Merge pull request #24708 from JuliaLang/sk/revstring

remove `RevString`; efficient generic `reverseind`

NEWS.md

@@ -308,6 +308,12 @@ This section lists changes that do not have deprecation warnings.
     `AbstractArray` types that specialized broadcasting using the old internal API will
     need to switch to the new API. ([#20740])
 
+  * The `RevString` type has been removed from the language; `reverse(::String)` returns
+    a `String` with code points (or fragments thereof) in reverse order. In general,
+    `reverse(s)` should return a string of the same type and encoding as `s` with code
+    points in reverse order; any string type overrides `reverse` to return a different
+    type of string must also override `reverseind` to compute reversed indices correctly.
+
 Library improvements
 --------------------
 
@@ -409,6 +415,15 @@ Library improvements
   * The `keys` of an `Associative` are now an `AbstractSet`. `Base.KeyIterator{<:Associative}`
     has been changed to `KeySet{K, <:Associative{K}} <: AbstractSet{K}` ([#24580]).
 
+  * New function `ncodeunits(s::AbstractString)` gives the number of code units in a string.
+    The generic definition is constant time but calls `endof(s)` which may be inefficient.
+    Therefore custom string types may want to define direct `ncodeunits` methods.
+
+  * `reverseind(s::AbstractString, i::Integer)` now has an efficient generic fallback, so
+    custom string types do not need to provide their own efficient defintions. The generic
+    definition relies on `ncodeunits` however, so for optimal performance you may need to
+    define a custom method for that function.
+
 Compiler/Runtime improvements
 -----------------------------
 

base/exports.jl

@@ -88,7 +88,6 @@ export
     Rational,
     Regex,
     RegexMatch,
-    RevString,
     RoundFromZero,
     RoundDown,
     RoundingMode,
@@ -756,6 +755,7 @@ export
     lstrip,
     match,
     matchall,
+    ncodeunits,
     ndigits,
     nextind,
     normalize_string,

base/precompile.jl

@@ -578,9 +578,6 @@ precompile(Tuple{typeof(Base.LineEdit.complete_line), Base.LineEdit.PromptState,
 precompile(Tuple{typeof(Base.LineEdit.input_string_newlines_aftercursor), Base.LineEdit.PromptState})
 precompile(Tuple{typeof(Base.LineEdit.complete_line), Base.REPL.REPLCompletionProvider, Base.LineEdit.PromptState})
 precompile(Tuple{getfield(Base, Symbol("#kw##parse")), Array{Any, 1}, typeof(Base.parse), String})
-precompile(Tuple{typeof(Base.isvalid), Base.RevString{String}, Int64})
-precompile(Tuple{typeof(Base.nextind), Base.RevString{String}, Int64})
-precompile(Tuple{typeof(Base.search), Base.RevString{String}, Array{Char, 1}, Int64})
 precompile(Tuple{typeof(Base.rsearch), String, Array{Char, 1}, Int64})
 precompile(Tuple{getfield(Base.REPLCompletions, Symbol("#kw##find_start_brace")), Array{Any, 1}, typeof(Base.REPLCompletions.find_start_brace), String})
 precompile(Tuple{typeof(Core.Inference.isbits), Tuple{Void, Void, Void}})

base/repl/REPLCompletions.jl

@@ -225,7 +225,7 @@ end
 # closed start brace from the end of the string.
 function find_start_brace(s::AbstractString; c_start='(', c_end=')')
     braces = 0
-    r = RevString(s)
+    r = reverse(s)
     i = start(r)
     in_single_quotes = false
     in_double_quotes = false
@@ -245,18 +245,21 @@ function find_start_brace(s::AbstractString; c_start='(', c_end=')')
                 in_back_ticks = true
             end
         else
-            if !in_back_ticks && !in_double_quotes && c == '\'' && !done(r, i) && next(r, i)[1]!='\\'
+            if !in_back_ticks && !in_double_quotes &&
+                c == '\'' && !done(r, i) && next(r, i)[1] != '\\'
                 in_single_quotes = !in_single_quotes
-            elseif !in_back_ticks && !in_single_quotes && c == '"' && !done(r, i) && next(r, i)[1]!='\\'
+            elseif !in_back_ticks && !in_single_quotes &&
+                c == '"' && !done(r, i) && next(r, i)[1] != '\\'
                 in_double_quotes = !in_double_quotes
-            elseif !in_single_quotes && !in_double_quotes && c == '`' && !done(r, i) && next(r, i)[1]!='\\'
+            elseif !in_single_quotes && !in_double_quotes &&
+                c == '`' && !done(r, i) && next(r, i)[1] != '\\'
                 in_back_ticks = !in_back_ticks
             end
         end
         braces == 1 && break
     end
     braces != 1 && return 0:-1, -1
-    method_name_end = reverseind(r, i)
+    method_name_end = reverseind(s, i)
     startind = nextind(s, rsearch(s, non_identifier_chars, method_name_end))
     return (startind:endof(s), method_name_end)
 end

