Fix #11464 map on a UTF16String should return a UTF16String

JuliaLang · Jun 17, 2015 · 4ad0aa0 · 4ad0aa0
1 parent dbe94d1
commit 4ad0aa0
Show file tree

Hide file tree

Showing 2 changed files with 31 additions and 0 deletions.
diff --git a/base/utf16.jl b/base/utf16.jl
@@ -143,3 +143,24 @@ function utf16(p::Union{Ptr{UInt16}, Ptr{Int16}})
     while unsafe_load(p, len+1) != 0; len += 1; end
     utf16(p, len)
 end
+
+function map(fun, str::UTF16String)
+    buf = UInt16[]
+    sizehint!(buf, length(str.data))
+    for ch in str
+        c2 = fun(ch)
+        !isa(c2, Char) &&
+            throw(UnicodeError(UTF_ERR_MAP_CHAR, 0, 0))
+        uc = reinterpret(UInt32, c2)
+        if uc < 0x10000
+            push!(buf, UInt16(uc))
+        elseif uc <= 0x10ffff
+            push!(buf, UInt16(0xd7c0 + (uc >> 10)))
+            push!(buf, UInt16(0xdc00 + (uc & 0x3ff)))
+        else
+            throw(UnicodeError(UTF_ERR_INVALID_CHAR, 0, uc))
+        end
+    end
+    push!(buf, 0)
+    UTF16String(buf)
+end
diff --git a/test/strings.jl b/test/strings.jl
@@ -1732,3 +1732,13 @@ d = UTF32String(c)
 c[1] = 'A'
 @test d=="A"
 
+# issue # 11464: uppercase/lowercase of UTF16String becomes a UTF8String
+
+@test typeof(uppercase("abcdef")) == ASCIIString
+@test typeof(uppercase(utf8("abcdef"))) == UTF8String
+@test typeof(uppercase(utf16("abcdef"))) == UTF16String
+@test typeof(uppercase(utf32("abcdef"))) == UTF32String
+@test typeof(lowercase("ABCDEF")) == ASCIIString
+@test typeof(lowercase(utf8("ABCDEF"))) == UTF8String
+@test typeof(lowercase(utf16("ABCDEF"))) == UTF16String
+@test typeof(lowercase(utf32("ABCDEF"))) == UTF32String