Skip to content

Latest commit

 

History

History
557 lines (341 loc) · 30.6 KB

2021-12-10.md

File metadata and controls

557 lines (341 loc) · 30.6 KB

< 2021-12-10 >

3,522,004 events, 1,844,183 push events, 2,728,389 commit messages, 215,480,139 characters

Friday 2021-12-10 01:47:14 by Terry Ott

Calling it for the night, remove itty router from the object, it's fucking my shit up


Friday 2021-12-10 01:57:38 by Paul

A Linux 150% zoom hack (#5611)

For some reason, fonts wiggle at 150 and only 150 zoom when you drag a slider. We don't know why. But if you force 150 to 149, it's all fine.

My theory is some rounding is hitting either side in a paint somewhere but that's deep in juce land if true.

Addresses #5565


Friday 2021-12-10 09:07:43 by AsherY8979

redid a whole bunch, changed username

fuck you leon


Friday 2021-12-10 10:14:53 by LemonInTheDark

Fixes random hell call-counts to Move (#63317)

Removes the parent call from mob/Login

Mothblocks pinged me with a profile of just a hellish amount of move calls Way too many, like 200000 Started looking into it, got distracted by how expensive macros were Turns out most of the cost of macros was in the nuke ops summon spawner

So I looked at why, only bit of it that was at all expensive was the login for the cyborgs Tested on a local, huh that is slow yeah

Looked at the profiler, huh there's that move count again So anyway, why is login so expensive

Spawned a few cyborgs in, dragged myself into them, nothing Spawned myself in with sdql magic using the summon spawner, man it really is still there I guess it has to do with move somehow, try and stick a breakpoint on it, get fucked by the debugger I notice the hanging comes during the Login parent call

Try again, this time with good breakpoints. We're trying to move, to (1,1,1), just like the reference says we will https://secure.byond.com/docs/ref/#/mob/proc/Login

But man, it just keeps happening, and we don't actually move Step through the code, we've got that null loc check in atom/movable/Move

So the move to (1,1,1) isn't working

Here's the exact line from the reference "If the mob has no location, place it near (1,1,1) if possible" Keyword is near

Talked to lummox about this behavior, figured it was a bug

It turns out by near, they mean inside that tile's area It'll keep trying to place you somewhere, in an attempt to effectively cover for shitty login systems, until you succeed in moving

That tile is space. There are 200,000 tiles with the same area as it OH.

So anyway, we're not calling parent on mob/login anymore. We can do all the work it did that we care about ourselves (IE: Just the statobj set) And this way we don't need to worry about 4 SECONDS OF OVERTIME WHENEVER SOME POOR FUCK MESSES UP SPAWN ORDER

So yeah, I'm a genius and not at all just malding at the existance of keybind macros, and hopefully another source of stutter bites the dust Not actually sure how widespread this is, but even if it's just spawn becacons that's pretty banging


Friday 2021-12-10 15:09:47 by vanstorm9

Adding songs 'Break The System', 'Concentrate!', 'Diabolic mulier', 'Eieisa', 'First Love Again', 'Itsu datte for you!', 'L! L! L! (Love the Life We Live)', 'LIVE with a smile!', 'Miracle STAY TUNE!', 'Silent Blaze', 'TO BE YOURSELF', 'Toy Doll', 'Turn it Up!', 'Yada!', 'Ye Mingzhu'


Friday 2021-12-10 15:28:24 by vanstorm9

Adding songs 'Akogare Ranraran', 'Cotton Candy Ei-Ei-Oh!', 'KU-RU-KU-RU Cruller!', 'Love Spiral Tower', 'Motto ne!', 'Namida×', 'Never giving up!', 'Not ALONE not HITORI', 'OKAWARI Happy life!', 'Perfect SEKAI', 'PURE PHRASE', 'Shiny Racers', 'SUKI for you, DREAM for you!', 'Tatehoko Tsubasa', 'Totsuzen GIRL'


Friday 2021-12-10 16:02:08 by Dave Woodall

First Automated Nav (with some top of minders)

One of the reasons blogging is essential is because this fantasy land we live is the result of evil taking away all physical materials and means to keep the press free and distributed freely without their oversight and regulations.

