Reliable crash in Jupyter notebook on arm64, AS M1

[Morpheu5]

I have some OHLC data here which I'm trying to plot and manipulate in a Jupyter notebook using the IJulia kernel. The following code reliably crashes the kernel, but shows no problems when running in CLI.

using DataFrames
using Plots
gr();

data = CSV.read("ohlc.csv", DataFrame);
plot(data[:, :close])

with crashlog extracted from Jupyter as follows.

10:43:07(write task): EXECUTING plot(data[end-24*60:end, :close])
10:43:07(write task): SENDING IPython Msg [ idents execute_input ] {
  parent_header = Dict{String, Any}("msg_id" => "d4ac6ed086d743558a866fd48908eda8", "msg_type" => "execute_request", "username" => "username", "date" => "2022-02-02T09:43:07.599Z", "version" => "5.2", "session" => "3e188fdcdb4a42e18f1cd0698322c974"),
  header = Dict{String, Any}("msg_id" => "052178aa-0295-40af-ae10-977a5ab837f1", "msg_type" => "execute_input", "username" => "username", "date" => Dates.DateTime("2022-02-02T10:43:07.612"), "version" => "5.3", "session" => "3e188fdcdb4a42e18f1cd0698322c974"),
  metadata = Dict{String, Any}(),
  content = Dict{String, Any}("code" => "plot(data[end-24*60:end, :close])", "execution_count" => 7)
}

