2,643,630 events, 1,201,064 push events, 1,978,672 commit messages, 158,096,880 characters
fix(pacer): Catch and retry on bad nonce
I can't recreate this error, though it's happening at night. My hope is that by just giving it a retry, things will magically get better. If not, we'll meet back here again tomorrow, I suppose.
Fixes: #1326 Fixes: COURTLISTENER-54
Deep Work (Cal Newport)
Damn. I read this book, what, maybe two years ago? And re-reading these highlights I made jolted me out of a stupor. Not quite that extreme, but yeah, I need some of that compartmentalizin', batchin' attitude in my life right now.
This work is more about productivity than creativity, but the two concepts are intertwined. You need to be able to do deep work to practice creativity. That much has been echoed throughout all the books I've documented here.
"To practice." I love the idea of creativitiy as a true practice, like a yoga or meditation practice. Or even a doctor's or a lawyer's practice. Why do they call it that? Going back to my original thought though, it's something you bring into action each day as a pursuit of something spiritual.
Editor's Note: I'm also about 4 weeks in to The Artist's Way, so that might be Julia Cameron channeling through me.
I ended up filing away some of my highlights from Deep Work into a totally separate note — splitting them between my actual, physical notebook and an Evernote note that I'm sure will be lost and then found several years from now. Those highlights in particular weren't a match for this guide, but were definitely relevant for how I go about my 9-5 work.
Side note: I love notes in the margins, and I love books that take me back places. I've been flipping through Creativity (Csikszentmihalyi) and I have all these notes scribbled from my time in Barcelona. It gives me all kinds of warm fuzzies just looking at them.
10:12pm Jun 23, tried to add some better documentation. I hate it, and honestly this code is some goofy lookin shit.
Adapt to breaking changes fron 0.32
Fuck you wm4 Thanks TheAMM
arcade battle bug fix (#51773)
-
fixies
-
use LAZYACCESS instead
-
holy shit how did I miss this
the XP wasn't properly distributed, I didn't change the gamerover_check that much from the original but still how in the fuck did I not see that
- I am big dumb
tfw I add more bug than I fix with each commit
-
h
-
double h
Co-authored-by: Caldony [email protected] Co-authored-by: caluany [email protected]
i ate for sex, and it felt so good; all that food let me put it all deep inside o' me; now i'm fat, and idk if i can ever go back, to the way things were, waist of the past; but it's all right, 'cause she can ride me all night, i don't need to lift a finger, not even to her pussy; 'cause that ain't sex; it's called foreplay, and it cums before sex, but it's alright to not do it if you both forget; 'specially in midst of the passionate heat of a foodfight, after hitting a bong; but because there was no foreplay before the really hot sex, it must last long; so take your time, there is no race, not even if you're in the middle of fuckin' her face; as long as it's in a hole, and not in her hand, because that's the very definition of sex, as i understand #hatting(it)off #hatherup #/(hat)+/
"9:25am. I am up. Let me do my morning chilling. I did the right thing by taking the time off to vacation, I really do feel that the thoughts I have made during this period will be crucial towards getting my life in order. They'll only affect a tiny part of my life, but it is dangerous to have chinks in the armor. I need to be an active player in all areas. I do not want to have any weaknesses.
10:25am. Let me start. Unlike in the past week I neither feel stressed nor distracted. Likewise, neither do I feel like not working on this. So till the day is over forget the distractors and focus on what is really important.
Programming - it all comes down to it in the end.
I can only prove my worth by finishing Spiral.
10:30am. It is finally the time to get back to my usual pace. Today feels quite regular.
https://code.visualstudio.com/api/language-extensions/syntax-highlight-guide
Let me start with this.
10:40am. I lied a bit. I am still slightly distracted, but this is around the usual for me. Still, going forward I want to exceed the level of focus that I had before. I'll let this develop naturally. Eventually I will get tired of the past story and start getting interested in the new one. I won't rush this thing.
https://code.visualstudio.com/api/language-extensions/syntax-highlight-guide#scope-inspector
I never tried this out before. Though before I do, let me read the article.
Starting with release 1.43, VS Code also allows extensions to provide tokenization through a Semantic Token Provider. Semantic providers are typically implemented by language servers that have a deeper understanding of the source file and can resolve symbols in the context of the project. For example, a constant variable name can be rendered using constant highlighting throughout the project, not just at the place of its declaration.
