News for hackers. Stuff that matters.
Re:Show HN: Lux – a luxurious package manager for Lua by tengbretson on Monday April 07, 2025 @04:07PM (#43615402)
Is there a good reason to introduce the halting problem to a package manifest?
Re:Show HN: Lux – a luxurious package manager for Lua by hgs3 on Tuesday April 08, 2025 @12:38PM (#43623699)
Lua has a sandboxed runtime so couldn't you kill rogue scripts after a timeout?
Re:Show HN: Lux – a luxurious package manager for Lua by giraffe_lady on Tuesday April 08, 2025 @01:02PM (#43623963)
Would you rather do that or use toml?
Re:Show HN: Lux – a luxurious package manager for Lua by hgs3 on Tuesday April 08, 2025 @01:39PM (#43624392)
Lua was designed for configuration, so I'd rather use Lua. Plus, it makes for a more elegant, unified design. TOML, and friends, are better paired with languages that were not designed for configuration.
Re:Show HN: Lux – a luxurious package manager for Lua by giraffe_lady on Tuesday April 08, 2025 @03:49PM (#43625763)
I know what lua was designed for and have professional experience with it. It's not that great as a config language imo, I can easily understand why a project would use something else.
Re:Show HN: Lux – a luxurious package manager for Lua by pmarreck on Sunday April 13, 2025 @12:49AM (#43670173)
It seems like every new language community and generation of developers needs to re-learn a set of the same lessons (why declarative instead of imperative config is the way to go, why functional is superior to imperative, why immutable is superior to mutable, why TDD (or just unit testing) is important, why hexagonal design is the best organization of logic, why static typing is important, why Nix is the only valid global package manager, etc. etc. etc.) and I suspect it's because these problems all share the same attributes:

1) the inferior solution seems easier at first (but won't be in the long run)

2) the people blithely picking the inferior solution have not yet encountered enough of the thorny nastiness that the initially-harder option completely prevents

the end result being that all the "greybeards" have settled on the "initially harder" solutions for things because they all learned the hard way that it just causes less headaches in the long run