base/shell.jl

@@ -14,7 +14,7 @@ function shell_parse(str::AbstractString, interpolate::Bool=true;
                      special::AbstractString="")
     s = lstrip(str)
     # strips the end but respects the space when the string ends with "\\ "
-    r = RevString(s)
+    r = reverse(s)
     i = start(r)
     c_old = nothing
     while !done(r,i)

base/strings/basic.jl

@@ -69,6 +69,10 @@ julia> 'j' * "ulia"
 
 one(::Union{T,Type{T}}) where {T<:AbstractString} = convert(T, "")
 
+# generic number of code units; implementations generally know how long a string
+# is though and should override this with a more efficient method
+ncodeunits(s::AbstractString) = nextind(s, endof(s)) - 1
+
 """
     length(s::AbstractString)
 
@@ -233,11 +237,11 @@ end
 ## Generic indexing functions ##
 
 """
-    thisind(str::AbstractString, i::Integer)
+    thisind(s::AbstractString, i::Integer)
 
-Get the largest valid string index at or before `i`.
-Returns `0` if there is no valid string index at or before `i`.
-Returns `endof(str)` if `i≥endof(str)`.
+If `i` is the index into a character in `s` then `thisind` returns the index of the
+start of that character. If `i < start(s)` then it returns `start(s) - 1`.
+If `i > ncodeunits(s)` then it returns `ncodeunits(s) + 1`.
 
 # Examples
 ```jldoctest
@@ -253,15 +257,21 @@ julia> thisind("αβγdef", 3)
 julia> thisind("αβγdef", 4)
 3
 
-julia> thisind("αβγdef", 20)
+julia> thisind("αβγdef", 9)
 9
+
+julia> thisind("αβγdef", 10)
+10
+
+julia> thisind("αβγdef", 20)
+10
 """
 function thisind(s::AbstractString, i::Integer)
     j = Int(i)
     isvalid(s, j) && return j
     j < start(s) && return 0
-    e = endof(s)
-    j >= endof(s) && return e
+    n = ncodeunits(s)
+    j > n && return n + 1
     prevind(s, j)
 end
 

base/strings/search.jl

@@ -194,12 +194,6 @@ end
 search(s::AbstractString, t::AbstractString, i::Integer=start(s)) = _search(s, t, i)
 search(s::ByteArray, t::ByteArray, i::Integer=start(s)) = _search(s, t, i)
 
-function rsearch(s::AbstractString, c::Chars)
-    j = search(RevString(s), c)
-    j == 0 && return 0
-    endof(s)-j+1
-end
-
 """
     rsearch(s::AbstractString, chars::Chars, [start::Integer])
 
@@ -212,44 +206,50 @@ julia> rsearch("aaabbb","b")
 6:6
 ```
 """
-function rsearch(s::AbstractString, c::Chars, i::Integer)
-    e = endof(s)
-    j = search(RevString(s), c, e-i+1)
-    j == 0 && return 0
-    e-j+1
+function rsearch(s::AbstractString, c::Chars, i::Integer=start(s))
+    if i < 1
+        return i == 0 ? 0 : throw(BoundsError(s, i))
+    end
+    n = ncodeunits(s)
+    if i > n
+        return i == n+1 ? 0 : throw(BoundsError(s, i))
+    end
+    # r[reverseind(r,i)] == reverse(r)[i] == s[i]
+    # s[reverseind(s,j)] == reverse(s)[j] == r[j]
+    r = reverse(s)
+    j = search(r, c, reverseind(r, i))
+    j == 0 ? 0 : reverseind(s, j)
 end
 
