`asyncmap` hides the "caused by" exception on re-throw

[arnaudh]

Version info:

julia> versioninfo()
Julia Version 1.8.2
Commit 36034abf260 (2022-09-29 15:21 UTC)
Platform Info:
  OS: macOS (x86_64-apple-darwin21.4.0)
  CPU: 8 × Intel(R) Core(TM) i7-1068NG7 CPU @ 2.30GHz
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-13.0.1 (ORCJIT, icelake-client)
  Threads: 1 on 8 virtual cores

Installed Julia from https://julialang.org/downloads/, drag-n-drop to Applications folder.

MWE:

julia> f(x) = try throw("root cause"); catch e; throw("additional info $x"); end
f (generic function with 2 methods)

julia> f(1)
ERROR: "additional info 1"
Stacktrace:
 [1] f(x::Int64)
   @ Main ./REPL[19]:1
 [2] top-level scope
   @ REPL[20]:1

caused by: "root cause"
Stacktrace:
 [1] f(x::Int64)
   @ Main ./REPL[19]:1
 [2] top-level scope
   @ REPL[20]:1

julia> asyncmap(f, 1:3)
ERROR: "additional info 1"
Stacktrace:
 [1] f(x::Int64)
   @ Main ./REPL[19]:1
 [2] (::Base.var"#929#934"{typeof(f)})(r::Base.RefValue{Any}, args::Tuple{Int64})
   @ Base ./asyncmap.jl:100
 [3] macro expansion
   @ ./asyncmap.jl:234 [inlined]
 [4] (::Base.var"#945#946"{Base.var"#929#934"{typeof(f)}, Channel{Any}, Nothing})()
   @ Base ./task.jl:484
Stacktrace:
 [1] (::Base.var"#939#941")(x::Task)
   @ Base ./asyncmap.jl:177
 [2] foreach(f::Base.var"#939#941", itr::Vector{Any})
   @ Base ./abstractarray.jl:2774
 [3] maptwice(wrapped_f::Function, chnl::Channel{Any}, worker_tasks::Vector{Any}, c::UnitRange{Int64})
   @ Base ./asyncmap.jl:177
 [4] wrap_n_exec_twice
   @ ./asyncmap.jl:153 [inlined]
 [5] #async_usemap#924
   @ ./asyncmap.jl:103 [inlined]
 [6] #asyncmap#923
   @ ./asyncmap.jl:81 [inlined]
 [7] asyncmap(f::Function, c::UnitRange{Int64})
   @ Base ./asyncmap.jl:80
 [8] top-level scope
   @ REPL[21]:1

It seems asyncmap doesn't print the caused by: "root cause" that would otherwise be shown.

Related work:

• Throw CapturedExceptions from asyncmap to avoid dropping the stacktrace #42105: seems that before that MR not even the "additional info 1" would be printed

[ericphanson]

I think that's what #42105 (comment) was pointing at too; we should be using the whole exception stack, instead of just the exception being thrown.