Errors in the REPL tests don't actually cause CI to fail

[vtjnash]

I saw this happening locally, though I don't know what it means

REPL                               (10) |        started at 2021-11-19T19:32:30.340
      From worker 10:   foo
      From worker 10:   foo foo
      From worker 10:   UNHANDLED TASK ERROR: IOError: stream is closed or unusable
      From worker 10:   Stacktrace:
      From worker 10:     [1] check_open
      From worker 10:       @ ./stream.jl:388 [inlined]
      From worker 10:     [2] uv_write_async(s::Base.PipeEndpoint, p::Ptr{UInt8}, n::UInt64)
      From worker 10:       @ Base ./stream.jl:1069
      From worker 10:     [3] uv_write(s::Base.PipeEndpoint, p::Ptr{UInt8}, n::UInt64)
      From worker 10:       @ Base ./stream.jl:1032
      From worker 10:     [4] unsafe_write(s::Base.PipeEndpoint, p::Ptr{UInt8}, n::UInt64)
      From worker 10:       @ Base ./stream.jl:1115
      From worker 10:     [5] unsafe_write
      From worker 10:       @ ./io.jl:675 [inlined]                                                                                                                               From worker 10:     [6] unsafe_write(s::Base.PipeEndpoint, p::Base.RefValue{UInt8}, n::Int64)
      From worker 10:       @ Base ./io.jl:673
      From worker 10:     [7] write
      From worker 10:       @ ./io.jl:676 [inlined]
      From worker 10:     [8] write(s::Base.PipeEndpoint, b::UInt8)
      From worker 10:       @ Base ./stream.jl:1153
      From worker 10:     [9] write(io::Main.Test10Main_REPL.REPLTests.FakeTerminals.FakeTerminal, byte::UInt8) (repeats 2 times)
      From worker 10:       @ Base ./io.jl:417
      From worker 10:    [10] write
      From worker 10:       @ ./io.jl:729 [inlined]
      From worker 10:    [11] print
      From worker 10:       @ ./char.jl:252 [inlined]
      From worker 10:    [12] println(io::IOContext{Main.Test10Main_REPL.REPLTests.FakeTerminals.FakeTerminal})
      From worker 10:       @ Base ./coreio.jl:5
      From worker 10:    [13] print_response(errio::IO, response::Any, show_value::Bool, have_color::Bool, specialdisplay::Union{Nothing, AbstractDisplay})
      From worker 10:       @ REPL /data/vtjnash/julia/usr/share/julia/stdlib/v1.8/REPL/src/REPL.jl:307
      From worker 10:    [14] (::REPL.var"#45#46"{REPL.LineEditREPL, Pair{Any, Bool}, Bool, Bool})(io::Any)
      From worker 10:       @ REPL /data/vtjnash/julia/usr/share/julia/stdlib/v1.8/REPL/src/REPL.jl:276
      From worker 10:    [15] with_repl_linfo(f::Any, repl::REPL.LineEditREPL)
      From worker 10:       @ REPL /data/vtjnash/julia/usr/share/julia/stdlib/v1.8/REPL/src/REPL.jl:527
      From worker 10:    [16] print_response(repl::REPL.AbstractREPL, response::Any, show_value::Bool, have_color::Bool)
      From worker 10:       @ REPL /data/vtjnash/julia/usr/share/julia/stdlib/v1.8/REPL/src/REPL.jl:274
      From worker 10:    [17] (::REPL.var"#do_respond#70"{Bool, Bool, REPL.var"#81#91"{REPL.LineEditREPL, REPL.REPLHistoryProvider}, REPL.LineEditREPL, REPL.LineEdit.Prompt})(s::REPL.LineEdit.MIState, buf::An
y, ok::Bool)
      From worker 10:       @ REPL /data/vtjnash/julia/usr/share/julia/stdlib/v1.8/REPL/src/REPL.jl:863
      From worker 10:    [18] #invokelatest#2
      From worker 10:       @ ./essentials.jl:731 [inlined]
      From worker 10:    [19] invokelatest
      From worker 10:       @ ./essentials.jl:729 [inlined]
      From worker 10:    [20] run_interface(terminal::REPL.Terminals.TextTerminal, m::REPL.LineEdit.ModalInterface, s::REPL.LineEdit.MIState)
      From worker 10:       @ REPL.LineEdit /data/vtjnash/julia/usr/share/julia/stdlib/v1.8/REPL/src/LineEdit.jl:2509
      From worker 10:    [21] run_frontend(repl::REPL.LineEditREPL, backend::REPL.REPLBackendRef)
      From worker 10:       @ REPL /data/vtjnash/julia/usr/share/julia/stdlib/v1.8/REPL/src/REPL.jl:1249
      From worker 10:    [22] (::REPL.var"#49#54"{REPL.LineEditREPL, REPL.REPLBackendRef})()
      From worker 10:       @ REPL ./task.jl:466
      From worker 10:
      From worker 10:   caused by: IOError: stream is closed or unusable
      From worker 10:   Stacktrace:
      From worker 10:     [1] check_open
      From worker 10:       @ ./stream.jl:388 [inlined]
      From worker 10:     [2] uv_write_async(s::Base.PipeEndpoint, p::Ptr{UInt8}, n::UInt64)