 function _rsearchindex(s, t, i)
     if isempty(t)
-        return 1 <= i <= nextind(s,endof(s)) ? i :
+        return 1 <= i <= nextind(s, endof(s)) ? i :
                throw(BoundsError(s, i))
     end
-    t = RevString(t)
-    rs = RevString(s)
+    t = reverse(t)
+    rs = reverse(s)
     l = endof(s)
-    t1, j2 = next(t,start(t))
+    t1, j2 = next(t, start(t))
     while true
-        i = rsearch(s,t1,i)
-        if i == 0 return 0 end
-        c, ii = next(rs,l-i+1)
+        i = rsearch(s, t1, i)
+        i == 0 && return 0
+        c, ii = next(rs, reverseind(rs, i))
         j = j2; k = ii
         matched = true
-        while !done(t,j)
-            if done(rs,k)
+        while !done(t, j)
+            if done(rs, k)
                 matched = false
                 break
             end
-            c, k = next(rs,k)
-            d, j = next(t,j)
+            c, k = next(rs, k)
+            d, j = next(t, j)
             if c != d
                 matched = false
                 break
             end
         end
-        if matched
-            return nextind(s,l-k+1)
-        end
-        i = l-ii+1
+        matched && return nextind(s, reverseind(s, k))
+        i = reverseind(s, ii)
     end
 end
 

base/strings/string.jl

@@ -87,6 +87,20 @@ codeunit(s::AbstractString, i::Integer)
     @gc_preserve s unsafe_load(pointer(s, i))
 end
 
+"""
+    ncodeunits(s::AbstractString)
+
+The number of code units in a string. For example, for UTF-8-like data such as
+the default `String` type, the number of code units is the number of bytes in
+the string, a.k.a. `sizeof(s)`. For a UTF-16 encoded string type, however, the
+code unit is `UInt16` so the number of code units is the number of `UInt16`
+words in the representation of the string. The expression `codeunit(s, i)` is
+valid and safe for precisely the range of `i` values `1:ncodeunits(s)`.
+
+See also: [`codeunit`](@ref).
+"""
+ncodeunits(s::String) = sizeof(s)
+
 write(io::IO, s::String) =
     @gc_preserve s unsafe_write(io, pointer(s), reinterpret(UInt, sizeof(s)))
 
@@ -109,8 +123,8 @@ end
 function thisind(s::String, i::Integer)
     j = Int(i)
     j < 1 && return 0
-    e = endof(s)
-    j >= e && return e
+    n = ncodeunits(s)
+    j > n && return n + 1
     @inbounds while j > 0 && is_valid_continuation(codeunit(s,j))
         j -= 1
     end
@@ -281,14 +295,6 @@ function first_utf8_byte(ch::Char)
     return b
 end
 
-function reverseind(s::String, i::Integer)
-    j = sizeof(s) + 1 - i
-    @inbounds while is_valid_continuation(codeunit(s, j))
-        j -= 1
-    end
-    return j
-end
-
 ## overload methods for efficiency ##
 
 isvalid(s::String, i::Integer) =
@@ -463,38 +469,6 @@ function string(a::Union{String,Char}...)
     return out
 end
 
-function reverse(s::String)
-    dat = Vector{UInt8}(s)
-    n = length(dat)
-    n <= 1 && return s
-    buf = StringVector(n)
-    out = n
-    pos = 1
-    @inbounds while out > 0
-        ch = dat[pos]
-        if ch > 0xdf
-            if ch < 0xf0
-                (out -= 3) < 0 && throw(UnicodeError(UTF_ERR_SHORT, pos, ch))
-                buf[out + 1], buf[out + 2], buf[out + 3] = ch, dat[pos + 1], dat[pos + 2]
-                pos += 3
-            else
-                (out -= 4) < 0 && throw(UnicodeError(UTF_ERR_SHORT, pos, ch))
-                buf[out+1], buf[out+2], buf[out+3], buf[out+4] = ch, dat[pos+1], dat[pos+2], dat[pos+3]
-                pos += 4
-            end
-        elseif ch > 0x7f
-            (out -= 2) < 0 && throw(UnicodeError(UTF_ERR_SHORT, pos, ch))
-            buf[out + 1], buf[out + 2] = ch, dat[pos + 1]
-            pos += 2
-        else
-            buf[out] = ch
-            out -= 1
-            pos += 1
-        end
-    end
-    String(buf)
-end
-
 function repeat(s::String, r::Integer)
     r < 0 && throw(ArgumentError("can't repeat a string $r times"))
     n = sizeof(s)

base/strings/strings.jl

@@ -1,7 +1,7 @@
 # This file is a part of Julia. License is MIT: https://julialang.org/license
 
 include("strings/errors.jl")
-include("strings/types.jl")
+include("strings/substring.jl")
 include("strings/basic.jl")
 include("strings/search.jl")
 include("strings/util.jl")