No check for valid precision and base in `setprecision(f::Function, ::Type{BigFloat}, prec::Integer; base::Integer=2)`

[Joel-Dahne]

The implementation of setprecision(::Type{BigFloat}, precision::Integer; base::Integer=2) checks that precision and base has valid values. More precisely it does

function setprecision(::Type{BigFloat}, precision::Integer; base::Integer=2)
    base > 1 || throw(DomainError(base, "`base` cannot be less than 2."))
    precision > 0 || throw(DomainError(precision, "`precision` cannot be less than 1."))
    DEFAULT_PRECISION[] = _convert_precision_from_base(precision, base)
    return precision
end

The version taking a function as argument is missing this check, it is currently implemented as

function setprecision(f::Function, ::Type{BigFloat}, prec::Integer; base::Integer=2)
    Base.ScopedValues.@with(CURRENT_PRECISION => _convert_precision_from_base(prec, base), f())
end

As a consequence we have e.g.

julia> setprecision(BigFloat, -1)
ERROR: DomainError with -1:
`precision` cannot be less than 1.
Stacktrace:
 [1] setprecision(::Type{BigFloat}, precision::Int64; base::Int64)
   @ Base.MPFR ./mpfr.jl:1067
 [2] setprecision(::Type{BigFloat}, precision::Int64)
   @ Base.MPFR ./mpfr.jl:1065
 [3] top-level scope
   @ REPL[15]:1

julia> setprecision(BigFloat, -1) do
           "No error!"
       end
"No error!"

The issue was introduced in #51362 when setprecision was changed to use ScopedValues. At the moment it seems like the missing check cannot actually lead to any issues because the constructor for BigFloat explicitly checks that precision >= 2. It does however still seem reasonable to include the check.

julia> versioninfo()
Julia Version 1.12.0-beta3
Commit faca79b503a (2025-05-12 06:47 UTC)
Build Info:
  Official https://julialang.org release
Platform Info:
  OS: Linux (x86_64-linux-gnu)
  CPU: 24 × AMD Ryzen 9 5900X 12-Core Processor
  WORD_SIZE: 64
  LLVM: libLLVM-18.1.7 (ORCJIT, znver3)
  GC: Built with stock GC
Threads: 1 default, 1 interactive, 1 GC (on 24 virtual cores)