Fix str_next and findfirst issues

Author: ScottPJones

src/compare.jl

@@ -54,7 +54,7 @@ end
     while pnt < fin
         str_done(b, pos) && return 1
         c1, pnt = _nextcp(C, pnt)
-        ch, pos = str_next(b, pos)
+        ch, pos = iterate(b, pos)
         c2 = ch%UInt32
         c1 == c2 || return ifelse(c1 < c2, -1, 1)
     end
@@ -93,7 +93,7 @@ function _cpeq(a::MaybeSub{T}, b) where {C<:CSE, T<:Str{C}}
     while pnt < fin
         str_done(b, pos) && return false
         c1, pnt = _nextcp(C, pnt)
-        ch, pos = str_next(b, pos)
+        ch, pos = iterate(b, pos)
         c1 == codepoint(ch) || return false
     end
     true

src/search.jl

@@ -130,6 +130,12 @@ Base.findlast(a::Str, b::AbstractString)    = nothing_sentinel(find(Last, a, b))
 Base.findnext(a::Str, b::AbstractString, i) = nothing_sentinel(find(Fwd, a, b, i))
 Base.findprev(a::Str, b::AbstractString, i) = nothing_sentinel(find(Rev, a, b, i))
 
+# Fix ambiguities caused by addition of new findfirst definition to base
+Base.findfirst(a::AbstractChar, b::Str)   = nothing_sentinel(find(First, a, b))
+Base.findlast(a::AbstractChar, b::Str)    = nothing_sentinel(find(Last, a, b))
+Base.findnext(a::AbstractChar, b::Str, i) = nothing_sentinel(find(Fwd, a, b, i))
+Base.findprev(a::AbstractChar, b::Str, i) = nothing_sentinel(find(Rev, a, b, i))
+
 function find(::Type{D}, fun::Function, str::AbstractString, pos::Integer) where {D<:Direction}
     pos < Int(D===Fwd) && (@boundscheck boundserr(str, pos); return 0)
     if pos > (len = ncodeunits(str))
@@ -189,7 +195,7 @@ function find(::Type{D}, needle::AbstractString, str::AbstractString,
     @inbounds is_valid(str, pos) || index_error(str, pos)
     (tlen = ncodeunits(needle)) == 0 && return pos:pos-1
     (cmp = CanContain(str, needle)) === NoCompare() && return _not_found
-    @inbounds ch, nxt = str_next(needle, 1)
+    @inbounds ch, nxt = iterate(needle, 1)
     is_valid(eltype(str), ch) || return _not_found
     # Check if single character
     if nxt > tlen
@@ -205,7 +211,7 @@ function find(::Type{T}, needle::AbstractString, str::AbstractString) where {T<:
     pos = T === First ? 1 : thisind(str, slen)
     (tlen = ncodeunits(needle)) == 0 && return pos:(pos-1)
     (cmp = CanContain(str, needle)) === NoCompare() && return _not_found
-    @inbounds ch, nxt = str_next(needle, 1)
+    @inbounds ch, nxt = iterate(needle, 1)
     is_valid(eltype(str), ch) || return _not_found
     # Check if single character
     if nxt > tlen
@@ -298,8 +304,8 @@ end
 """Compare two strings, starting at nxtstr and nxtsub"""
 @inline function _cmp_str(str, strpos, endpos, sub, subpos, endsub)
     while strpos <= endpos
-        c, strnxt = str_next(str, strpos)
-        d, subpos = str_next(sub, subpos)
+        c, strnxt = iterate(str, strpos)
+        d, subpos = iterate(sub, subpos)
         c == d || break
         subpos > endsub && return strpos
         strpos = strnxt

src/support.jl

@@ -264,7 +264,7 @@ function unsafe_check_string(str::T;
     totalchar = latin1byte = num2byte = num3byte = num4byte = invalids = 0
     pos = 1
     @inbounds while !str_done(str, pos)
-        chr, nxt = str_next(str, pos)
+        chr, nxt = iterate(str, pos)
         ch = chr%UInt32
         totalchar += 1
         if ch > 0x7f
@@ -288,7 +288,7 @@ function unsafe_check_string(str::T;
                     break
                 end
                 # next character *must* be a trailing surrogate character
-                chr, nxt = str_next(str, nxt)
+                chr, nxt = iterate(str, nxt)
                 if !is_surrogate_trail(chr)
                     accept_invalids || strerror(StrErrors.NOT_TRAIL, pos, chr)
                     invalids += 1

src/utf8.jl

@@ -387,9 +387,9 @@ _iterate(::MultiCU, ::Type{T}, str::SubString{<:Str{RawUTF8CSE}}, pos::Int) wher
 end
 
 _next(::MultiCU, ::Type{T}, str::Str{RawUTF8CSE}, pos::Int) where {T} =
-    str_next(str.data, pos)
+    iterate(str.data, pos)
 _next(::MultiCU, ::Type{T}, str::SubString{<:Str{RawUTF8CSE}}, pos::Int) where {T} =
-    str_next(SubString(str.string.data, str.offset + pos, str.offset + ncodeunits(str)), 1)
+    iterate(SubString(str.string.data, str.offset + pos, str.offset + ncodeunits(str)), 1)
 
 ## overload methods for efficiency ##
 

test/basic.jl

@@ -305,8 +305,8 @@ let
 
     @test lastindex(srep) == 7
 
-    @test str_next(srep, 3) == ('β',5)
-    @test str_next(srep, 7) == ('β',9)
+    @test iterate(srep, 3) == ('β',5)
+    @test iterate(srep, 7) == ('β',9)
 
     @test srep[7] == 'β'
     @test_throws StringIndexError srep[8]
@@ -340,8 +340,8 @@ end
     @test_throws MethodError codeunit(tstr, true)
     @test_throws MethodError isvalid(tstr, 1)
     @test_throws MethodError isvalid(tstr, true)
-    @test_throws MethodError str_next(tstr, 1)
-    @test_throws MethodError str_next(tstr, true)
+    @test_throws MethodError iterate(tstr, 1)
+    @test_throws MethodError iterate(tstr, true)
     @test_throws MethodError lastindex(tstr)
 
     gstr = GenericString("12")
@@ -611,7 +611,7 @@ end
     for st in ("Hello", "Σ", "こんにちは", "😊😁")
         local s
         s = ST(st)
-        @test str_next(s, lastindex(s))[2] > sizeof(s)
+        @test iterate(s, lastindex(s))[2] > sizeof(s)
         @test nextind(s, lastindex(s)) > sizeof(s)
     end
 end
@@ -915,7 +915,7 @@ function testbin(::Type{ST}) where {ST}
                  b"\xf8\x9f\x98\x84", b"\xf8\x9f\x98\x84z")),
         s in lst
         st = ST(s)
-        @test str_next(st, 1)[2] == 2
+        @test iterate(st, 1)[2] == 2
         @test nextind(st, 1) == 2
     end
 
@@ -930,7 +930,7 @@ function testbin(::Type{ST}) where {ST}
         (s, r) in lst
         st = ST(s)
         (ST === BinaryStr || ST === Text1Str) && (r = 2)
-        @test str_next(st, 1)[2] == r
+        @test iterate(st, 1)[2] == r
         @test nextind(st, 1) == r
     end
 end