the time getting back was nearly noop, maybe manually reverting one migration at most. Not to say they wasn't mean and petty all along, but there was no vendorlock at least.

2023-12-19 19:29:11

- reply

@npub1xan…jlvy
https://eientei.org/media/081ed137ed61c04e182edfb1ed96f2259b472f014e60f632ae5aefd067f21d35.png?name=image.png
https://eientei.org/media/4ef35c938e4c4072ebbfb5b923f11b22c4aeed5157375ffd17653c16b41c05ba.jpg?name=__guu_jungle_wa_itsumo_hare_nochi_guu__dd5234df97caf0b38482149bd77255c3.jpg

2023-12-19 19:26:32

- reply

@npub1xan…jlvy
https://eientei.org/media/6c035cec570a829edd4a2bb1c0bbdea2c3d603f39a9bc91c60d06f21492ab2a2.jpg?name=WaFY0.jpg

2023-12-15 15:33:32

- reply

@npub1n78…87ev Looks like list deconstructor

2023-12-13 20:22:57

- reply

@npub1xan…jlvy @npub1z9a…zn6y @npub1t8l…pgly
https://eientei.org/media/b3bdfee82a23d44d8be8ee6614bc3333e2f12bb84304f23fe87717af8de7135a.png?name=image.png

2023-12-13 20:08:14

- reply

@npub1ajw…aj8d also pleroma[fe] does not treat onion as a valid TLD naturally

2023-12-13 20:07:26

- reply

@npub1ajw…aj8d Apparently this can be in general slightly complicated by RFC 7686-compliant software like curl, and if you have a httpd already, another reverse proxy might be needed to re-direct 80 to the privoxy, but for now a temporary setup on eientei side is complete.

2023-12-13 19:48:53

- reply

@npub1ajw…aj8d Don't see where DNS caching comes into it, unless you mean to do the *.onion wildcard via it's static entry.

2023-12-13 19:34:29

- reply

@npub1ajw…aj8d Not normally, but I can install a dnsmasq temporarily. What period do you want it on?

2023-12-11 15:54:14

- reply

@npub10ph…mlal You seriously spent all the time hand-picking 41 posts? @npub1xan…jlvy do you still have "once I report you your life is fucking over" handy?
https://eientei.org/media/c426188aff0826087eca5fba4144aee7fb3702eb3ab842ed8bc3272cff15c6a3.png?name=image.png

2023-12-03 20:38:48

- reply

@npub1xan…jlvy Spine correction needed.
https://eientei.org/media/2e4b93b392f66bcd6b5f55aeb208418e1232de7485a4457c06c9e2bdf793f969.jpg?name=__nazrin_touhou__feae65bb676fd2f094587ac39bc4caf3.jpg

2023-12-03 20:24:10

- reply

@npub1xan…jlvy cyka krisa
https://eientei.org/media/be45d355bf365e67b1fb49a3dca6e895f21249bf286441ef31a9de66d976a5be.jpg?name=__hakurei_reimu_kirisame_marisa_and_nazrin_touhou_drawn_by_ao_usagi__2b2791adbd5dcb64494dba6ca48d6ac0.jpg

2023-12-02 09:43:47

- reply

@npub1x70…cxkp @npub1fr8…lmwg There are no original characters, with exception of miku, all are taken from existing imageboard mascots and renamed at most.
https://eientei.org/media/ca1919f83f22c2b79b6a6c65b12568374d461083a5f7d748e30056746f8254ba.jpg?name=fdcb2ecc894d5c5d95a2d1e9110f9ff5.jpg

2023-11-19 12:53:18

ノリノリ実況
#ノリノリ〇〇シリーズ #帝京平成大学のここがすごい #棒歌ロイド
https://www.nicovideo.jp/watch/sm42957090
https://eientei.org/media/f1c9715c1904f648abf6f74eef142c6f69907ac8cd8e4f9a54ec242dada19c02.mp4?name=sm42957090-max-h264_1080p-aac_192kbps.mp4

2023-11-16 16:47:20

- reply

@npub1n78…87ev @npub1wg0…t0yf At least they made nice modems.

