3,069,924 events, 1,553,567 push events, 2,404,304 commit messages, 177,588,179 characters
i fucking hate computers
raw socket programming is fucking stupid
fuck all this, going OOP
I can't keep track of all these edges and shit with a functional approach, going to try writing a bunch of silly classes and seeing if that screws my head on straighter.
fixed deviantt presence effect (friendly npcs take more damage) still existing without the debuff icon mutant eye/sparkling adoration love rays use local iframes (will not interfere with your attacks anymore) wof can't deathray for a second after entering p3 (should never instant deathray in singleplayer anymore) sand elemental bonus life removed and spawns bonus sandnados slower added recipe for heart chocolate flying dutchman nerfed, longer delay between bullet barrages and bullet barrages are much shorter twin rangs will always produce one burst of projectiles upon hitting an enemy or solid tile dubious circuitry probes: probes and lasers now do 35 damage (was 50 and 80(??????)) probes can fire lasers twice as often manual probe only fires when you are using a weapon and its not a mining tool lasers cannot break larvae(?) new mutant's gift description sibling weapons: have new recipes sparkling love, styx gazer no longer dropped by devi/abom devi/abom drop a new material on nohit, no longer first attempt locked mutant drops material on nohit or 10% chance eye of cthulhu always drops 5 life crystals plantera always drops 5 life fruits deviantt more delay between teleports on rune wizard attack random velocity for spectres light orbs are denser when splitting increased skull charm drop rate to 5% (was 2%) golem boulders fade in much faster geysers spawn on ceiling with boulders, other stomp produces floor geysers removed bats spawning in hordes spectre healing lasts longer than before and moves a bit faster again fixed(?) ice queens spawning multiple sets of super flockos ice queen super flockos fire tight spreads of frost waves, but fire slower than before nerfed martian saucer core laser spray duration and random spread prime spin dark star timeleft reduced, but MUCH denser than before twins retinazer burning aura shrinks while doing deathray retinazer aura inflicts oiled and on fire too fixed the despawn meme post plantera dungeon armored bones inherit angry bones bone spray and baby guardians cultist's nebula spheres now rubberband (only speed up when they're a certain distance away from you) active summon weapons: flesh hand, destroyer gun max requirement reduced to 3 fish stick max reduced to 4 all tooltips now state how many empty slots to get their full power stinger is no longer an ammo added echdeath
bugfixes & issues
inventory works at keeping track of items in real-time, and also can save to file (though i've removed this feature until i figure out how i'm going to handle saving progress of all other game data). stackable items stack properly and a new stack is generated when an item's max stack is reached. however i'm thinking of removing this feature and instead keep a maximum of 1 stack of a particular item type, rather than give the player the option to fit every inventory slot with a stack of potions. to do: the next important thing to figure out is removing items from the inventory. also on picking up an item, a notice should appear, indicating it's been placed in the inventory. the inventory UI should display a sprite of the current slot's item. since originally the player's inventory was saved to a file, this means i will have to keep track of the inventory another way. currently, the Exit script takes the exiting player's inventory and stores it in variable new_inventory, which is passed as part of the attributes argument in the global goto_scene function. the player is then re-instantiated. after hours of debugging, i cannot seem to reconnect the GameManager's initialize_player() function to the inventory UI. it does so successfully on first instantiation, ie. when the program starts. but when the player gets reinstantiated in a new scene, no methods i've tried seem to work to reconnect the UI. it disappears. i am going fucking insane trying to fix this. i have made 0 progress for 3 fucking hours, i have 0 fucking errors, and there is zero fucking discernible reason as to why this doesn't fucking work. the inventoryUI's _on_player_inventory_changed function simply stops running. it runs once. then it doesn't ever again. despite creating a separate "reinitialization" signal. despite making another connection there. what fucking trash. fucking garbage waste of fucking time. connecting the GameManager's reinitialize_player function to a key, works however! fuck this fuck this fuck this. it's not fucking complicated to get a function to run. you tell it to fucking run. i've told it to fucking run in so many fucking ways, and it won't. it won't even err. it's infuriating. but when i bind it to a fucking keystroke it works. i don't it to be hidden by a keystroke, i want it to run like a fucking computer program should. why the fuck doesn't the player get fucking reinitialized when i fucking tell it to? so many fucking times?