      From worker 10:       @ Base ./stream.jl:1069
      From worker 10:     [3] uv_write(s::Base.PipeEndpoint, p::Ptr{UInt8}, n::UInt64)
      From worker 10:       @ Base ./stream.jl:1032
      From worker 10:     [4] unsafe_write(s::Base.PipeEndpoint, p::Ptr{UInt8}, n::UInt64)
      From worker 10:       @ Base ./stream.jl:1115
      From worker 10:     [5] unsafe_write(io::Main.Test10Main_REPL.REPLTests.FakeTerminals.FakeTerminal, p::Ptr{UInt8}, nb::UInt64) (repeats 2 times)
      From worker 10:       @ Base ./io.jl:419                                                                                                                                                                        From worker 10:     [6] write
      From worker 10:       @ ./strings/io.jl:244 [inlined]
      From worker 10:     [7] print
      From worker 10:       @ ./strings/io.jl:246 [inlined]
      From worker 10:     [8] with_output_color(f::Function, color::Symbol, io::IOContext{Main.Test10Main_REPL.REPLTests.FakeTerminals.FakeTerminal}, args::String; bold::Bool, underline::Bool, blink::Bool, reverse::Bool, hidden::Bool)
      From worker 10:       @ Base ./util.jl:79
      From worker 10:     [9] #printstyled#883
      From worker 10:       @ ./util.jl:129 [inlined]
      From worker 10:    [10] display_error(io::IOContext{Main.Test10Main_REPL.REPLTests.FakeTerminals.FakeTerminal}, stack::Base.ExceptionStack)
      From worker 10:       @ Base ./client.jl:102
      From worker 10:    [11] #invokelatest#2
      From worker 10:       @ ./essentials.jl:731 [inlined]
      From worker 10:    [12] invokelatest
      From worker 10:       @ ./essentials.jl:729 [inlined]
      From worker 10:    [13] print_response(errio::IO, response::Any, show_value::Bool, have_color::Bool, specialdisplay::Union{Nothing, AbstractDisplay})
      From worker 10:       @ REPL /data/vtjnash/julia/usr/share/julia/stdlib/v1.8/REPL/src/REPL.jl:289
      From worker 10:    [14] (::REPL.var"#45#46"{REPL.LineEditREPL, Pair{Any, Bool}, Bool, Bool})(io::Any)
      From worker 10:       @ REPL /data/vtjnash/julia/usr/share/julia/stdlib/v1.8/REPL/src/REPL.jl:276
      From worker 10:    [15] with_repl_linfo(f::Any, repl::REPL.LineEditREPL)
      From worker 10:       @ REPL /data/vtjnash/julia/usr/share/julia/stdlib/v1.8/REPL/src/REPL.jl:527
      From worker 10:    [16] print_response(repl::REPL.AbstractREPL, response::Any, show_value::Bool, have_color::Bool)
      From worker 10:       @ REPL /data/vtjnash/julia/usr/share/julia/stdlib/v1.8/REPL/src/REPL.jl:274
      From worker 10:    [17] (::REPL.var"#do_respond#70"{Bool, Bool, REPL.var"#81#91"{REPL.LineEditREPL, REPL.REPLHistoryProvider}, REPL.LineEditREPL, REPL.LineEdit.Prompt})(s::REPL.LineEdit.MIState, buf::Any, ok::Bool)
      From worker 10:       @ REPL /data/vtjnash/julia/usr/share/julia/stdlib/v1.8/REPL/src/REPL.jl:863
      From worker 10:    [18] #invokelatest#2
      From worker 10:       @ ./essentials.jl:731 [inlined]
      From worker 10:    [19] invokelatest
      From worker 10:       @ ./essentials.jl:729 [inlined]
      From worker 10:    [20] run_interface(terminal::REPL.Terminals.TextTerminal, m::REPL.LineEdit.ModalInterface, s::REPL.LineEdit.MIState)
      From worker 10:       @ REPL.LineEdit /data/vtjnash/julia/usr/share/julia/stdlib/v1.8/REPL/src/LineEdit.jl:2509
      From worker 10:    [21] run_frontend(repl::REPL.LineEditREPL, backend::REPL.REPLBackendRef)
      From worker 10:       @ REPL /data/vtjnash/julia/usr/share/julia/stdlib/v1.8/REPL/src/REPL.jl:1249
      From worker 10:    [22] (::REPL.var"#49#54"{REPL.LineEditREPL, REPL.REPLBackendRef})()
      From worker 10:       @ REPL ./task.jl:466
REPL                               (10) |    72.70 |   1.61 |  2.2 |    2663.52 |   684.44

[aviatesk]

Seems like this first appeared at 2c2e724.

[DilumAluthge]

Why doesn't this test failure cause CI to fail? That in itself seems to be a bug.

[BioTurboNick]

FWIW the underlying issue in the OP was fixed as part of #43173

[DilumAluthge]

Now that the underlying bug is fixed, I think we should shift the focus of this issue to figuring out why these failures didn't actually cause CI to fail.

[KristofferC]

Wasn't that also fixed in #43173 (a missing wait)?

[DilumAluthge]

Good catch; I didn't see the addition ofBase.wait(repltask).

I think this issue is resolved then.