I really should take advantage of this.
11am.
{
"scopeName": "source.abc",
"patterns": [{ "include": "#expression" }],
"repository": {
"expression": {
"patterns": [{ "include": "#letter" }, { "include": "#paren-expression" }]
},
"letter": {
"match": "a|b|c",
"name": "keyword.letter"
},
"paren-expression": {
"begin": "\\(",
"end": "\\)",
"beginCaptures": {
"0": { "name": "punctuation.paren.open" }
},
"endCaptures": {
"0": { "name": "punctuation.paren.close" }
},
"name": "expression.group",
"patterns": [{ "include": "#expression" }]
}
}
}Now that I am really studying this, I see that this is actually a mutually recursive definition.
11:05am. Focus me, stop taking a peek if you've gotten a reply to that parser combinator thread.
11:10am. \l is really easy to write. If I am going to use Unicode in order to get the forall, I might as well introduce the Unicode lambda.
11:20am. Let me take a short break.
11:50am. I am back. Let me resume reading the guide.
It is coming to me. Though I'll use the semantic token provider, I should implement basic keyword highlighting.
12pm.
There is one additional complication for injection languages embedded languages: by default, VS Code treats all tokens within a string as string contents and all tokens with a comment as token content.
Definitely going to need to think about this...
12:10pm. Done studying the document. I wish I could say I am focusing my whole being on this, but it feels like I am giving it 20-30% of my all at best. Well, it is fine. Reading documentation was never exactly a great locus for my motivation.
At this point, I am mostly daydreaming rather than working. Still, I think I get most of the things in the document, what I do have to do next is get familiar with...
https://macromates.com/manual/en/language_grammars https://macromates.com/manual/en/regular_expressions https://www.apeth.com/nonblog/stories/textmatebundle.html
12:25pm. Ok, I've decided. Forget this. I am going to go with the server side tokenization, not because I do not want to use regular expressions, but simply because I do not want to duplicate effort. If I start using Textmate grammars, then I will be putting stuff on the client side which I do not want. I should strive to minimize the amount of code I write on the Typescript side.
12:30pm. Now that I've confessed to this the goal I should take on is clear - do the tokenizer for Spiral. I already have one, but the one I need needs to be incremental and fit for use in the server.
I admit, I haven't though that I would be doing highlighting using the tokenizer, but that is great as it gives me a natural thing to strive for. It seems a bit convoluted to go through the server, but as syntax highlighting is the very first step, this is a great opportunity to do some ground up work.
12:35pm. If only I could get the rest of my mind to look straight that would be great.
Once I start programming, my focus should go up. And that is what I am going to do after the break.
In the tokenizer, I am going to get rid of the dependency on FParsec so I can detect tabs properly. At this point I do not actually understand how to do deal with changes just yet. So before I start, that is the first thing I should check out. Can I get the initial file as one big change? That would be great. If not, I'll just do a bit caching in the plugin to make sure that the server gets the proper info.
...I'll know if I try to save a file with no changes whether VS Code is paying attention to this kind of thing. Probably not...
12:45pm. I'll leave it for later. Let me get some food."
repo: remove Dependabot configuration
Maintaining dependency upgrades with Dependabot in a Gatsby project is just too much pain in the ass.
It's not you, but it's you. Sorry!
Updated license to "do what the fuck you want to but with conditions"
LS IS NOT PWD, YOU IDIOT SANDWICH!!!
Ok, maybe I'm starting to get a bit insane. But all I did for last few days was staring at this goddamnit code.
Signed-off-by: GuzioMG [email protected]
Take 4, syndie items changes (#1278)
-
Take 4, holy shit im gonna die edition
-
radbow/ebow
a compromise
-
small tweak
-
Update uplink_items.dm
-
Update special.dm
rm Fody, P.S. Fuck you! init caliburn
Signed-off-by: Shark [email protected]
"https://hinjakuhonyaku.com/novels/rose-princess-of-hellrage/rose-princess-v1c12/
While her conscience was as good as dead, being able to plan and carry out massacres, her sense of shame seemed to be quite livid.
Kek.
1:50pm. Just a bit more and then I will start.
2:15pm. Let me resume.
Yes, so the first thing I need to check is how changes work.
2:25pm. I am playing with this and it seems only the changes would get set over the wire.
