Starting this thread anew to avoid confusion, because I changed my mind on which approach I support partway through a long thread (sorry!). It’s also useful to distill some of the last thread’s learnings. Thank you to all who participated—especially those who rightfully criticized my ignorance! Man…

Like the last thread, I’m terminating this thread (or at least, talking about function chaining) because I’ve arrived at an even better idea . Please refrain from further comments on this post, unless they’re about partial function application and function composition!

"uniment: Moreover, every time you define a lambda, it’s compiled from scratch (this includes defining a struct to hold the fixed arguments). This takes typically about ten milliseconds. It’s pretty wasteful to have a lot of these in your codebase, for functions that will only ever be used once…

[image] bertschi: Why not my_arr |> x->meth(x,1) |> ⬚->my_func(x, ⬚, y) if you are already chaining anyways? Although the operations on my_arr might be most naturally thought of as a call chain, the call to my_func might not be. I’d rather not be locked into a long chain, if that’s not how…

[image] uniment: Although the operations on my_arr might be most naturally thought of as a call chain, the call to my_func might not be. I’d rather not be locked into a long chain, if that’s not how it was constructed in my head. Each chain sends an object through a sequence of transformations.…

[image] jar1: My initial instinct is that the commas are just annoying and don’t add much value. I’ve decided that I agree with you, and I’ve updated the spec and examples for block expressions accordingly. In block mode the commas are now optional, similar to how Vectors work. I’m also play…

[image] bertschi: @generated function Base.|>(x, f::Base.Fix1) :(f.f(f.x, x)) end Would that be as fast as using a macro in the first place, i.e., option 2? My test says no: julia> @generated function Base.:|>(x, f::Base.Fix1) :(f.f(f.x, x)) end julia> f = (x,y) -> x…

I’m really not sure about these benchmarks … julia> @generated function Base.:|>(x, f::Base.Fix1) :(f.f(f.x, x)) end julia> f = (x,y) -> x+y; f(1,2); julia> @time 1 |> Base.Fix1(f, 2) 0.026216 seconds (4.66 k allocations: 281.242 KiB, 99.03% compilation time) 3 …

That was another reason why I wrote Chain, because you can avoid compilation of Fix instances or anonymous functions or whatever you need to make |> work…

Indeed, measuring compile time is a difficult thing to do. I detailed my misadventures with it in this post . The runtime of this function is basically instantaneous, which you can confirm using @btime (a proper measurement will show single-digit nanoseconds)—so even though compile time and runtime …

[image] uniment: In block mode the commas are now optional Optional commas is unnecessary and confusing imo. Commas or no commas, not both.

I somewhat disagree with this, but you made me realize that the behavior I had specified of allowing but not forcing commas in :blocks causes inconsistent behavior when you wish to have a single expression that returns a tuple. So I have no choice but to oblige :wink: What I want is syntax which al…

Fixing the Piping/Chaining/Partial Application Issue (Rev 2)

Internals & Design

jar1 November 19, 2022, 7:20pm 31

Optional commas is unnecessary and confusing imo. Commas or no commas, not both.

Topic		Replies	Views
Summary of piping/chaining proposal? Internals & Design	23	3658	August 13, 2024
Fixing the Piping/Chaining Issue (Rev 3) Internals & Design multithreading , syntax , piping , chaining , threading	89	9195	April 5, 2024
Fixing the Piping/Chaining Issue Internals & Design proposal , piping , chaining , partial-evaluation , threading	212	9865	January 16, 2023
Seeking feedback on Chainables before registering it Package Announcements package , feedback , chain	16	532	March 14, 2026
[RFC] PipelessPipes.jl (now Chain.jl) Package Announcements	61	5569	March 25, 2021

Fixing the Piping/Chaining/Partial Application Issue (Rev 2)

Related topics