Basic repl interaction not fully precompiled

[LilithHafner]

If I type this at the repl: 1+1[up][up][enter]1[up][down][left][right][down][home][backspace][del]one([tab], I get a lot of precomilations

This test

export JULIA_DEPOT_PATH="~/this_file_does_not_exist"
julia --trace-compile=precompiles -q --startup-file=no
[ type the following: "1+1[up][up][enter]1[up][down][left][right][down][home][backspace][del]one([tab]"  ]
cat precompiles

Yields

precompile(Tuple{typeof(Base.repeat), Char, Int64})
precompile(Tuple{typeof(Base.seek), Base.GenericIOBuffer{GenericMemory{:not_atomic, UInt8, Core.AddrSpace{Core}(0x00)}}, Int64})
precompile(Tuple{typeof(Base.setindex!), Array{String, 1}, String, Int64})
precompile(Tuple{REPL.LineEdit.var"#27#28"{REPL.LineEdit.var"#156#212", String}, Any, Any})
precompile(Tuple{REPL.LineEdit.var"#156#212", REPL.LineEdit.MIState, Any, Vararg{Any}})
precompile(Tuple{REPL.LineEdit.var"#27#28"{REPL.LineEdit.var"#155#211", String}, Any, Any})
precompile(Tuple{REPL.LineEdit.var"#155#211", REPL.LineEdit.MIState, Any, Vararg{Any}})
precompile(Tuple{REPL.LineEdit.var"#27#28"{REPL.LineEdit.var"#265#269"{REPL.LineEdit.PrefixHistoryPrompt}, String}, Any, Any})
precompile(Tuple{typeof(Base.isequal), UInt8})
precompile(Tuple{REPL.LineEdit.var"#265#269"{REPL.LineEdit.PrefixHistoryPrompt}, REPL.LineEdit.MIState, Any, Vararg{Any}})
precompile(Tuple{REPL.LineEdit.var"#27#28"{REPL.LineEdit.var"#148#204", String}, Any, Any})
precompile(Tuple{REPL.LineEdit.var"#148#204", REPL.LineEdit.MIState, Any, Vararg{Any}})
precompile(Tuple{typeof(Base.min), Float64, Float64})
precompile(Tuple{REPL.LineEdit.var"#27#28"{REPL.LineEdit.var"#161#217", String}, Any, Any})
precompile(Tuple{REPL.LineEdit.var"#161#217", REPL.LineEdit.MIState, Any, Vararg{Any}})
precompile(Tuple{typeof(Base.show), Base.GenericIOBuffer{GenericMemory{:not_atomic, UInt8, Core.AddrSpace{Core}(0x00)}}, Method})
precompile(Tuple{typeof(Base.show), Base.IOContext{Base.GenericIOBuffer{GenericMemory{:not_atomic, UInt8, Core.AddrSpace{Core}(0x00)}}}, Core.TypeofBottom})
precompile(Tuple{typeof(Base.show_signature_function), Base.IOContext{Base.GenericIOBuffer{GenericMemory{:not_atomic, UInt8, Core.AddrSpace{Core}(0x00)}}}, Any, Bool, String, Bool})
precompile(Tuple{Type{Pair{A, B} where B where A}, Symbol, TypeVar})
precompile(Tuple{typeof(Base.string_with_env), Base.ImmutableDict{Symbol, Any}, TypeVar})
precompile(Tuple{typeof(REPL.Terminals.cmove_down), REPL.Terminals.TTYTerminal, Int64})
precompile(Tuple{typeof(REPL.Terminals.cmove_col), REPL.Terminals.TTYTerminal, Int64})
precompile(Tuple{typeof(Base.write), Base.TTY, Char})
precompile(Tuple{REPL.LineEdit.var"#27#28"{REPL.LineEdit.var"#248#256", String}, Any, Any})
precompile(Tuple{REPL.LineEdit.var"#248#256", REPL.LineEdit.MIState, REPL.LineEdit.ModeState, Any})
precompile(Tuple{typeof(Base.print), Base.TTY, String})

On Julia 1.11

julia> versioninfo()
Julia Version 1.11.0-beta1
Commit 08e1fc0abb9 (2024-04-10 08:40 UTC)
Build Info:
  Official https://julialang.org/ release
Platform Info:
  OS: Linux (aarch64-linux-gnu)
  CPU: 8 × unknown
  WORD_SIZE: 64
  LLVM: libLLVM-16.0.6 (ORCJIT, apple-m2)
Threads: 1 default, 0 interactive, 1 GC (on 8 virtual cores)
Environment:
  JULIA_DEPOT_PATH = ~/this_file_does_not_exist

See also #52783, #51532

I find these precompilations annoying when I'm trying to determine whether my packages are fully precompiled at the REPL and have to parse out what is my fault and what is just the REPL.

[IanButterworth]

This might just take entering your cheat code into this script

julia/stdlib/REPL/src/precompile.jl

Lines 40 to 62 in 7ba1b33

	repl_script = """
	2+2
	print("")
	printstyled("a", "b")
	display([1])
	display([1 2; 3 4])
	foo(x) = 1
	@time @eval foo(1)
	; pwd
	$CTRL_C
	$CTRL_R$CTRL_C#
	? reinterpret
	using Ra\t$CTRL_C
	\\alpha\t$CTRL_C
	\e[200~paste here ;)\e[201~"$CTRL_C
	$UP_ARROW$DOWN_ARROW$CTRL_C
	123\b\b\b$CTRL_C
	\b\b$CTRL_C
	f(x) = x03
	f(1,2)
	[][1]
	cd("complete_path\t\t$CTRL_C
	"""

[LilithHafner]

Thank you! This is why I opened an issue instead of a PR. PR forthcoming.

[LilithHafner]

Unfortunately, that won't be enough. For example, typing 2+2 at the REPL currently triggers compilation even though it is already in the precompile script.

x@fedora:~/.julia/dev/SplitREPL$ julia +nightly --startup-file=no --trace-compile=stderr
               _
   _       _ _(_)_     |  Documentation: https://docs.julialang.org
  (_)     | (_) (_)    |
   _ _   _| |_  __ _   |  Type "?" for help, "]?" for Pkg help.
  | | | | | | |/ _' |  |
  | | |_| | | | (_| |  |  Version 1.12.0-DEV.344 (2024-04-15)
 _/ |\__'_|_|_|\__'_|  |  Commit 7ba1b332f87 (0 days old master)
|__/                   |

julia> 2+precompile(Tuple{typeof(Base.repeat), Char, Int64})
2

[IanButterworth]

#51106 would be helpful for diagnosing that, but it needs a rebase (cc. @JeffBezanson)

[KristofferC]

The time these take are all of the order of where they take less time than is perceptible by a human and since this is for interactive work they don't really make a difference. If something is found that takes a long time it can be added to the script.

[LilithHafner]

As I mentioned in the OP, the runtime isn't what bothers me; it's the noise when I'm trying to test the precompilation of my own packages. An alternative solution would be to suppress or provide an option to suppress these precomiplation statements from being emitted.

The fact that 2+2 is in the precompile script and yet typing 2+2 in the REPL is not fully precompiled is bad. As @KristofferC correctly notes, it is not a catastrophic bug, but it should still be fixed.