But freedom of the press is not limited to paper. A blog is the new newspaper to speak freely. If I speak disrespectfully in your house then you can kick me out. If you don't like what I say in my home, you can leave. Having my own home means I can have the courage to speak up and potentially be kicked out of your house without losing my own bed.

Wars are waged in two ways: weapons and words. wordS are a sWord. By speaking clearly, sharply, and with conviction you will inevitably encounter an enemy who disaproves with your point of view. They will rage against you. Some of those enemies are in powerful places of this grid system. If I choose to rant on their platform then they will delete me - as YT did of Alex Jonze, SGTReport, and other Alternative Media minds who question the agenda / propaganda / narratives.

Building A blog from Bash, up means they delete my domain name(s), account names, profiles, hosting, and I can just reboot without content loss with another set of vendor. If I have your email address, then I can email you when my address changes. It's just like the old days when your best friend moved, you told them your new address. It AKA AKA's. Along my life I've been known as - Dave Woodall, FakeFarm, DadZoo, and.or.at, Sortpop, wwwoodall, dubpipod, ABD, 351-1301, 7603 Clubhouse, my SSN, etc. The power of being literate is that you can speak, read, and lead with words. The benefit of building a blog from scratch is that I can speak freely until the last computer is confiscated.

My occupation is building complex systems as efficient and effective as possible. Fake farm is where I am learning to speak for myself. Fake Farm is my real growth. I have had many blogs over the years. Many times I tried managing navigation by hand before having the skills do automate. It is incredibly tedious and time consuming. Today marks the first time of implementing an automated navbar with my personal blog.

P.S. Email me if you want to be on my email list:

[email protected]


Friday 2021-12-10 16:40:04 by advil

feat: clicking for inventory item actions

One thing the action panel has revealed is how minimally implemented clicking is in the webtiles UI. I realize that lots of advanced players are perfectly fine with never using a mouse, but allowing it makes things easier for newer players who aren't used to huge keybindings. This commit enables clicking on the actions in an item description popup (which is one of my first targets for this kind of change because it's so obviously missing when you right click from the action panel).

The implementation is entirely client-side: it converts any strings like "e(v)oke" into a span with the data-hotkey attribute set to the hotkey in parens. This might be described as hacky (perhaps sending structured action data from the server would be better), but it was really easy to implement and actually seems pretty robust/useful. Part of the trick was that someone (aidanh I assume) has previously implemented an amazing thing where if you just set this data attribute, an on click handler magically springs into existence that sends the hotkey.


Friday 2021-12-10 17:10:50 by Emil Karlström

you may now cat any file from any where, this only took like 100 lines of code to implement and 2 functions but it fuckin works, fuck you rättare


Friday 2021-12-10 18:00:09 by DrunkWinter

[chore] remove eslint & husky

Currently both are a pain in the ass and hard to work with in terms of developer experience, they are more like a nuisance than a helper.

ESlint is hard to work with when unconfigured. We might add this back in later when we have enough time to write the rules.

Husky is reasonable, but with Prettier it tends to change files which are not in the commit. Although this is solvable with lint-staged, the fact that your code looks different on the commit vs what you just wrote is annoying. Also the error/warning messages are a lap of text, with bunch of useless information and it also does not play nicely with a GUI based Git. My proposal as an alternative is having CI testing on a GitHub pull request and when it fails, the dev must manually run lint. At that point it is clear to the dev that transformation has happened to the original code and is able to review it.


Friday 2021-12-10 18:18:22 by DrunkWinter

[chore] remove eslint & husky

Currently both are a pain in the ass and hard to work with in terms of developer experience, they are more like a nuisance than a helper.

ESlint is hard to work with when unconfigured. We might add this back in later when we have enough time to write the config rules.

Husky is reasonable, but with Prettier it tends to change files which are not in the commit. Although this is solvable with lint-staged, the fact that your code looks different on the commit vs what you just wrote is annoying. Also the error/warning messages are a lap of text, with bunch of useless information and it also does not play nicely with a GUI based Git.

My proposal (#97) as an alternative is having CI testing on GitHub on each pull request, when it fails, the dev must manually run npm run format. At that point it is clear to the dev that transformation has happened to the original code and is able to review it.


