compatibility with StringView in Base

Project.toml

@@ -1,7 +1,7 @@
 name = "StringViews"
 uuid = "354b36f9-a18e-4713-926e-db85100087ba"
 authors = ["Steven G. Johnson <stevenj@alum.mit.edu>"]
-version = "1.3.7"
+version = "2.0"
 
 [compat]
 julia = "1.6"

src/StringViews.jl

@@ -8,8 +8,20 @@ Unlike Julia's built-in `String` type (which also wraps UTF-8 data), the
 instance, and does not take "ownership" of or modify the array.   Otherwise,
 a `StringView` is intended to be usable in any context where you might
 have otherwise used `String`.
+
+In Julia 1.14, the `StringView` type will be included in the Julia `Base` module,
+at which point the `StringViews` module will become a trivial stub.
 """
 module StringViews
+
+# no longer needed after https://github.com/JuliaLang/julia/pull/60526
+if isdefined(Base, :StringView)
+
+const SVRegexMatch = RegexMatch
+export SVRegexMatch
+
+else
+
 export StringView, SVRegexMatch
 
 """
@@ -27,27 +39,31 @@ in the buffer.
 """
 struct StringView{T<:AbstractVector{UInt8}} <: AbstractString
     data::T
+
+    function StringView{T}(data::T) where {T <: AbstractVector{UInt8}}
+        # For now, StringViews code assumes one-based indexing
+        Base.require_one_based_indexing(data)
+
+        # Prevent someone constructing e.g. a `StringView{AbstractVector{UInt8}}`,
+        # the existence of which will complicate the implementation and provide
+        # no usability benefit.
+        if !isconcretetype(T)
+            throw(ArgumentError("StringView must be parameterized with a concrete type"))
+        end
+
+        new{T}(data)
+    end
 end
 
 const DenseStringView = StringView{<:Union{DenseVector{UInt8},<:Base.FastContiguousSubArray{UInt8,1,<:DenseVector{UInt8}}}}
 const StringAndSub = Union{String,SubString{String}}
 const StringViewAndSub = Union{StringView,SubString{<:StringView}}
 const DenseStringViewAndSub = Union{DenseStringView,SubString{<:DenseStringView}}
 
-Base.Vector{UInt8}(s::StringView{Vector{UInt8}}) = s.data
+StringView(v::AbstractVector{UInt8}) = StringView{typeof(v)}(v)
 Base.Vector{UInt8}(s::StringViewAndSub) = Vector{UInt8}(codeunits(s))
 Base.Array{UInt8}(s::StringViewAndSub) = Vector{UInt8}(s)
 Base.String(s::StringViewAndSub) = String(copyto!(Base.StringVector(ncodeunits(s)), codeunits(s)))
-StringView(s::StringView) = s
-StringView{S}(s::StringView{S}) where {S<:AbstractVector{UInt8}} = s
-StringView(s::String) = StringView(codeunits(s))
-
-# iobuffer constructor (note that buf.data is always 1-based)
-@inline function StringView(buf::IOBuffer, r::OrdinalRange{<:Integer,<:Integer}=Base.OneTo(buf.ptr-1))
-    @boundscheck issubset(r, Base.OneTo(buf.size)) || throw(BoundsError(buf, r))
-    StringView(@view buf.data[r])
-end
-
 Base.copy(s::StringView) = StringView(copy(s.data))
 
 Base.Symbol(s::DenseStringViewAndSub) =
@@ -88,9 +104,10 @@ end
 Base.:(==)(s1::StringViewAndSub, s2::StringAndSub) = s2 == s1
 
 Base.typemin(::Type{StringView{Vector{UInt8}}}) = StringView(Vector{UInt8}(undef,0))
-Base.typemin(::Type{StringView{Base.CodeUnits{UInt8, String}}}) = StringView("")
+Base.typemin(::Type{StringView{Base.CodeUnits{UInt8, String}}}) = StringView(codeunits(""))
 Base.typemin(::T) where {T<:StringView} = typemin(T)
 Base.one(::Union{T,Type{T}}) where {T<:StringView} = typemin(T)
+Base.oneunit(::Union{T, Type{T}}) where {T<:StringView} = typemin(T)
 
 if VERSION < v"1.10.0-DEV.1007" # JuliaLang/julia#47880
     Base.isvalid(s::DenseStringViewAndSub) = ccall(:u8_isvalid, Int32, (Ptr{UInt8}, Int), s, sizeof(s)) ≠ 0
@@ -141,4 +158,6 @@ include("parse.jl")
 include("util.jl")
 include("search.jl")
 
