Move back isascii() from Unicode to Base (#25076)

This function is not really Unicode-related and its definition does not depend
on the Unicode standard version. This fixes a bug introduced when moving
isascii() to Base.Unicode, as it was not exported from Unicode stdlib module
as it should have been.

Author: nalimilan

NEWS.md

@@ -766,7 +766,7 @@ Deprecated or removed
 
   * Unicode-related string functions have been moved to the new `Unicode` standard
     library module ([#25021]). This applies to `normalize_string`, `graphemes`,
-    `is_assigned_char`, `textwidth`, `isascii`, `islower`, `isupper`, `isalpha`,
+    `is_assigned_char`, `textwidth`, `islower`, `isupper`, `isalpha`,
     `isdigit`, `isxdigit`, `isnumber`, `isalnum`, `iscntrl`, `ispunct`, `isspace`,
     `isprint`, `isgraph`, `lowercase`, `uppercase`, `titlecase`, `lcfirst` and `ucfirst`.
 

base/char.jl

@@ -147,7 +147,7 @@ function show(io::IO, ::MIME"text/plain", c::Char)
         isoverlong(c) && print(io, "[overlong] ")
         u = UInt32(c)
         h = hex(u, u ≤ 0xffff ? 4 : 6)
-        print(io, (Unicode.isascii(c) ? "ASCII/" : ""), "Unicode U+", h)
+        print(io, (isascii(c) ? "ASCII/" : ""), "Unicode U+", h)
     else
         print(io, ": Malformed UTF-8")
     end

base/exports.jl

@@ -726,6 +726,7 @@ export
     hex2bytes,
     hex2bytes!,
     info,
+    isascii,
     ismatch,
     isvalid,
     join,

