I just created a
sanctuary tag on Stack Overflow. You may like to subscribe to it. Also, I posted a question: Custom equality semantics for Immutable.js data structures. I'd love to be able to use maps and sets when working with Sanctuary, but for various reasons neither the native types nor the Immutable.js equivalents are currently suitable.
:point_up: September 17, 2016 1:41 PM @scott-christopher are there (more specific) common use cases you could name, to explain it a bit more?
here is test case from Free
for a bit context the Free was not stack safe and for it to be stack safe target monds should have some common interface so that it could depend on it. now we have ChainRec, in purescript it's MonadRec, and Free and other structures could use it for doing safe monadic recursion. if you are interested in that here is a paper about that https://github.com/functorial/stack-safety-for-free/blob/gh-pages/index.markdown
Hi everyone. Could anybody share a link to open source project which widely uses sanctuary ? I'm interested to see how exactly communication with outside world (like DB, external services) is implemented also interested in the way how error handling is done. Would be awesome if such a project is already used in real production with nodejs for instance. Thanks in advance
The
generate script which builds the Sanctuary website is one example of handling errors with the Either type, @btbvoy. For DB interactions one could use a library such as Fluture. The only large project I'm aware of that makes heavy use of Sanctuary is not open source, but reading this blog post will give you an idea of how I used Ramda and Sanctuary in that project.
Yes it is, @Bradcomp. I was working on updating #216 last night. I got stuck on the laws after changing
sequence to traverse. I think I'll .skip those tests and ask @scott-christopher for assistance. One question is whether Maybe#fantasy-land/map should replace Maybe#map or whether we should have both. What's your opinion on this, Brad?
but if someone is creating that functor object in it's codebase and there is no way for it to receive other one it could use the normal map i think
And also namespacing those methods I think is good because for some object we might want to use word
map as name of it's property where we are storing some Map object and it's not a function. but as FL methods are namespaced we could have the Map object live in ['map'] and Functor map in ['fantasy-land/map']. so libs which require FL Functor will not care about our Map object. or for example if we want to create more user friendly object maybe we want to call transform to map and andThen to chain. so basically i see FL namespaced methods as the way for libs to interact with each other. but for user who is using some specific implementation of maybe I think there should also be normal non namespaced methods
The main source of slowness is polymorphism. Functions with types such as
Integer -> String -> Maybe Integer are fine; it's functions with types such as a -> a which are problematic. This means S.I is one of the slow functions when type checking is enabled!
ok looking at the uglified code maybe it's not necessary
@davidchambers Could you please elaborate also about sanctuary-js's Either and ramda-fantasy's Either. What's the rule one should follow in order to understand which one to pick :-) As I understand sanctuary-js was created first but ramda-fantasy just recently appeared. Are these libraries competitors ? :-)
RF started as a laboratory for experimentation, and in many respects this is still the case. Sanctuary started as a collection of safe versions of unsafe Ramda functions. The libraries do overlap, but not significantly. Sanctuary overlaps more with Ramda itself.
If you want to use Either or Maybe, Sanctuary should serve you well. If you want Future/Task, I suggest Fluture. If you want IO, Identity, Reader, State, or Tuple, RF should be able to provide it for you.
I have written my point on FL namespaced methods vs non namespaced at origamitower/folktale#39 what you think on that? /cc @davidchambers
Nice one. I certainly think Folktale should keep the unprefixed method names, since Folktale is closer to Scala than Haskell so using it in an OO manner feels quite natural. I'm not convinced this is right for Sanctuary, though. I'll create a Sanctuary issue a bit later (I'm currently reworking the property tests for FL methods and finally making good progress).
@davidchambers I almost never use method chaining, preferring instead to use the functions in Ramda. My only concerns are in regards to compatibility. It would be nice to have adoption of the FL namespacing in Ramda and Sanctuary (and others of course) be fairly simultaneous, or else have the old names deprecated but maintained until Ramda gets updated.