2023-11-11 15:25:10

- reply

@npub1dph…t7jj @npub1sdr…zct8 @npub14mk…sqx8
https://eientei.org/media/2a8272a946754dab33411449d69831252651db1684a0432f56a3d44b90c32aeb.png?name=image.png

2023-10-13 16:56:00

- reply

@npub1x70…cxkp @npub1t4s…ud4h @npub188f…nn4f @npub1xan…jlvy There is only one true way.
https://eientei.org/media/6e21a27740bfbacf2fde2092bc141f406e6790a299e53a00578c3cf5d894b5fa.png?name=image.png

2023-10-08 20:29:08

- reply

@npub1msm…0dqu We do a little trolling.
https://eientei.org/media/5e8179dbc043510d8ed5555074e259ce686bf56b3d9983c829d4d2f6c2704eed.png?name=image.png

2023-10-05 19:36:08

- reply

@npub1n78…87ev @npub1jax…nvuh @npub1wg0…t0yf @npub1kwu…qdsv That can be work-arounded by predicating on NF: awk '/pernias_passwords/ && NF >= 2 && $0=$2', but It also won't handle files with whitespaces and newlines in names properly. find . -path '*pernias_passwords*' -exec du -a {} \; might be required.

2023-10-05 19:23:46

- reply

@npub1n78…87ev @npub1jax…nvuh @npub1wg0…t0yf @npub1kwu…qdsv That would be true (1), but in reverse order should be fine.

2023-10-04 21:12:02

- reply

@npub1xan…jlvy tux racer
https://eientei.org/media/4e3f49a35d303023f082264919e19a9361370fc89d9a522d5d2029eb91b38f2d.png?name=image.png

2023-10-04 21:02:36

- reply

@npub1xan…jlvy Looks like some kind of a pinball game

2023-10-04 15:34:45

- reply

@npub1xan…jlvy @npub1erc…9xy8 @npub1pa6…p7r6 @npub15ft…ef0r NIPA~H
https://eientei.org/media/06519a3a3b351ad134e3a4b505403760b74ffa17baaeab480f2f4e6dcca890a8.png?name=image.png

2023-10-04 15:30:20

- reply

@npub1xan…jlvy @npub1erc…9xy8 @npub1pa6…p7r6 @npub15ft…ef0r
https://eientei.org/media/f7662d68f7f5e745f4ca2bd9551477add210476278f918ba1c34bf7ec7fd53da.png?name=image.png

2023-10-02 17:29:26

Being demented is fun, you run shutdown -P on random servers and wonder why the one you wanted to is not shutting down.
https://eientei.org/media/8e2ba3f0fc80a0b131a2ace443d9588b68c215e6d6a5964f2fbb7764772294fe.png?name=image.png

2023-09-30 21:00:26

- reply

@npub14ug…8jay
https://eientei.org/media/59436f94db209682382831f69827c57ff109b4e1c912c6adb1efa8ec711e0bb1.png?name=image.png

2023-09-24 07:32:55

- reply

@npub1n78…87ev @npub1wg0…t0yf Just shifts ascii range around a bit, basically
tr ' -/:-@[-`{-}' '`-{/" -'
using s/// to initialize the buffer, y/// to do the replacement and then s/// again to do the system call. If sed did support ranges and unequal sets in y///; it could've been just sed.

2023-09-23 21:40:05

- reply

@npub1n78…87ev @npub1wg0…t0yf
echo "test... test... test..." | perl -e '$??s:;s:s;;$?::s;;=]=>%-{<-|}}./}:<%=%<@%}|}<&|`{;;y; -/:-@[-`{-};`-{/" -;;s;;$_;see'
the movie.
https://eientei.org/media/23ab1eab48a4184f64386f1f9cd76548544a18230192de0b5776c07f968a3812.png?name=image.png

2023-09-23 06:36:33

- reply

@npub1xan…jlvy фальшивка
https://eientei.org/media/2f7de82b2f239cd8c0dfe44384658f69ed70abf84872f9ba830b1f2aab769199.png?name=image.png

2023-09-23 06:26:56

- reply

@npub1xan…jlvy UYO OYU потрачено
https://eientei.org/media/51d5e46329bd24c794025d1b55f7823f5d221d9bb4a41ece321bc70a0f2f78c0.png?name=image.png

2023-09-22 22:27:52

- reply

@npub128m…fzga
https://eientei.org/media/0de1cc4c70f7cc7dba2d8d194488260dcf85e269f6b9e047e3cb625114aa22a7.png?name=image.png

2023-09-22 14:47:19

- reply

@npub1xan…jlvy Unfortunate indeed.