Goodbye, Artzie Music
I don't think a git repo is where you're supposed to eulogize a YouTube channel, but I also heard your not supposed to use it to store binary files - and I've very clearly ignored that. What's one more?
Artzie Music was a Youtube channel that superimposed FutureFunk and Vaporwave tracks over anime GIFs. It sounds dumb and simple; I guess on a certain level it was. But that was the beauty of it - Artzie packaged together music and visuals into a cohesive aesthetic. The visuals informed the music and the music informed the visuals. Again, not revolutionary by any means but when I found Artzie Music it was revolutionary to me.
It's been said that music has a special place on YouTube because it almost resembles a records store. Songs aren't categorized by artist or genre, but are instead presented almost entirely on the basis of their thumbnails. In that environment, Artzie was like a candy store. Each thumbnail and subsequent GIF was hand-picked - the best designs, the best colors, the best animation.
That's where I found Vaporwave and FutureFunk. I'm not sure what about it struck a cord. It just felt like, for the first time, music actually resonated with me. I'd liked music before, but the music on Artzie felt like it drew ideas out of me, out of thin air. I'm not kidding about that - when I found Artzie is when my programming really kicked into overdrive. Webservers, rendering projects, robots, machine learning. It was Artzie driving it all. I was even listening to Artzie Music while working on this project!
And there was this whole community of people who had the same vibe. I always felt like my music was too weird for most people, and I make a general point of not mentioning it. But in Artzie there was a great mass of people who actually liked the same nonsense I did. It really was the "#1 Destination for Unique Music".
But, there was a hard truth to Artzie Music. A lot of Vaporwave and FutureFunk really pushed the limit on what constituted sampling and what was literally just a slowed-down or sped-up version of a song. And in nearly all cases the songs obscured the songs and artists that were being sampled from - which seems like a genuine ethical problem. Frequently, the sampling credits were the top comments on each video, as everyone came together to try and decipher where and how a song came to be. And those anime GIFs were presented more or less unedited. It was a copyright minefield, and a mass DMCA based on those GIFs was what did Artzie in.
I know why copyright is important. The sampling issues I outlined above are a really good example of that! I just - I wish my favorite weird little virtual record store had been spared. That we could have just had that.
But we had it for a while. And a lot of good artists I first saw on Artzie are still out there, making music. And Artzie broadened my horizons. Turns out, a lot of what I liked in Vaporwave and FutureFunk was in the samples. It was in Disco, R&B, Jazz Fusion, City Pop. And there are still other channels that do work similar to Artzie, and there may come another successor (here's hoping)! But it's not a guarantee, and I guess that's the scary part.
Some people may proclaim the death of FutureFunk, of Vaporwave. To me, it seems like each "death" (because they apparently die about every other year or so) is a transition to another phase, another wave of artists. And it's true that maybe A E S T H E T I C music might die out as a subculture - the memes and aesthetic principles going towards the wayside. But I think that what resonated with me originally in Vaporwave and FutureFunk will resonate with others going forward. I think there will always be a market for this media, even if the media can't trace it's heritage back to the starting days of Vaporwave and FutureFunk.
And besides, the Dodgers won the World Series. Hard to be sad when that happens.
update to portugal and rubber
- rubber from 8 to 6 per refinery, removed rubber outta africa
- portugal gets max infra in its core states.