base/io.jl

@@ -967,7 +967,7 @@ julia> countlines(io, '.')
 ```
 """
 function countlines(io::IO, eol::Char='\n')
-    Unicode.isascii(eol) || throw(ArgumentError("only ASCII line terminators are supported"))
+    isascii(eol) || throw(ArgumentError("only ASCII line terminators are supported"))
     aeol = UInt8(eol)
     a = Vector{UInt8}(uninitialized, 8192)
     nl = 0

base/loading.jl

@@ -68,7 +68,7 @@ elseif Sys.isapple()
 
         # If there is no match, it's possible that the file does exist but HFS+
         # performed unicode normalization. See  https://developer.apple.com/library/mac/qa/qa1235/_index.html.
-        Unicode.isascii(path_basename) && return false
+        isascii(path_basename) && return false
         Vector{UInt8}(Unicode.normalize(path_basename, :NFD)) == casepreserved_basename
     end
 else

base/repl/latex_symbols.jl

@@ -56,7 +56,7 @@ open(fname) do f
                 split(replace(L, r"[{}\"]+", "\t"), "\t"))
         c = Char(parse(Int, x[2], 16))
         if (Base.is_id_char(c) || Base.isoperator(Symbol(c))) &&
-           string(c) ∉ latex_strings && !Unicode.isascii(c)
+           string(c) ∉ latex_strings && !isascii(c)
             tabcomname = escape_string(x[3])
             if startswith(tabcomname, "\\\\math")
                 tabcomname = string("\\\\", tabcomname[7:end])

base/socket.jl

@@ -636,7 +636,7 @@ Gets all of the IP addresses of the `host`.
 Uses the operating system's underlying getaddrinfo implementation, which may do a DNS lookup.
 """
 function getalladdrinfo(host::String)
-    Unicode.isascii(host) || error("non-ASCII hostname: $host")
+    isascii(host) || error("non-ASCII hostname: $host")
     req = Libc.malloc(_sizeof_uv_getaddrinfo)
     uv_req_set_data(req, C_NULL) # in case we get interrupted before arriving at the wait call
     status = ccall(:jl_getaddrinfo, Int32, (Ptr{Void}, Ptr{Void}, Cstring, Ptr{Void}, Ptr{Void}),

base/strings/basic.jl

@@ -486,6 +486,30 @@ next(e::EachStringIndex, state) = (state, nextind(e.s, state))
 done(e::EachStringIndex, state) = done(e.s, state)
 eltype(::Type{EachStringIndex}) = Int
 
+"""
+    isascii(c::Union{Char,AbstractString}) -> Bool
+
+Test whether a character belongs to the ASCII character set, or whether this is true for
+all elements of a string.
+
+# Examples
+```jldoctest
+julia> isascii('a')
+true
+
+julia> isascii('α')
+false
+
+julia> isascii("abc")
+true
+
+julia> isascii("αβγ")
+false
+```
+"""
+isascii(c::Char) = bswap(reinterpret(UInt32, c)) < 0x80
+isascii(s::AbstractString) = all(isascii, s)
+
 ## string map, filter, has ##
 
 function map(f, s::AbstractString)

base/strings/search.jl

@@ -28,7 +28,7 @@ function search(a::Union{String,ByteArray}, b::Union{Int8,UInt8}, i::Integer = 1
 end
 
 function search(a::ByteArray, b::Char, i::Integer = 1)
-    if Unicode.isascii(b)
+    if isascii(b)
         search(a,UInt8(b),i)
     else
         search(a,Vector{UInt8}(string(b)),i).start
@@ -59,7 +59,7 @@ function rsearch(a::Union{String,ByteArray}, b::Union{Int8,UInt8}, i::Integer =
 end
 
 function rsearch(a::ByteArray, b::Char, i::Integer = length(a))
-    if Unicode.isascii(b)
+    if isascii(b)
         rsearch(a,UInt8(b),i)
     else
         rsearch(a,Vector{UInt8}(string(b)),i).start

base/strings/unicode.jl

@@ -597,32 +597,6 @@ true
 """
 isgraph(c::Char) = UTF8PROC_CATEGORY_LU <= category_code(c) <= UTF8PROC_CATEGORY_SO
 
-"""
-    isascii(c::Union{Char,AbstractString}) -> Bool
-
-Test whether a character belongs to the ASCII character set, or whether this is true for
-all elements of a string.
-
-# Examples
-```jldoctest
-julia> using Unicode
-
-julia> isascii('a')
-true
-
-julia> isascii('α')
-false
-
-julia> isascii("abc")
-true
-
-julia> isascii("αβγ")
-false
-```
-"""
-isascii(c::Char) = bswap(reinterpret(UInt32, c)) < 0x80
-isascii(s::AbstractString) = all(isascii, s)
-
 """
     isxdigit(c::Char) -> Bool
 

stdlib/Distributed/src/Distributed.jl

@@ -17,7 +17,7 @@ using Base: Process, Semaphore, JLOptions, AnyDict, buffer_writes, wait_connecte
             binding_module, notify_error, atexit, julia_exename, julia_cmd,
             AsyncGenerator, acquire, release, invokelatest,
             shell_escape_posixly, uv_error, coalesce, notnothing
-using Base.Unicode: isascii, isdigit, isnumeric
+using Base.Unicode: isdigit, isnumeric
 
 # NOTE: clusterserialize.jl imports additional symbols from Base.Serializer for use
 

stdlib/SuiteSparse/src/cholmod.jl

@@ -611,7 +611,7 @@ end
 
 ### cholmod_check.h ###
 function print_sparse(A::Sparse{Tv}, name::String) where Tv<:VTypes
-    Unicode.isascii(name) || error("non-ASCII name: $name")
+    isascii(name) || error("non-ASCII name: $name")
     set_print_level(common_struct, 3)
     @isok ccall((@cholmod_name("print_sparse", SuiteSparse_long),:libcholmod), Cint,
             (Ptr{C_Sparse{Tv}}, Ptr{UInt8}, Ptr{UInt8}),

test/arrayops.jl

@@ -464,7 +464,7 @@ end
     s = "julia"
     @test find(c -> c == 'l', s) == [3]
     g = Base.Unicode.graphemes("日本語")
-    @test find(Base.Unicode.isascii, g) == Int[]
+    @test find(isascii, g) == Int[]
     @test find(!iszero, (i % 2 for i in 1:10)) == collect(1:2:9)
 end
 @testset "findn" begin

test/strings/io.jl

@@ -66,7 +66,7 @@
         cp, ch, st = cx[i,:]
         @test cp == convert(UInt32, ch)
         @test string(ch) == unescape_string(st)
-        if Base.Unicode.isascii(ch) || !Base.Unicode.isprint(ch)
+        if isascii(ch) || !Base.Unicode.isprint(ch)
             @test st == escape_string(string(ch))
         end
         for j = 1:size(cx,1)