signal (11): Segmentation fault: 11
in expression starting at /Users/morpheu5/.julia/packages/IJulia/e8kqU/src/kernel.jl:53
^ at ./math.jl:0 [inlined]
bounding_order_of_magnitude at /Users/morpheu5/.julia/packages/PlotUtils/rTTui/src/ticks.jl:11
optimize_ticks_typed at /Users/morpheu5/.julia/packages/PlotUtils/rTTui/src/ticks.jl:197
unknown function (ip: 0x1156280b3)
jl_apply_generic at /Applications/Julia-1.7.app/Contents/Resources/julia/lib/julia/libjulia-internal.1.7.dylib (unknown line)
#optimize_ticks#42 at /Users/morpheu5/.julia/packages/PlotUtils/rTTui/src/ticks.jl:154
optimize_ticks##kw at /Users/morpheu5/.julia/packages/PlotUtils/rTTui/src/ticks.jl:152
unknown function (ip: 0x115623e07)
jl_apply_generic at /Applications/Julia-1.7.app/Contents/Resources/julia/lib/julia/libjulia-internal.1.7.dylib (unknown line)
optimal_ticks_and_labels at /Users/morpheu5/.julia/packages/Plots/Xv2qA/src/axes.jl:170
unknown function (ip: 0x11562314b)
jl_apply_generic at /Applications/Julia-1.7.app/Contents/Resources/julia/lib/julia/libjulia-internal.1.7.dylib (unknown line)
do_apply at /Applications/Julia-1.7.app/Contents/Resources/julia/lib/julia/libjulia-internal.1.7.dylib (unknown line)
get_ticks at /Users/morpheu5/.julia/packages/Plots/Xv2qA/src/axes.jl:320
jl_apply_generic at /Applications/Julia-1.7.app/Contents/Resources/julia/lib/julia/libjulia-internal.1.7.dylib (unknown line)
#get_ticks#95 at /Users/morpheu5/.julia/packages/Plots/Xv2qA/src/axes.jl:241
get_ticks at /Users/morpheu5/.julia/packages/Plots/Xv2qA/src/axes.jl:224 [inlined]
_update_min_padding! at /Users/morpheu5/.julia/packages/Plots/Xv2qA/src/backends/gr.jl:888
jl_apply_generic at /Applications/Julia-1.7.app/Contents/Resources/julia/lib/julia/libjulia-internal.1.7.dylib (unknown line)
iterate at ./generator.jl:47 [inlined]
_collect at ./array.jl:744
unknown function (ip: 0x115617d4b)
collect_similar at ./array.jl:653 [inlined]
map at ./abstractarray.jl:2849 [inlined]
_update_min_padding! at /Users/morpheu5/.julia/packages/Plots/Xv2qA/src/layouts.jl:290
jl_apply_generic at /Applications/Julia-1.7.app/Contents/Resources/julia/lib/julia/libjulia-internal.1.7.dylib (unknown line)
prepare_output at /Users/morpheu5/.julia/packages/Plots/Xv2qA/src/plot.jl:224
show at /Users/morpheu5/.julia/packages/Plots/Xv2qA/src/output.jl:213
jl_apply_generic at /Applications/Julia-1.7.app/Contents/Resources/julia/lib/julia/libjulia-internal.1.7.dylib (unknown line)
do_apply at /Applications/Julia-1.7.app/Contents/Resources/julia/lib/julia/libjulia-internal.1.7.dylib (unknown line)
#sprint#426 at ./strings/io.jl:114
sprint at ./strings/io.jl:108 [inlined]
_ijulia_display_dict at /Users/morpheu5/.julia/packages/Plots/Xv2qA/src/ijulia.jl:47
display_dict at /Users/morpheu5/.julia/packages/Plots/Xv2qA/src/init.jl:92
jl_apply_generic at /Applications/Julia-1.7.app/Contents/Resources/julia/lib/julia/libjulia-internal.1.7.dylib (unknown line)
jl_f__call_latest at /Applications/Julia-1.7.app/Contents/Resources/julia/lib/julia/libjulia-internal.1.7.dylib (unknown line)
#invokelatest#2 at ./essentials.jl:716 [inlined]
invokelatest at ./essentials.jl:714 [inlined]
execute_request at /Users/morpheu5/.julia/packages/IJulia/e8kqU/src/execute_request.jl:112
jl_apply_generic at /Applications/Julia-1.7.app/Contents/Resources/julia/lib/julia/libjulia-internal.1.7.dylib (unknown line)
jl_f__call_latest at /Applications/Julia-1.7.app/Contents/Resources/julia/lib/julia/libjulia-internal.1.7.dylib (unknown line)
#invokelatest#2 at ./essentials.jl:716 [inlined]
invokelatest at ./essentials.jl:714 [inlined]
eventloop at /Users/morpheu5/.julia/packages/IJulia/e8kqU/src/eventloop.jl:8
#15 at ./task.jl:423
unknown function (ip: 0x115351ed7)
jl_apply_generic at /Applications/Julia-1.7.app/Contents/Resources/julia/lib/julia/libjulia-internal.1.7.dylib (unknown line)
start_task at /Applications/Julia-1.7.app/Contents/Resources/julia/lib/julia/libjulia-internal.1.7.dylib (unknown line)
Allocations: 174538587 (Pool: 174480553; Big: 58034); GC: 107
[I 10:43:13.684 NotebookApp] KernelRestarter: restarting kernel (1/5), keep random ports
kernel 6665f423-b38f-4a16-8dd4-be2f1b514ff9 restarted
PROFILE = Dict{String, Any}("key" => "11d37313-a6793232823424c4cc9f0654", "transport" => "tcp", "signature_scheme" => "hmac-sha256", "shell_port" => 55755, "hb_port" => 55758, "control_port" => 55759, "ip" => "127.0.0.1", "stdin_port" => 55757, "iopub_port" => 55756, "kernel_name" => "ijulia-1.7")
10:43:16(): SENDING IPython Msg [ idents status ] {
  parent_header = Dict("username" => "jlkernel", "session" => "f0233114-d42b-4d38-b664-ad3a9525986b"),
  header = Dict{String, Any}("msg_id" => "a9505189-ece0-446a-a0f6-f8b5fd18dd6e", "msg_type" => "status", "username" => "jlkernel", "date" => Dates.DateTime("2022-02-02T10:43:16.899"), "version" => "5.3", "session" => "f0233114-d42b-4d38-b664-ad3a9525986b"),
  metadata = Dict{String, Any}(),
  content = Dict("execution_state" => "starting")
}
Starting kernel event loops.
10:43:17(write task): got msg part 3e188fdcdb4a42e18f1cd0698322c974
10:43:17(write task): got msg part <IDS|MSG>
10:43:19(write task): RECEIVED IPython Msg [ idents 3e188fdcdb4a42e18f1cd0698322c974 ] {
  parent_header = Dict{String, Any}(),
  header = Dict{String, Any}("msg_id" => "5f1e0c5471b74b659e2b56afe0dc0212", "msg_type" => "kernel_info_request", "username" => "username", "date" => "2022-02-02T09:43:17.162Z", "version" => "5.2", "session" => "3e188fdcdb4a42e18f1cd0698322c974"),
  metadata = Dict{String, Any}(),
  content = Dict{String, Any}()
}
10:43:19(write task): SENDING IPython Msg [ idents status ] {
  parent_header = Dict{String, Any}("msg_id" => "5f1e0c5471b74b659e2b56afe0dc0212", "msg_type" => "kernel_info_request", "username" => "username", "date" => "2022-02-02T09:43:17.162Z", "version" => "5.2", "session" => "3e188fdcdb4a42e18f1cd0698322c974"),
  header = Dict{String, Any}("msg_id" => "731a698b-555b-44bc-8770-125492dc6430", "msg_type" => "status", "username" => "username", "date" => Dates.DateTime("2022-02-02T10:43:19.272"), "version" => "5.3", "session" => "3e188fdcdb4a42e18f1cd0698322c974"),
  metadata = Dict{String, Any}(),
  content = Dict("execution_state" => "busy")
}
10:43:19(write task): SENDING IPython Msg [ idents 3e188fdcdb4a42e18f1cd0698322c974 ] {
  parent_header = Dict{String, Any}("msg_id" => "5f1e0c5471b74b659e2b56afe0dc0212", "msg_type" => "kernel_info_request", "username" => "username", "date" => "2022-02-02T09:43:17.162Z", "version" => "5.2", "session" => "3e188fdcdb4a42e18f1cd0698322c974"),
  header = Dict{String, Any}("msg_id" => "bd156f15-607e-4f55-853d-f309b55d4c8c", "msg_type" => "kernel_info_reply", "username" => "username", "date" => Dates.DateTime("2022-02-02T10:43:19.314"), "version" => "5.3", "session" => "3e188fdcdb4a42e18f1cd0698322c974"),
  metadata = Dict{String, Any}(),
  content = Dict{String, Any}("help_links" => [Dict("text" => "Julia Home Page", "url" => "http://julialang.org/"), Dict("text" => "Julia Documentation", "url" => "http://docs.julialang.org/"), Dict("text" => "Julia Packages", "url" => "http://pkg.julialang.org/")], "language_info" => Dict("file_extension" => ".jl", "name" => "julia", "mimetype" => "application/julia", "version" => "1.7.1"), "status" => "ok", "implementation" => "ijulia", "banner" => "Julia: A fresh approach to technical computing.", "protocol_version" => "5.0")
}
10:43:19(write task): SENDING IPython Msg [ idents status ] {
  parent_header = Dict{String, Any}("msg_id" => "5f1e0c5471b74b659e2b56afe0dc0212", "msg_type" => "kernel_info_request", "username" => "username", "date" => "2022-02-02T09:43:17.162Z", "version" => "5.2", "session" => "3e188fdcdb4a42e18f1cd0698322c974"),
  header = Dict{String, Any}("msg_id" => "e8c2ce06-2703-4b9f-b1ae-52e78d11ee95", "msg_type" => "status", "username" => "username", "date" => Dates.DateTime("2022-02-02T10:43:19.411"), "version" => "5.3", "session" => "3e188fdcdb4a42e18f1cd0698322c974"),
  metadata = Dict{String, Any}(),
  content = Dict("execution_state" => "idle")
}