I do not understand how deletions work.
doc.forEach(x => window.showInformationMessage(`Chaged: ${JSON.stringify(x)}`));Ah, this is only positive when I erase something.
2:30pm. Ok, I understand the way the changes work now.
2:35pm. Hrmmm, should I just send the whole file along with the latest changes. That seems like it would be the easiest solution.
Ah wait...
This might not trigger on empty changes.
const spiProcess = async (file_path : string, doc : readonly TextDocumentContentChangeEvent []) => {
window.showInformationMessage(`File: ${file_path}`);
doc.forEach(x => window.showInformationMessage(`Chaged: ${JSON.stringify(x)}`));I forgot something obvious.
What I want to do here is keep track of the latest changes from the server. Once I get the first reply I can just send the changes, otherwise I should just pass whole files back and forth.
One thing I should change is the way I do ranges.
type CfgPos = { Line: number, Column: number }
type CfgRange = { Start: CfgPos, End: CfgPos }As it turns out, I can serialize the regular range just fine.
Chaged: {"range":[{"line":1,"character":5},{"line":1,"character":6}],"rangeOffset":18,"rangeLength":1,"text":""}
Is this how it is laid out internally, as a tuple?
Let me open it in the debugger.
2:45pm. How strange, why is the debugger not working?
...It is not working and I do not know why. It worked for me before. Right now the breakpoints are just not triggering.
2:50pm. I decided to try restarting it. Now the stupid thing is downloading packages. Why does it always do that? Can't it put them in some place and sit on them?
In the debugger it looks different. Hmmm, I have no idea. At least for sending things backwards I should use the different range.
Chaged: {"range":[{"line":1,"character":5},{"line":1,"character":6}],"rangeOffset":18,"rangeLength":1,"text":""}
Still, let me adopt the things here as a convention.
2:55pm. Ok...since keeping track of what the client is receiving is an optimization I will eschew it for now. I'll leave it for when the tokenizer is finished.
3pm. Hmmm...
{"range":[{"line":1,"character":5},{"line":1,"character":6}],"rangeOffset":18,"rangeLength":1,"text":""}
For the position, let me adopt the {"line":1,"character":5} as a convention.
After that I will make sure the files are being sent properly. After that I can start work on the tokenizer.
type CfgPos = { Line: number, Column: number }
type CfgRange = { Start: CfgPos, End: CfgPos }
Since this is not true, let me try getting rid of it. Instead let me see if I can just JSON.parse to the VS Code's Range.
3:10pm. Let me try it out.
3:15pm. The debugger is fucked up again.
(x : [string, [Position,Position] | null][]) => {
window.showInformationMessage(JSON.stringify(x));
const diag : Diagnostic [] = [];
x.forEach(([error, range]) => {
if (range) {
diag.push(new Diagnostic(new Range(range[0],range[1]), error, DiagnosticSeverity.Error))Let me just do this.
3:25pm. Ok, I finally did it. I've been gripping with stupid errors. It turns out I cannot just parse the same form back. Seriously, OO is the worst. The object methods make them non-serializable.
Now where was I?
Yeah, I need to deal with the tokenizer.
Seriously, why did I even mess with the ranges?
So I can send those changes more succinctly. Let me modify the file sent so that I am also sending the changes along.
3:35pm.
type Changes = { range: Range; rangeOffset: int; rangeLength: int; text: string }Half my mind is not into it so I am slow as a snail at this.
| File of {|spiPath : string; spiText : string; spiChanges : Changes []|}Let me make the file this.
3:45pm. Ok, it works. Now which step is next?
3:55pm. Before I can even start the tokenizer, I need to bring out the block structure of the code.
4:05pm. I am still thinking about it. What would be the best way to do the initial split into blocks?
...I feel like stepping away from the screen for a bit.
4:10pm. I've been thinking of adding the doc keyword, but I should go against it. I think only the top level functions should have embedded documentation. Meaning the block should include the docs. Also it should include the and statements.
4:35pm. Ok, I've had some time to think about it. Forget tokenization for the time being and just focus on blockization.
Let me check out something in F#.
// Identity
let f x = xThe above is just a normal comment, but...
/// Identity
let f x = xThe triple slash turns it into documentation.
I've been thinking and have come to the conclussion that this is not necessarily the easiest thing to do. Rather...