- indsutrial advisors availible at start for them
- partialat start. war eco at war start, total mob in april 41
- meme fleet added in can remove if u like since no one gave a shit and ppl were amused
- final oob 2 light, 2 med and 2 heavy tanks. 2 marine, 2 mountain, 2 inf (up 1 heavy and 1 med)
- 1k air, 600 fighters, 400 tacts
- buffed fuel idea and gave portugal max refining fuel tech to help fuel new oob
- added in 2 generals with high stats incase i have to give a fuck to dislodge fuck tards (only used to contian memes or engsieg/end game)
- tech added to portugal
"9am. https://fundresearch.fidelity.com/fund-screener/
So it turns out Fidelity has an excellent screener for mutual funds, not just for its own family, but across the board. I've been trying to find this for days. I've been trying to talk myself into ditching individual stocks, but I was not sure that ETFs on their own would give me a wide enough breadth of options. With this, that issue has been settled.
Well, it is not true that ETFs are not broad enough in their offerings. That might have been true one or two decades ago, but today things are different. It is true that there are some offerings that the two do not intersect with, here I am just fulfilling a psychological need.
9:10am. Even if it is going to be a decent way into the future that it is going to be useful, I will commit myself to this new way of thinking today. If you don't want to spent time doing it, actually trading a basket is not easy for a lone trader - it is much easier to go through an intermediary.
Back when I was starting out, it was my dream that trading should not take more than 15-30m of my time a day. If I internalize this new way of thinking, that will be a success.
9:20am. Let me read Mahoako and Hero's Daughter, then I will get to work. I need to deal with the mess that are the validator and individual file servers. I might make the validator a regular function since it does not need to be concurrent.
9:25am. This recent rally, and the action of the past 6.5 years that I've been away from market really did fire me up. The fact that I'll be able to make peace with my position in the financial world is something that I did not expect. My effort was just way too much. Had I given up quicker, I'd have come to the answer sooner, but then again, if I were the type to give up, I'd never have attained the needed experience.
9:30am. The same goes for programming.
Let me do my morning reading so I can do my morning programming. I am wasting too much time.
10:15am. Let me start.
10:30am. I am thinking.
Basically everything is a huge mess. By everything, I really mean the way errors are handled.
I have the individual file server, and it is doing validation, but then I am also making and validating in the validator.
To make matters worse, the validator is supposed to be propagating errors, and yet it does not have the capability to see most of the errors because they are being handled in isolation in the file server. I am doing the same function in multiple places.
10:40am. Ah, yes, I nearly forgot. If the individual file server fails to load, it will raise a fatal error instead of just showing up as a regular one. That is nasty.
11:30am. I am thinking. I actually have a lot of inspiration right now. I have a great idea how to separate loading and validation. I figured out how to do concurrent depth first traversal.
12:30pm. Everything is just coming to me. Now only do I know how to do the parts that have been giving me trouble thus far, but I know how to do the entire compilation pipeline without any of the design complexities and disadvantages of my current approach.
12:50pm. For suppressing parser inputs to the typechecker, I do not have an elegant solution, but I can think of couple of rough ones. Well, this is just a small fragment of the overall thing, so I do not need to worry about this.
12:55pm. Yeah, I can do this. It all fits now. I really was short of inspiration, but with the ideas I have in mind now, I do not have to make a hugely complicated chain of state machines. For build itself, I can do a concurrent DFS much like for loading files and validating them.
Let me stop here. I am overdue for breakfast."
调整空行 diff --git a/index.html b/index.html index 17300f0..f95b03b 100644 --- a/index.html +++ b/index.html @@ -20,12 +20,19 @@
-
<br /> <span class="space"/><span class="comments">* don't be dismal ,don't be wild!</span><br /> -
<br /> <span class="space"/><span class="comments">* in the day of grief,be mild</span><br /> -
<br /> <span class="space"/><span class="comments">* Merry days will come,believe</span><br /> -
<br /> <span class="space"/><span class="comments">* Heart is living in tomorrow</span><br /> -
<br /> <span class="space"/><span class="comments">* Present is dejected here;</span><br /> -
<br /> <span class="space"/><span class="comments">* In a moment , passes sorrow;</span><br /> -
<br /> <span class="space"/><span class="comments">* That which passes will be dear.</span><br /> </div> <div id="loveHeart">
Finish the god damn Data mining code, fuck the Association rule
fix(bootstrap): same-account modern bootstrapping still requires policy ARNs (#9867)
The current bootstrapping experience has sharp edges. It requires you
to pass --cloudformation-execution-policies in all cases, even if
you just want a "simple", same-account bootstrap stack. If you bootstrap
your own accounts, you probably don't have a centralized SecOps
team that is limiting you, and you're probably intending to deploy
everything using the CDK, which means you're looking for AdministratorAccess.
