442
POV: You maintain a JS library that 80% of modern web infrastructure uses as a dependency. (lemmy.blahaj.zone)
submitted 8 months ago by to c/programmer_humor@programming.dev
49 comments fedilink hide all child comments
🤔 1
 
you are viewing a single comment's thread
view the rest of the comments
[–] 36 points 8 months ago (7 children)

i mean its still good to use an abstraction layer in case you ever have to change the underlying call; it's far easier to change it in one place instead of replacing every call

  • permalink
  • fedilink
  • source
  • hideshow 7 child comments
    •
  • [–] 18 points 8 months ago (4 children)

    Is this a joke?

    If you need a different random function, you write a different random function either way. Having one function do nothing but call another function does nothing.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 4 child comments
    •
  • [–] 31 points 8 months ago (2 children)

    There are several legit reasons why you'd do this. Unit tests, for example: override getRandom() with an implementation that always returns the same series of numbers, and now you have repeatable tests without touching the production code.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •
  • [–] 3 points 8 months ago* (1 child)

    Can you override Math.random within a local scope?

    At my shop we do create generic covers for vendor specific functionality, for the reasons you stated. Though the practice was started in case we ever needed to swap vendors.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 1 child comment
    •
  • [–] 12 points 8 months ago

    You can, but you shouldn’t. You don’t know what else relies on Math.random. That’s why there’s the wrapper function. That you can override in unit tests without worrying about other, unrelated code.

  • permalink
  • fedilink
  • source
  • parent
  • [–] 8 points 8 months ago

    It's not about a different function providing different randomness, but providing a compatible implementation for environments not supporting the "regular" implementation.

    If this screenshot is legit, I guarantee you that either the library is older and there was some weird branching for IE or it's brand new and had branching for the hot new JS runtime / cross compiling.

    Supporting a metric fuckton of browsers and environments takes the same amount of shims.

  • permalink
  • fedilink
  • source
  • parent