Friday 2021-12-10 18:28:36 by Sean Gallagher

move search api call to backend

We're just doing the search from our own backend now. Because CORS is a pain in the ass and the whole damn thing adds complexity that I don't need. So now we'll do the search from the backend and just pass the results on to the client; it's not as performant, but it's less prone to breakage.


Friday 2021-12-10 19:22:54 by Marko Grdinić

"11:15am. I am up. I've been thinking up scenes I can paint. Right now, somewhat ironically, the biggest problem is that I do not have a complete model. The base mesh isn't rigged, and neither is it good enough. I need to work on it more. But let me forget about the Luna model just for a bit. I think I'll just paint the still life scenery as it is. The abyssal limbo should be a decent challenge for my first paintings.

11:30am. Let me chill just a bit and then I will start. I want to do a bit of experimentation on clay vs clay strips again. After that I'll play around with creating the scenery.

12:05pm. Let me have breakfast here. Time for the usual routine of breakfast + chores. Then I start. It is time to bolster my will.

It is one thing to develop an excelelnt visual memory, but for a more technical 3d + 2d approach that I want, I should be able to build up skill a lot quicker. Getting to a good level should be relatively simple and painless.

1:15pm. Done with breakfast and chores. Let me finish the chapter of Demon Sword Maiden and I will start.

Let me take a moment to throw some mud on another work.

https://re-library.com/translations/reborn-as-a-transcendent/

While Demon Sword Maiden is solid, the above is absolute trash, and I must have wasted 2h reading 40 chapters yesterday up to 98 when I could have played LoboCorp. None of the characters have an IQ above 80, and the story has the slowest plodding pace I've ever seen. People voted for this over some boring trap story, but neither of them should have been picked up. The annoying thing is that it has an interesting premise, but that can't save it when the author is absolutely talentless. It is so boring.

I wish Re-Library picked up Hellrage instead.

Let me finish the chapter and I will start.

I just can't focus. I had a lot more motivation when I was programming despite the hardship. What I am doing now feels more like a side thing, and I am doing it for the love of the routine. I really have little urge to get up early so I can get extra work in. Maybe if I get some success that would change.

1:25pm. Let me start. If I could get even a consistent 100$ a month that would relieve some of my burden. So far I haven't been able to do anything to affect the outside world with my own power. Better hardware and ML algos cannot come soon enough, but until then I have to do what I can.

First, let me play with brushes. I want to see if the clay brush can be made as manageable as the clay strips one. Let me put down that fractalized plane

1:40pm. It is not a coincidence. Even if I try to fiddle with the settings, clay and clay strips are very different brushes. If it is just adding volume, I should be using clay strips just as I see people doing.

Hmmmm, Let me try it on a sphere. Maybe the most notable difference is how it handles plane edges.

...No, it is the same thing. They are different brushes. Clay strips really feel like strips, while clay tends to inflate.

Ok, enough of that.

2:30pm. Now I am messing around with a walkway. I am really making slow progress. Forget doing the railings directly, I should get them from the asset store. The same goes for furniture. Let me start laying some of that down.

2:50pm. Ah, right. It actually isn't that hard to create a mirror. All you need is to ramp up the metalicity in the shader settings and maybe something else.

2:55pm. It has been a while since I've gone through the CG Fast Track course. I forgot most of the shader stuff.

One thing I want is to put the Spiral logo on the floor somewhere. How do I do that?

Ah, you can import .svg. That is very nice.

2:55pm. I've imported it, but what do I do now with it?

I'll leave that for later. Let me put a sofa in front of the mirror.

3:10pm. A red sofa caught my eye. It is an awesome sofa with 1m polys. The fan really starts spinning when I use it. Evee is really bad with reflections for some reason. I am going to have to look into how to make it better.

At any rate I've disabled the object so as to give my compuer some reprieve. For the logo I am thinking of just slapping it on an image somehow. Let me take a short break and I will do that. After that I'll put some wardrobes next to the mirror.

3:35pm. I am back. Let me resume.