2023-09-22 14:45:30

- reply

@npub1xan…jlvy If first seen date is same as last seen, it doesn't even had to be some other user. ikwyd just randomly attributes unrelated hashes this way.

2023-09-21 19:27:43

- reply

@npub1wg0…t0yf Such test is only good to confirm. If cargo supports some plugin loading mechanism, or starting new processes, it does not necessarily has to be a part of binary. And if not - resource in question could be compressed before embedding.

2023-09-21 08:20:27

- reply

@npub1xan…jlvy I can't fix myself, let alone anyone else.
https://eientei.org/media/72d5173dcea1e34e5a236334e90efe80d55c6ade6068fba5b6315345bb9e91ab.png?name=image.png

2023-09-19 22:31:22

- reply

@npub1vr0…glwg Lithium batteries can produce a lot of spectacular gas and heat when short-circuiting or contacting even air moisture, especially if discharged. Got the result above on ryzen 7 4750U with no energy optimizations and ~20 of docker containers with scylladb/postgres/kafka/redis also running in background, so with some moderation on scaling and scheduling I should be fine for several days straight at same workload, which should be enough to reach power outlet one way or another.
Still would prefer having swappable battery on laptop, those are more straight-forward to manage.
https://eientei.org/media/47b1818f1c3146c73bb48e1cc138291f297034d72eb77a13f86e4efbff073493.png?name=image.png

2023-09-19 21:13:56

Well, 300wh bomb lasts about 15 hours on full CPU power and occasional recompilation, indexing and test runs. Should do for several days with throttling and tlp.
https://eientei.org/media/41e4b70d0f29e6b0f824c2ce0fdd30110a8d4d367d8600d41f88e5881ca64085.png?name=image.png

2023-09-16 21:08:32

Жрать. Мозги. :cheesesausage:
https://eientei.org/media/41703576e3563f1422957e2b9fe1458ec49f2e29592ecff5ef981e13c0d83257.png?name=image.png

2023-09-16 20:01:35

- reply

@npub1n78…87ev @npub13wh…wvah @npub1wg0…t0yf I'm more gered towards TIS-100/Shenzhen/EXAPUNKS, pseudoassembly is easier to play around with than visual turtles.

2023-09-16 19:45:33

- reply

@npub1n78…87ev @npub1wg0…t0yf @npub13wh…wvah But zachtronics games are very fun optimization puzzles, though global leaderboard might be required in case you missed some less obvious optimization paths. And usually you end up with 2-3 solutions per problem to meet various optimization criteria.

2023-09-16 17:33:48

- reply

@npub1xan…jlvy There is no buses in Gensokyo.
https://eientei.org/media/2286cc2f37bc721357f8e8aced0ec09e4b0f233772ba7d15712622a4bb0e5c93.png?name=image.png

2023-09-11 17:22:24

- reply

@npub1xan…jlvy Брежнефу
https://eientei.org/media/f22a97fbb029e12240ce0c48320f7f66a6a35274954d2c6f948d7cfd50519ece.png?name=image.png

2023-09-11 13:08:06

- reply

@npub1xan…jlvy Even in death freedom is not guaranteed.
CP
oa
ns
ss
cs
ii
on
ug
s
n
e
s
s
https://eientei.org/media/26c3a091be9ec6384145ddeaf489e0ca25171056300d1af020a593eae1b9d6a6.png?name=image.png

2023-09-10 21:00:03

- reply

@npub1n78…87ev
https://eientei.org/media/1da1c56a4503fd35fbd9399e25186110e7ef6287a6d78c7b6950e0bcce477eeb.png?name=image.png

2023-09-06 17:20:34

- reply

@npub1xan…jlvy ВИUY
https://eientei.org/media/9ae2b632e50201e451b5368019b0ec42e29276ffb12c124833386ae4c1ed7f05.png?name=image.png

2023-09-04 10:15:56

- reply

@npub1aw2…70dd @npub1zhj…kmgu @npub1xan…jlvy @npub1ajw…aj8d removing them with sql delete marks them as unused and subsequent vacuum full would rebuild tables without unused log entries. But it is a blocking operation and can take considerable time.

