etienneburdet.com
😱
69 posts
51 followers
193 following
Regular Contributor
Active Commenter
comment in response to
post
Salvaging end of season too, enjoying down time in between jobs. Perfect time to learn React Query for the next job too 👌
comment in response to
post
Just in case, or most likely for other readers: Leslie Vaillant in Probably approximately correct offers a unifying theory for how nature is "smart" and humans are "smart". You could then argue that intelligence was embedded in matter since the beginning and the "algorithm" just self improved.
comment in response to
post
Can't think of something "real world" that I couldn't achieve with :first-child and :last-child (excluding :has). If you're first and last, well…
Maybe some justify-self stuff?
comment in response to
post
It's an headless select!
(well no, showdown is good enough if the answer and rightfully so)
comment in response to
post
That's a convoluted way to say "I shot myself in the foot"
comment in response to
post
I think this is the first practical usage function*/yield* in JS I ever see 😂 It's a good one… !
comment in response to
post
I mean, if you are going the way of requiring a full build tool suite anyways, you might as well go the full way and compile for real, moving away from Babel plugin etc.
But the compiler will have to remain optional because of how much legacy React is already around.
comment in response to
post
Isn't this the deprecated peer dependency problem with platform specific version of Nx?
I got this all the time due to some minor resolving, resulting in nx-arm64 or nx-darwin not being found on CI.
comment in response to
post
xkcd.com/1319/
comment in response to
post
I've been waiting for that for a while. Shipping dual is exhausting for what we do: monorepo that uses ESM in local dev, but has to ship UMD for our core platfform.
But type: module still breaks our main app (which has some debt indeed…).
comment in response to
post
You could easily make a training board for you that's still fun for the kids with a setup like that!
comment in response to
post
The exciting part is now keeping old phones alive! Especially when version support drop and you to start hacking custom builds…
comment in response to
post
Un des rares articles qui remet les problèmes dans le bon ordre d'importance : l'urbanisation et le transport sont de loin les plus gros leviers. Et pourtant dans les stations qui ont de l'altitude, ça construit au delà de tout sens commun. Et plutôt des spas que des transports...
comment in response to
post
Continue kinda have this problem as well with tab autocomplete, it's like I have to set up a mental filter on the suggestions until I really want it.
Room for UX here…
comment in response to
post
Wait what, it's my gym! Ha no it's the one from the 15th… come to St-Lazare, there's half Paris dev scene climbing there, we could hold Vite conf in between routes…
comment in response to
post
Designer:Â You want icons as SVG or Figma?
Me: I want them as Library/Iconame
🤓
comment in response to
post
Careful with the running. You start jogging around your block and before your whole life is organized around training schedule and running mountains for hours...
Thanks for React Query! Reading a lot about it, it's great inspiration!
comment in response to
post
Yes and we have already done this accumulator stuff a few times already. I'm sure they had something in mind... Hope there will be official solutions!
comment in response to
post
Well yestarday I started with something like that. Since my Var is a union, IÂ ended up with a union when matching multiple vars declaration in a row.
I think that was the intended "tricky part". I might be wrong… But the "string" solution is simpler indeed!
comment in response to
post
I feel like we're missing the interesting bits by not explicitely looking for 'let' | 'const' | 'var' but rather checking for = … ;
I'm pretty sure the authors had something else in mind.
comment in response to
post
And here is a contrieved one without NoInfer (that's basically reimplementing NoInfer the old way…)
I really though there would be a solution with curried function and type guards, but it doesn't seem so…
comment in response to
post
Ok, seen @tkdodo.eu and @grad.dev solutions… didn't know about no infer. I knew it was cheating a bit to hard code Color, it wouldn't work with other colors combinations…
comment in response to
post
I still think it could have used some incompatible functions tests.
I am kinda disappointed it's so hard to type [ReturnType<F>] extends Params<G>.
I'm gonna dig, I'm sure I'm not the first one to try that. Must be a 5 years old issue with 1M answers somwhere.
comment in response to
post
Now echo chamber? Progressives are diverse enough, known for debate, nitpicks of all kinds…
Universities for exemple are by far and large progressive, but more known for heated debate about just anything than "echo-ing".
comment in response to
post
I don't have any data, but my guess is people who care about choosing their algorithm and recovering are gonna be "progressive"—almost by definition, it's progress from the social media status quo.
comment in response to
post
For reference, ts-toolbelt curry used by Ramda:
github.com/millsp/ts-to...
comment in response to
post
I would think it's an error. Currying would preserve the order, no? (but return a function indeed, which I guess is the point of the test). Else I really don't get how spikedcurry would figure out order.
There are valid answers though (haven't look at them yet, I have to do mine 😅)
comment in response to
post
To me it's "normal":Â older frameworks produce more complex, real world problems over time and people are more likely to hit the wall.
I just don't know how you would normalize that…
comment in response to
post
😱
Honestly, I think 'no-explicit-any' should be turned on and cause errors. I don't think I've seen one where it wouldn't lead to better solutions.
comment in response to
post
Don't think mine is more elegant, just more factored with subtypes. But I pretty much do the same: check for "", clean left/right, find next destination, count fuel…
comment in response to
post
Used to like a mix of Twist & Todoist. Todoist is very personnal list, simple and efficient but can be linked to a lot of things. Twist make it easy to share & all. But it's hard not to sprawl over github projects, shortcut/jira etc.
Of course, you could do it in Notion…
comment in response to
post
Yeah I came across yours with return type. I actually tried too after resting some blog on invariance but didn't think of the second param with a default.
comment in response to
post
Hardcoded the naughty/nice too, I couldn't find any logic. Just assumed it was a given list by some reinder…
comment in response to
post
No, I did a bit research for typing "new" stuff and then went and check some others (the $K = Extract<…> is pretty nice). Still reads a bit alien to me, never actually used this solution in real life 🙈
comment in response to
post
Yeah, it's ramping up quickly!
comment in response to
post
Haaaa, Extract! Nice one. I knew mine was overcomplicated somehow.
comment in response to
post
Quake 3 (DeFrag). By far.