In effect, you are forced to constantly Google and copy/paste the ARN
for AdministratorAccess (because who can remember that)
and it's a bad experience.
In the case of bootstrapping an account just for use "by itself", however (which is the 90% use case for bootstrapping), the trust boundary is very clear and there's no risk of privilege escalation between accounts.
Add an optimization where in the case of a "simple", non-cross account bootstrap, we'll default to the AdministratorAccess ARN for you.
Once you establish --trust with another account, we'll still
force you to spell out the execution policy you'll want to use though.
Fixes #8571.
By submitting this pull request, I confirm that my contribution is made under the terms of the Apache-2.0 license
api art photography cli aws llvm computer-graphics computer-vision ocr ray-tracing svg ethereum databases dynamodb postgresql sqlite 3d-modeling blender color design-inspiration design-systems design fonts icons user-experience distributed-systems e-commerce environment css html games spatial-analysis displays hardware ideas keyboards parenting macOS pytorch generative-adversarial-networks homotopy-theory math statistics decentralization dns http matrix networking rabbitmq tcp neuroscience github-actions github docker kubernetes file-systems linux nixos operating-systems quantum-computing podcasts bash c cpp flutter erlang go-libraries haskell java javascript d3js jest blitz nextjs react-hooks svelte julia lua processing python-libraries python r ruby rust-libraries swift-libraries swiftui tcl hashing json probabilistic-programming reverse-engineering software-testing git marketing blogs sleep text-editors vim-plugins tools config-management fish shell video browsers google-chrome cms web-components web-performance web-scraping web consultancies writing
Keyboard press fix
fuck you Unity you stupid bitch
Migrated to Next.js v10 and away from Gatsby v1 🎉
For about 3 years I used Gatsby as a static site generator for jonrh.is. I started using the v0 beta and then migrated to v1 later on. Gatsby dependencies were unstable to the point that in 2020 I could no longer make changes.
I have used Next.js v9 in two smaller projects and the developer experience was significantly better. Especially when used in combination with Vercel.
Deleted Rollbar error tracking for now. I may add it back later.
Deleted html.js. layout/index.js and _app.js now serves that purpose.
Deleted blog-post.js file that served as a blog post template. In future commits I will either convert all markdown files into React JavaScript files or use MDX.
Modified the Font Awsome minified font slightly. Changed the URL of the fonts that it loads. Before the URLs were assuming the font files were located relative to the CSS file in a sibling /fonts folder. However for now they have been relocated to the public/fonts/ folder so the URL path is now absolute. A bit of a dirty hack but I'm okay with it for now. I hope to move away from Font Awesome in the future and use something lighter.
Moved the required contents of typography.js (sets the font sizes, etc) into layout/index.js. I would like to move away from this dependency later on but for now I am focused on porting the site from GatsbyJS to Next.js with ideally no visual changes.
Files deleted that were only relevant for GatsbyJS:
- .babelr
- gatsby-config.js
- gatsby-node.js
Greatly simplified package.json. Removed all dependencies that were no longer required. Most of them were Gatsby plugins or utility libraries required to work with Gatsby.
about.js / blog.js / dashboard-consulting.js / index.js / portfolio.js. Refactored to use new Layout. Removed React prop types because most of them were very vague anyway (just telling an input is an object). Removed GraphQL query strings that were used by GatsbyJS.
New file: pages/_app.js. This is the root file of a Next.js v10 app (see docs). This is sort of the equivalent to the html.js file that Gatsby v1 used (no longer used in v2 if I remember correctly).