3:40pm. Hmmm, I can't just add an image it seems. How do I use that svg?

https://www.youtube.com/watch?v=ctwiT8LYjVQ How to import a SVG & work with it | Blender 2.81 | Tutorial

Blessed Youtube.

https://youtu.be/ctwiT8LYjVQ?t=136

Ahhh, I see.

3:45pm. Damn it, unlike the web version the strokes have zero width.

3:50pm. Agh, let me draw it directly on the surface. I'll put yesterday's tutorial to good use. I can just trace over the finished graphic.

Oh, the stroke grease pencil has solidity rather than just being 2d. How did that work?

4:15pm. It always had it, it is just that I did not notice it. At any rate, it is really easy to use the grease pencil. I had to normalize the opacity to 1 and join the curves, but it did work. It is not hard at all to make a spiral using the curve tool and then join the curves together. Nice. Now there is a Spiral logo on the walkway entry. I haven't put in the portal yet, but that does not matter right now.

Let me lay down the wardrobes. I am ultra slow. If I had a job as an pro artist, I'd have gotten fired for today's work.

But let me continue going. Let me lay down the wardrobe.

4:40pm. Let me do the railings.

4:50pm. Sigh, what the hell is going on with this carpet. It is grinding my computer to a halt.

4:55pm. I do not get it. This other carpet is only 20k vertices, but it is still to much for my computer to handle.

5pm. Turning ff the visibility helps a little, but it is still slow to drag. Nevermind, let me just go with this carpet. Now railings.

https://youtu.be/WcNiL_rwF9A Modeling a Railing ALONG A PATH in Blender

Where would I be without Youtube.

5:15pm. https://youtu.be/WcNiL_rwF9A?t=456

Ughhh...this is interesting, but not what I had in mind. Let me check out Archipack.

https://www.youtube.com/watch?v=l-aom9PyosM Architectural Modeling in Blender IS HERE! Architectural Modeling with ArchiPack!

Let me watch this.

https://youtu.be/l-aom9PyosM?t=74

This has something for fences. I hope it will work for 3.0.

https://youtu.be/l-aom9PyosM?t=238

This is pretty cool.

https://youtu.be/l-aom9PyosM?t=582

Here are the fences.

5:35pm. Let me take a short break.

https://youtu.be/lmACxKsDtlk Blender 2.8 | Easy handrail and coping tutorial

I'll watch this after I get back.

5:40pm. https://www.reddit.com/r/MachineLearning/comments/r0g2o2/d_ai_safety_needs_great_engineers/

///

Hi Marko,

Thanks for applying to Anthropic!

I'd like to invite you to do a one hour technical interview with us. We recommend Google Colab (Python 3.7.11), but you're welcome to use whatever environment you feel most comfortable with. We'll ask you to share your screen in Google Meet. You should be comfortable with Python's standard library; feel free to use StackOverflow and Google during the interview. We'll save some time at the end to talk about Anthropic and answer questions you might have.

Grab a slot here: snip

Best, Catherio

///

20h ago, I actually got an invite from that AI company I applied to 3 weeks ago. I felt it was strange I did not get a reply, so I was actually taking the time to check my mailbox every once in a while.

5:50pm. Sigh...sigh...sigh...

Ok, actual AI related work is something I'd consider. Let me finish the workday according to plan, and after that I'll grab a slot. If I can pass the interview and get a decent offer, then I'll seriously consider it. Since this is SV company, I'll be expecting something like 150k/year.

Me doing programming for money certainly makes more sense than me doing art for such a purpose.

And I do want money quite a bit. To freeze mom after her cancer become critical, in case of my own emergencies, to upgrade my rig, to get the AI chips.

But mostly for my worn-down ego. There are many reasons to get money, but I really want it for the worst one. Because of my insecurity.

Being aware of it does nothing to change how I feel.

Not able to make money directly via RL was a particularly big blow to my self-esteem. I'll do anything to make the old path alive again.

6pm. Let me leave that aside, I need to watch that vid.

https://youtu.be/lmACxKsDtlk Blender 2.8 | Easy handrail and coping tutorial

Let me watch this.

https://youtu.be/lmACxKsDtlk?t=407

Ahhhhh, I had a lot of trouble lowering the walkway, but I just realized that I could have duplicated a part, filled it, and then boolean difference'd it.

