Update Getting-Started.md

[BebeSparkelSparkel]

Updated getting started for psc-package instead of bower

Caused me lots of frustration so hope this makes it easier for others.

Related issue #187

[justinwoo]

You can find the packages curated packages on [Pursuit (https://pursuit.purescript.org/)

This seems confusing, that it would make people think that all Pursuit packages are in package sets. Also that package sets aren't actually curated, just community contributed.

[BebeSparkelSparkel]

@jbrownson Please advise what would be a more accurate statement

[hdgarrood]

I’m not sure we should position psc-package as the default option, because many packages aren’t available in package sets, and because it’s a relatively uncommon model (whereas people are more likely to have come across something vaguely resembling bower before).

[BebeSparkelSparkel]

I guess my problem with bower was that I couldn't even install purescript-config and the default prelude for version 0.12.0 without having to manually/randomly pick the package versions

Unable to find a suitable version for purescript-prelude, please choose one by typing one of the numbers below:
    1) purescript-prelude#^3.0.0 which resolved to 3.3.0 and is required by purescript-config#0.0.6, purescript-control#3.3.1, purescript-eff#3.2.1, purescript-either#3.2.0, purescript-functions#3.0.0, purescript-invariant#3.0.0, purescript-leibniz#4.1.0, purescript-maybe#3.1.0, purescript-newtype#2.0.0, purescript-node-streams#3.3.0, purescript-nonempty#4.3.0, purescript-posix-types#3.0.0, purescript-proxy#2.1.0, purescript-tuples#4.1.0
    2) purescript-prelude#^4.0.0 which resolved to 4.1.0 and is required by purescript-argonaut-core#4.0.1, purescript-arrays#5.0.0, purescript-bifunctors#4.0.0, purescript-console#4.1.0, purescript-const#4.0.0, purescript-contravariant#4.0.0, purescript-control#4.1.0, purescript-datetime#4.0.0, purescript-distributive#4.0.0, purescript-effect#2.0.0, purescript-either#4.0.0, purescript-enums#4.0.0, purescript-exceptions#4.0.0, purescript-foldable-traversable#4.0.0, purescript-foreign#5.0.0, purescript-foreign-object#1.0.0, purescript-form-urlencoded#4.0.1, purescript-functions#4.0.0, purescript-functors#3.0.1, purescript-gen#2.1.0, purescript-http-methods#4.0.2, purescript-identity#4.0.0, purescript-integers#4.0.0, purescript-invariant#4.1.0, purescript-lazy#4.0.0, purescript-lists#5.0.0, purescript-maybe#4.0.0, purescript-media-types#4.0.1, purescript-newtype#3.0.0, purescript-nonempty#5.0.0, purescript-ordered-collections#1.0.0, purescript-orders#4.0.0, purescript-parallel#4.0.0, purescript-psci-support#4.0.0, purescript-refs#4.1.0, purescript-st#4.0.0, purescript-strings#4.0.0, purescript-tailrec#4.0.0, purescript-transformers#4.1.0, purescript-tuples#5.0.0, purescript-unfoldable#4.0.0
    3) purescript-prelude#^4.1.0 which resolved to 4.1.0 and is required by rm2qb

Prefix the choice with ! to persist it to bower.json

? Answer 3

[hdgarrood]

I accept that the UI could be a lot better in that case, but that is it working as expected; there does not appear to be a version of purescript-config which can work alongside version 4.1.0 of the prelude. Therefore switching to psc-package wouldn't help you there: purescript-config most likely wouldn't even make it into any 0.12.x package set at all in its current state.

[BebeSparkelSparkel]

So there is no way to access env vars right now without downgrading to 0.11.7

[hdgarrood]

That’s not true, see https://pursuit.purescript.org/packages/purescript-node-process/6.0.0/docs/Node.Process#v:lookupEnv

[BebeSparkelSparkel]

Thanks

Is https://purescript-users.ml/t/working-with-0-11-7-in-a-0-12-world/180 still valid or is it out of date at this point?

[hdgarrood]

A lot more of the ecosystem has caught up by this point in comparison to when that was written. I think I’d probably recommend 0.12 unless you have a specific reason to use 0.11.7.

[hdgarrood]

I think switching the getting started guide to use psc-package rather than bower is potentially worth considering but needs more discussion first -- for instance, I'd argue we'd also need to a) explain the differences between the psc-package model vs the more standard centralised package repo + dependency solving approach which bower (and most other package managers take) and b) decide on and commit to a standard way of shipping the psc-package binary. Suggest closing this (and potentially also opening a new issue for this discussion)?

[JordanMartinez]

I'd argue we'd also need to a) explain the differences between the psc-package model vs the more standard centralised package repo + dependency solving approach which bower (and most other package managers take)

@hdgarrood I've already done most of this (aside from the dependency solving approach Bower uses) in the related page in my learning repo

[hdgarrood]

@JordanMartinez That page looks good but on reflection I'm not sure it's appropriate to expose all that to someone who is just trying to get something working during their first experience with the language. I'm going to close this PR for now, but feel free to open issues if you want to revive this discussion.