Is NPMX hosted in Europe or only its pod instance? Also, is there an org behind NPMX? This is hard to find these info on the npmx website.
Posts by Victorien Elvinger
Did OpenAI and Anthropic use the same #AI to generate their announcement? Or is there some AI convergence, resulting in poor diversity?
This makes me concerned about the future maintainability of such code. I mean, this applies to any large amount of AI-generated code.
Three years ago, I opened an issue to bridge the syntax gap between TypeScript and the Flow type checker. It is now closed :) The Flow team have done an amazing job for bringing Flow closer to TypeScript while retaining their unique features like exact object types, opaque types, nominal classes.
Regarding miette, you could easily migrate to the fork maintained by OXC: github.com/oxc-project/...
Otherwise, annotate-snippets-rs (github.com/rust-lang/an...) that is the next-gen diagnostic lib for therust compiler seems to be a good choice.
This could make an excellent case for a Clippy lint rule :)
According to the articles you shared, free will makes us better individuals. However, I believe that, as a society, we should consider determinism if we are to build a fairer, more pragmatic and compassionate society.
Personally, I think that, given the amount of scientifically discovered determinism (physical, social, etc.), it is difficult to believe in free will. I can believe in some randomness at small scale, however if we define free will as random will, it is no longer free will...
Taking your words, if you made your choices based on a framework of thoughts that is determined by your experience/knowledge, where does that leave free will?
I don't get the paradox part. If you believe in full determinism, then free will is an illusion as well as the choice of believing or not in that.
You can also be interested in:
- alloc_instead_of_core
- default_numeric_fallback
- float_arithmetic
- format_collect
- inefficient_to_string
- large_include_file
- large_stack_arrays
- std_instead_of_alloc
std_instead_of_core
- struct_excessive_bools
- trivially_copy_pass_by_ref
Node excitement 😉
Congrats to @marcoippolito.dev on today's landmark Node 22.18 release. It is the first LTS release to ship unflagged TypeScript support 🎉
This means Node can run *.ts files. It pairs well with TypeScript's "erasableSyntaxOnly" flag 👍
Many folk contributed 🙏
Yes, this is BARE's first announcement. I later helped improve its specifications and shape its current form.
Yeah, I agree. Many libraries don't comply with `exactOptionalPropertyTypes`. I had to disable it because of this.
Also, I agree that `noUncheckedIndexedAccess` is really picky. I am unsure whether it is possible to easily write code that deals with arrays when this option is enabled.
There are some rationales here in the relevant issue.
> these options are annoying to enable later, but straightforward to comply with in a greenfield situation. It'e best to start with them on and turn them off if you don't like them.
I think exposing types from dependencies to consumers breaks encapsulation. Of course, there are cases where it is acceptable. In general, though, it seems like poor practice to me.
I am a bit surprised by the number of NPM dependencies of tsdown (about 40). Because it builds on Rolldown, I was expecting many fewer. However, this is still an improvement on the unbuild, which had hundreds of dependencies.
I watched three seasons. I might come back one day. The end of the third season was pretty tough.
It seems that the Records & Tuples proposal is withdrawn in favor of Composite Keys.
docs.google.com/presentation...
It could be nice to see if codegen could be easier and quicker with Facet avoiding proc-macro and the like.
Do the use cases always use runtime reflection or do they generate code at compile time using the const data provided by Facet?
It also allows you to write enum flags more easily: you can declare a composite flag that corresponds to the union of several flags.
From an implementation point of view it is not totally true. Engines do many optimizations to use actual arrays. They even have dedicated optimizations for holey arrays (arrays with holes aka sparse arrays).
They are good days because you realize that you still have so much to learn. This is the prerequisite for acquiring new knowledge :)
True. Note thta 21.x is no longer maintained.