2023-09-04 10:01:13

- reply

@npub1xan…jlvy @npub1zhj…kmgu @npub1ajw…aj8d Currently pleroma DB is 130 gigs, with daily growth averaging at about 200+ megabytes. The size is also somewhat inflated by FTS indices. Have been running since February of 2021 for ~2+ years now and have survived domain migration.
https://eientei.org/media/c66aee01156a1557cef45947b1fa0d4372103957c12f938073789d965816ee2e.png?name=image.png
https://eientei.org/media/961dff7c1043a64eb66bd03f8388dfa441485691a54e54632ae70634f4ddc8ad.png?name=image.png

2023-09-02 17:10:25

- reply

@npub1xan…jlvy
https://eientei.org/media/b9b8f86a9c5a1d14187ef0ec050e15da271376bec9a9c3db0c82d7b989ef1cc2.png?name=image.png

2023-09-02 10:02:57

- reply

@npub1xan…jlvy @npub1srh…8g90 @npub1mjr…w8z3 @npub1y6t…9wl9 @npub1zhj…kmgu Scaling factor setting should now be available.
https://eientei.org/media/eaa0e5fd38ec4d1dbef0d0ac6dbcfd4c23dce86ccae72971ae9d26ac2e2b6bda.png?name=image.png

2023-09-02 01:42:38

- reply

@npub13v6…hy07 @npub1uvf…teda @npub1xan…jlvy @npub1c9k…727g @npub1zna…wtpl What are you talking about, it was there for past 10 years starting with xen patches and then upstream linux adopting vfio and ACS patches. Just pass iommu group as a whole and there shouldn't be any issues, granted your cpu supports vt-d/amd-vi

2023-08-30 06:51:58

- reply

@npub1xan…jlvy @npub1srh…8g90 @npub1mjr…w8z3 @npub1y6t…9wl9 @npub1zhj…kmgu Maybe an out-of-layout tooltip would be better suited for hovering (same as media viewer? Simper one?), rather than changing the in-layout element size, which would fuck with selections.

2023-08-30 06:46:16

- reply

@npub1xan…jlvy @npub1srh…8g90 @npub1mjr…w8z3 @npub1y6t…9wl9 @npub1zhj…kmgu Such setting is possible to add to PleromaFE, but emote exceeding line height won't look great in multiple lines of text. I suppose there can be setting to also increase line height, but it may still look awkward after some point. I'll take the time on weekend.

2023-08-27 20:57:55

- reply

@npub1uel…ea56 As usual, suffering builds character.

2023-08-27 15:57:00

- reply

@npub1pu7…pdj6 Euler angles lock at poles, use quaternions and a translation vector to build a proper orientation matrix.
https://eientei.org/media/7aed1672b64b3ad2fab213921f87211b5ff3265939c387634c987f0fd619bddd.png?name=image.png

2023-08-26 08:43:55

- reply

@npub1xan…jlvy Just call him a retard.

2023-08-22 16:05:36

- reply

@npub1xan…jlvy @npub1n78…87ev This one probably, if you mean competitive.
https://eientei.org/media/ffc1f5730a2f633ca45a27b89df39c500dbd32719a6d2627c7c45a121ecbb9da.png?name=image.png

2023-08-21 17:25:29

- reply

@npub1n78…87ev @npub1adk…52kc @npub1z5r…rjcc
> the longest *scripting* you do on cli is something like `ls -1 | xargs file | grep 'image'` so it makes sense to use whatever's the fastest.
Cannot relate, I tend to do 1-2kb one-liners rather often. Having bash helps to leverage it's interactive extensions while not straying too far from POSIX sh, if it ever needs to be saved as a script. Most of my scripts were one-liners that got too unwieldy to edit once.

2023-08-21 12:20:41

- reply

@npub1xan…jlvy Ага, его в /to/ ычана форсили, этот чарт вроде целиком от туда

2023-08-21 12:18:57

- reply

@npub1xan…jlvy О, шики-кун из коридора тоцноц.гц

2023-08-21 06:42:50

- reply

@npub1xe5…xg59 @npub1snh…ywgf @npub1lwt…q3ry Even if place blocks everything, it should allow some form of DNS resoluion, in which case, depending on strictness of blocking, you can connect to your preferred VPN implementation running on UDP port 53, or use an actual DNS tunneling, e.g. iodine