Hmmm, ahh, actually I could have filled the inner paers. That would make ripping them apart much easier. Agh.

Well, let me move on.

6:40pm. Had to leave for lunch. Let me resume. I need to finish the video.

https://youtu.be/lmACxKsDtlk?t=711

I seriously have no idea what this guy is doing. The video is pretty hard to follow. That invite has made me distracted too.

He keeps convertinb between curve and mesh.

https://youtu.be/lmACxKsDtlk?t=832

This is interesting. So that is what that option does.

https://youtu.be/lmACxKsDtlk?t=563

Am I being dumb? Just what are these copings.

https://en.wikipedia.org/wiki/Coping_(architecture)

So stone railings? I see.

https://youtu.be/lmACxKsDtlk?t=774

This thing where he adds depth to a curve...maybe I could have handled the Spiral logo like this?

Let me try importing it again.

7:05pm. Yeah, I could have definitely done it like this. Even if I had a curve, I just needed to add some depth to the bevel and I'd have my stroke weight. Tsk.

7:10pm. Well, things like these happen when you do not know the full capabilities of what you are using. My current solution was no bad by any means.

Let me go back to the video.

7:15pm. I've gone through it.

Hmmmm, if only I could do something like evenly space object along a curve. Also I could start with a straight railing and somehow deform it. If that would be possible, that would be great.

Also rather than being open like now, I'd prefer the limbo to be a transparent dome.

https://www.youtube.com/results?search_query=blender+curves

https://www.youtube.com/watch?v=_3NFtRBDCV0 Blender Curves Array Modifier Beginner Tutorial

This seems on point in what I am interested in.

7:20pm. Ok...

Let me also check out Blender carpets. I really find it bizzare how they grind my PC to a halt.

https://www.youtube.com/results?search_query=blender+carpet

These tutorials seem very on point.

https://www.youtube.com/watch?v=zKvm1bJk8s4 How to make realistic carpet in Blender

I'll leave this for tomorrow.

7:25pm. At this point I forgot what I was supposed to be doing. Railings.

Ah, yes...

If I am doing railings, then I won't need the transparent dome now will I?

I'll need them for the walkway, but there the railings can be straight. Ok, that is one thing.

My first priority then should be to try out archipack.

7:35pm. I am thinking.

I want to do some action scenes of Luna where she is slashing away puppets with a sword.

7:35pm. I should do some scenes of Sol. Like when he comes back to Earth after a trek through hell and finds everybody dead and it being overrun by demons. Also that scene I keep dreaming about of him killing 4m demons in one shot with the soul annihilator. A fierce explossion and then whole mountains are razed. The monster counter says 0/4,321,821. And the level ends, except the intermission screen is no longer a struggle of Sol, but of his hunt.

An image wouldn't capture the sounds, but I want to capture the feeling of cold hatred.

7:40pm. Sigh, which order is it going to all go. Who knows. Let me close Blender here and I'll grab a slot. Let's not waste this opportunity. I should not put too much stock in it. If it goes nowhere, I'll continue my art adventure, but if I get it that can only benefit me.

This is actually the ideal way to apply. I am not desperate for job by any means, and those who apply to a thousand places only to get rejected are simply pitiful. Instead I am picking and chosing at my leisure, seeing where the threads led me.

7:45pm. Ok. I've booked it at 6pm on Monday, Dec 13th.

8pm. https://colab.research.google.com/?utm_source=scs-index#scrollTo=5fCEDCU_qrC0

They recommend I use this. I am not sure what they will be testing me on. Sonce they are doing research, I thought they would be PyTorch users, but maybe they are TF guys. If they really want the regular data science package in TF, I might end up failing the interview. And I am not that familiar with the Python's standard library.

8:05pm. If they are overly critical of my Python specific skills I could fail. I am also not that familiar with things like csv files, and using regular stuff.

The stuff I was working on was all on the bleeding edge of things. If they want me to adjust exotic hardware, implement paper ideas or scale things that will be a better bet. Though don't know how to scale ML models across clusters just yet.

8:10pm. Yeah, there are plenty of way to fail. Even want to prepare, who knows what the questions will be. This isn't a school assignment where I get an A for figuring out what the teachers want me to say. I'll be candid with my strengths and weaknesses and let the chips fall where they may.