// Identity
let f x = xI will make the above valid documentation. The above will be a part of the same block.
// Identity
let f x = xThis on the other hand won't. The empty space will act as a blocker making it a part of the previous block.
4:40pm. Let me do it. I could do it in the next hour...no actually, applying the changes as well, I am not sure that I could really do it.
4:45pm. Arghhhh...I am not sure. As changes are made, I am going to need to split and merge lines, and once I have blocks that extends to them naturally as well.
If it was merely processing a file from start to finish that would be easy, but just what do I do with an already finished structure to which changes need to be made?
4:45pm. I've decided that I'll only allow documentation for the outer functions. And I will have them be non-indented.
5pm. Let me go back to bed for a while longer. It does not seem I will be doing any programming today.
5:05pm. Blocks are a natural idea, but now that I am sitting here and thinking how to actually do them, I am coming up empty. There is no benefit to doing them as I envisioned it originally.
Damn, on paper I know how to do all of this, but how to do arbitrary changes cleanly with only small amounts of work expended is new to me.
7pm. Shit, it is already this late? I've been busting my head in bed the whole time.
Honestly, nothing is coming to me. I can't figure out a way to deal with the problem in front of me elegantly.
I mean, splitting the lines into blocks is the easiest thing ever, but considering blocks are going to be servers, I am really having issue figuring out how to do it without regenerating the whole thing.
Come to think of it, the supervisor in the prototype was complex as well. Blocks are nasty to modify as any change can split or merge them.
7:05pm. There is no helping it. Tomorrow I will just start with small optimizations and lead into the bigger ones. At the very least, all this thinking warmed up towards doing the basic implementation.
If you can't do it, just think. If you can't think, just do it. Since I am in the later phase, I'll take this advice and just dive into it.
Once I have working blocks that will be the big enabler of everything that will come afterwards."
TUTOR shit
i did all this with my tutor!!! now for the real suck ass stuff on my own
Prettier file names with path. IT WAS NEEDLESSLY HARD, FUCK YOUR PATHS, RUST!
warped fork okrs
With remarks:
-
I'm simply not scoring the
go-ipfs-related OKR. To the best of my understanding, the efforts to roll go-ipld-prime forward into more of the go-ipfs codebase are currently stalled because those efforts were deprioritized during the go-ipfs v0.5 release cycle for reasons that had to do with go-ipfs and are unrelated to IPLD. (And yes, that is the reason why I insisted on phrasing this OKR the way it's phrased; this was fairly predictable.) We did do a fair amount of work to chart out how this migration could be performed, and I'm at least happy with that; but I wouldn't say it's on track to go further until we have go-ipfs increasing prioritization and consistently carrying a baton on that side (particularly considering that the main result of our how-to-migrate research was confirming that yes, the internal package graph of go-ipfs is such that this change is very likely to either require acting in a very large and decisive sweep (which requires maintainer energy and active engagement); or, requires prolonged efforts in advance to break things down into smaller and less densely inter-related pieces (which might be the way to go, but again has a lot more to do with go-ipfs maintainer energy than any other limiting factor)). -
Codegen gets a partial score. It's going along, but continues to be a massive task. Some demo usage has been made, but we're still lacking a CLI and a finished end-user-accessible experience, so it's hard to call that complete. The feature table of all schema features we intend to support is also still less than solid checkmarks (although, hey, we have a detailed feature table now; that too is progress). The good news is we're iterating and getting close enough to real demo usage that we're starting to get some interesting practical design feedback from that process.
-
Link loading gets review: let's give that partial credit. It's gotten review! There's a sizable writeup issue on it now. That was the goal; I suppose I could mark it 1.0 with no shame. I've also made a couple tries on drafting a better interface... however, none of that has born fruit yet; something definitively better remains illusive. I'm refraining from making changes just for the sake of making changes without confidence that it would result in a better solution than the current situation.
-
Traversal and link loading docs: yes, we have examples now. They use the standard godoc example/test system. Yay. Done. Could also use more? Definitely. But we no longer have a void.
Final, typical disclaimer: none of these are really OKRs; they're not really measurable; we don't separate objectives and results; and we just plain don't have a team big enough and unitary enough that OKRs make sense in the first place, so they degenerate to task lists, which in turn they are spectacularly not good at being. The same laments on the inaccuracy-fermentation this process creates apply every quarter; I note them duly.