PlotUtils appears to be where the crash happens, and that's why I'm here, but I'm not too experienced navigating Julia's stack traces so feel free to point me somewhere else :)

This is my configuration:

$ jupyter --version
Selected Jupyter core packages...
IPython          : 8.0.0
ipykernel        : 6.7.0
ipywidgets       : 7.6.5
jupyter_client   : 7.1.0
jupyter_core     : 4.9.1
jupyter_server   : 1.13.2
jupyterlab       : 3.2.8
nbclient         : 0.5.10
nbconvert        : 6.4.0
nbformat         : 5.1.3
notebook         : 6.4.7
qtconsole        : 5.2.2
traitlets        : 5.1.1

$ echo "using Pkg; Pkg.status()" | julia
      Status `~/.julia/environments/v1.7/Project.toml`
  [336ed68f] CSV v0.10.2
  [324d7699] CategoricalArrays v0.10.2
  [35d6a980] ColorSchemes v3.16.0
  [717857b8] DSP v0.7.4
  [a93c6f00] DataFrames v1.3.2
  [1313f7d8] DataFramesMeta v0.10.0
  [4b11ee91] Gaston v1.0.4
  [f67ccb44] HDF5 v0.16.1
  [09f84164] HypothesisTests v0.10.6
  [7073ff75] IJulia v1.23.2
  [682c06a0] JSON v0.21.2
  [50d2b5c4] Lazy v0.15.1
  [76087f3c] NLopt v0.6.4
  [bac558e1] OrderedCollections v1.4.1
  [91a5bcdd] Plots v1.25.5
  [d330b81b] PyPlot v2.10.0
  [2913bbd2] StatsBase v0.33.14
  [2cb19f9e] StatsKit v0.3.1
  [f3b207a7] StatsPlots v0.14.30

I can only add that if I run the x86_64 build of Julia 1.7.1 via Rosetta 2, the crash mostly disappears (I think I only managed to trigger it once but I can't remember the conditions). OTOH, running via Rosetta 2 in incredibly, painstakingly slow when it comes to reading CSVs into DataFrames, so I'd rather avoid it. I mean, this one I read in a few seconds with the native build, but it takes several minutes with the x86_64 build, and this file is small at just over 200 kB. I have a few CSVs that easily reach into the 10s of MB, and at least a couple closer to 0.5 GB so… :)