+end
+
 end # module

src/decoding.jl

@@ -68,15 +68,15 @@ end
 Base.getindex(s::StringView, r::UnitRange{<:Integer}) = s[Int(first(r)):Int(last(r))]
 
 @inline function Base.getindex(s::StringView, r::UnitRange{Int})
-    isempty(r) && return ""
+    isempty(r) && return StringView(s.data[1:0])
     i, j = first(r), last(r)
     @boundscheck begin
         checkbounds(s, r)
         @inbounds isvalid(s, i) || Base.string_index_err(s, i)
         @inbounds isvalid(s, j) || Base.string_index_err(s, j)
     end
     j = nextind(s, j) - 1
-    return StringView(@view s.data[i:j])
+    return StringView(s.data[i:j])
 end
 
 Base.length(s::StringView) = length_continued(s, 1, ncodeunits(s), ncodeunits(s))

test/runtests.jl

@@ -6,11 +6,11 @@ s = StringView(b)
 ss = SubString(s, 2, 5) # "ooba"
 abc = StringView(0x61:0x63)
 invalid = StringView([0x8b, 0x52, 0x9b, 0x8d])
-su = StringView("föôẞαr")
+stringview(s::String) = StringView(codeunits(s)) # convenience constructor
+su = stringview("föôẞαr")
 
 @testset "construction/conversion" begin
-    @test StringView(s) === s
-    @test Vector{UInt8}(s) === Array{UInt8}(s) === codeunits(s) === b
+    @test Vector{UInt8}(s) == Array{UInt8}(s) == codeunits(s) === b
     @test Vector{UInt8}(StringView(@view b[1:3])) == b[1:3]
     @test codeunits(String(s)) == s.data
     @test Vector{UInt8}(abc) == collect(0x61:0x63)
@@ -21,21 +21,11 @@ su = StringView("föôẞαr")
     @test c == "foobar"
     @test c.data !== s.data
 
-    buf = IOBuffer()
-    write(buf, s)
-    @test StringView(buf) == s == StringView(buf, 0x01:0x06)
-    @test StringView(buf, 3:5) == "oba" == StringView(buf, 0x03:0x01:0x05)
-    write(buf, "baz")
-    @test StringView(buf) == s * "baz"
-    @test String(take!(buf)) == s * "baz"
-    @test StringView(buf) == ""
-    @test_throws BoundsError StringView(buf, 3:4)
-
-    @test StringView("foo") isa StringView{Base.CodeUnits{UInt8,String}}
+    @test stringview("foo") isa StringView{Base.CodeUnits{UInt8,String}}
 
     @test s isa StringViews.DenseStringView
     @test StringView(@view b[1:3]) isa StringViews.DenseStringView
-    @test StringView("foo") isa StringViews.DenseStringView
+    @test stringview("foo") isa StringViews.DenseStringView
     @test StringView(@view codeunits("foobar")[1:3]) isa StringViews.DenseStringView
 
     @test pointer(s) == pointer(b) == Base.unsafe_convert(Ptr{UInt8}, s)
@@ -63,8 +53,8 @@ end
 
     @test Base.print_to_string(ss) == "ooba"
 
-    @test cmp("foobar","bar") == cmp(ss,"bar") == -cmp("bar",ss) == cmp(ss,StringView("bar"))
-    @test ss == StringView("ooba") == "ooba" == ss == "ooba"
+    @test cmp("foobar","bar") == cmp(ss,"bar") == -cmp("bar",ss) == cmp(ss,stringview("bar"))
+    @test ss == stringview("ooba") == "ooba" == ss == "ooba"
     @test isvalid(ss)
 end
 
@@ -94,17 +84,17 @@ end
     @test findnext(r"[aeiou]+", s, 1) == 2:3
     @test findnext(r"[aeiou]+", ss, 1) == 1:2
 