I'll think about this like a loterry ticket and try to leave it out of mind until the time comes.

I'll proceed with cultivating my art as if there is nothing the matter.

Tomorrow I will do the glass screen and the railings.

One thing I am really missing in Blender is things like being able to open doors. I'd like to be able to open that wardrobe.

Railings are a surprisingly tough challenge. I could make a piece, but I do nto know how to connect them smoothly. There is really a lot I do not know about.

Now let me close here. Time for some fun."


Friday 2021-12-10 22:27:14 by Sean Christopherson

KVM: x86: Inject #UD on "unsupported" hypercall if patching fails

Inject a #UD if patching in the correct hypercall fails, e.g. due to emulator_write_emulated() failing because RIP is mapped not-writable by the guest. The guest is likely doomed in any case, but observing a #UD in the guest is far friendlier to debug/triage than a !WRITABLE #PF with CR2 pointing at the RIP of the faulting instruction.

Ideally, KVM wouldn't patch at all; it's the guest's responsibility to identify and use the correct hypercall instruction (VMCALL vs. VMMCALL). Sadly, older Linux kernels prior to commit c1118b3602c2 ("x86: kvm: use alternatives for VMCALL vs. VMMCALL if kernel text is read-only") do the wrong thing and blindly use VMCALL, i.e. removing the patching would break running VMs with older kernels.

One could argue that KVM should be "fixed" to ignore guest paging protections instead of injecting #UD, but patching in the first place was a mistake as it was a hack-a-fix for a guest bug. There are myriad fatal issues with KVM's patching:

  1. Patches using an emulated guest write, which will fail if RIP is not mapped writable. This is the issue being mitigated.

  2. Doesn't ensure the write is "atomic", e.g. a hypercall that splits a page boundary will be handled as two separate writes, which means that a partial, corrupted instruction can be observed by a vCPU.

  3. Doesn't serialize other CPU cores after updating the code stream.

  4. Completely fails to account for the case where KVM is emulating due to invalid guest state with unrestricted_guest=0. Patching and retrying the instruction will result in vCPU getting stuck in an infinite loop.

But, the "support" so awful, especially #1, that there's practically zero chance that a modern guest kernel can rely on KVM to patch the guest. So, rather than proliferate KVM's bad behavior any further than the absolute minimum needed for backwards compatibility, just try to make it suck a little less.

Cc: Hou Wenlong [email protected] Signed-off-by: Sean Christopherson [email protected]


Friday 2021-12-10 23:13:14 by Dav999-v

Simplify XML structure in language files

The main change is that translations are no longer in between opening and closing tags, but as an attribute next to English. This has a couple of advantages:

  • It's more consistent and easier to understand
  • The translation is right next to the English string, instead of being separated by the other arguments like the explanation
  • I (and maybe others) don't have to be annoyed anymore by XML parsers automatically deciding to self-close tags that were supposed to be left open, whenever programmatically editing the file
  • Makes it easier to let the language files be imported and exported by Excel. This was my original inspiration for doing this.

So, I've been experimenting with Excel, I knew you could work with XML via macros, but recently I found out it also has a feature to add XML schemas (XSD) so you can import and export XML files directly without having to change your XML structure. Sounds great, except XSD is convoluted, and it's hard to get Excel to do what you want. Having only attributes, or only tag content (or child tags), is doable, and exactly what Excel was designed to let you do with relative ease. Having both attributes and tag content requires special care to get right in XSD, and once you do get it working, it will result in a weird column ordering, the translation header will be the tag's name, exporting as XML also doesn't work most of the time...

Another simplification I did was to remove comment attributes. It would require special code in the language sync feature to load them in and map them out so they can be kept when syncing, which means more complexity, so I just decided to scrap it in favor of manually keeping notes. It's also a question whether translator's comments should be kept around in the end product and if not, how to remove them (an option in the menu?) and when (by whom)... Plus, translators may always use CAT tools anyway.


Friday 2021-12-10 23:16:02 by Favouriteless

Rite of sanctity

cool as fuck particles omg i love it so much


< 2021-12-10 >