-    sv = StringView(codeunits("foo 1234 bar"))
+    sv = stringview("foo 1234 bar")
     @test match(r"[0-9]+", sv).match.string === sv
     @test eltype(eachmatch(r"[0-9]+", sv)) == SVRegexMatch{typeof(sv)}
 
     # Regex match of substring of stringview
-    strv = only(match(r"^([a-z]+)$", SubString(StringView((b"abc")))))
+    strv = only(match(r"^([a-z]+)$", SubString(StringView(b"abc"))))
     @test typeof(strv) == SubString{StringView{typeof(b"abc")}}
 end
 
 @testset "named subpatterns" begin
-    m = match(r"(?<a>.)(.)(?<b>.)", StringView(codeunits("xyz")))
+    m = match(r"(?<a>.)(.)(?<b>.)", stringview("xyz"))
     @test haskey(m, :a)
     @test haskey(m, 2)
     @test haskey(m, "b")
@@ -118,7 +108,7 @@ end
 @testset "parsing" begin
     for val in (true, 1234, 1234.5, 1234.5f0, 4.5+3.25im)
         sval = string(val)
-        for str in (StringView(sval), SubString("foo"*sval*"bar", 4, 3+length(sval)))
+        for str in (stringview(sval), SubString("foo"*sval*"bar", 4, 3+length(sval)))
             @test parse(typeof(val), str) === val
         end
     end
@@ -139,18 +129,18 @@ end
         @test findnext(==("ba"), str, 1) === findnext(==("ba"), sS, 1)
         @test findprev(==("ba"), str, n) === findprev(==("ba"), sS, n)
     end
-    @test chomp(StringView("foo\n")) == "foo"
+    @test chomp(stringview("foo\n")) == "foo"
 
     # issue #5
     let v = [0x32, 0x30, 0x32, 0x31, 0x2d, 0x31, 0x31, 0x2d, 0x31, 0x30, 0x20, 0x32, 0x31, 0x3a, 0x34, 0x32, 0x3a, 0x30, 0x35, 0x2e, 0x31, 0x31, 0x35, 0x38, 0x30, 0x37],
         pat = r"(\.[\d]{3})\d+" => s"\g<1>"
         @test replace(String(copy(v)), pat) == replace(StringView(v), pat)
     end
 
-    @test findfirst(==('ø'), StringView("abc")) === nothing
-    @test findfirst(==('ø'), StringView("abæø")) == 5
-    @test findlast(==('ø'), StringView("abc")) === nothing
-    @test findlast(==('ø'), StringView("abæø")) == 5
+    @test findfirst(==('ø'), stringview("abc")) === nothing
+    @test findfirst(==('ø'), stringview("abæø")) == 5
+    @test findlast(==('ø'), stringview("abc")) === nothing
+    @test findlast(==('ø'), stringview("abæø")) == 5
 end
 
 @testset "replace" begin
@@ -164,8 +154,8 @@ end
 end
 
 @testset "miscellaneous" begin
-    @test cmp("foobar","bar") == cmp(s,"bar") == -cmp("bar",s) == cmp(s,StringView("bar"))
-    @test s == StringView("foobar") == "foobar" == s == "foobar" != StringView("bar")
+    @test cmp("foobar","bar") == cmp(s,"bar") == -cmp("bar",s) == cmp(s,stringview("bar"))
+    @test s == stringview("foobar") == "foobar" == s == "foobar" != stringview("bar")
     @test cmp(abc, "bar") == cmp("abc","bar")
 
     @test Base.typemin(s) isa StringView{Vector{UInt8}}
@@ -176,7 +166,7 @@ end
     @test oneunit(su) == oneunit(typeof(su)) == one(su) == ""
 
     @test isascii(s)
-    @test !isascii(StringView("fööbār"))
+    @test !isascii(stringview("fööbār"))
 
     @test isvalid(s)
     @test isvalid(abc)
@@ -188,5 +178,5 @@ end
     end
 
     # issue #12
-    @test_throws StringIndexError StringView(codeunits("fooα"))[1:5]
+    @test_throws StringIndexError stringview("fooα")[1:5]
 end