From 219729c0419420e4981c5c095159eee98e0f34eb Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Bogumi=C5=82=20Kami=C5=84ski?= <bkamins@sgh.waw.pl>
Date: Tue, 31 Oct 2017 18:53:20 +0100
Subject: [PATCH] make next behave like nextind for String

---
 base/strings/string.jl | 20 +++++++++++++++-----
 test/strings/basic.jl  | 14 ++++++++++++++
 2 files changed, 29 insertions(+), 5 deletions(-)

diff --git a/base/strings/string.jl b/base/strings/string.jl
index 45d1a05b352ca..5fdf82c8764e6 100644
--- a/base/strings/string.jl
+++ b/base/strings/string.jl
@@ -231,12 +231,22 @@ end
     end
     trailing = utf8_trailing[b + 1]
     if l < i + trailing
-        return '\ufffd', i+1
-    end
-    c::UInt32 = 0
-    @inbounds for j = 1:(trailing + 1)
+        for j in (i+1):l
+            @inbounds if !is_valid_continuation(codeunit(s, j))
+                return '\ufffd' j
+            end
+        end
+        return '\ufffd', l+1
+    end
+    @inbounds c::UInt32 = codeunit(s, i)
+    i += 1
+    @inbounds for j = 1:trailing
+        b = codeunit(s, i)
+        if !is_valid_continuation(b)
+            return '\ufffd', i
+        end
         c <<= 6
-        c += codeunit(s, i)
+        c += b
         i += 1
     end
     c -= utf8_offset[trailing + 1]
diff --git a/test/strings/basic.jl b/test/strings/basic.jl
index 0c7cdb54824ae..8fe6b07e1cc7e 100644
--- a/test/strings/basic.jl
+++ b/test/strings/basic.jl
@@ -198,10 +198,24 @@ end
     @test lcfirst("")==""
     @test lcfirst("*")=="*"
 end
+
+# test next for invalid UTF-8 AbstractStrings
+@testset "next for invalid UTF-8" begin
+    srand(1)
+    let s = "1"*String(rand(0x00:0xff, 2^16)), i = 1, j = 1
+        while !done(s, i)
+            c, i = next(s, i)
+            j = nextind(s, j)
+            @test i == j
+        end
+    end
+end
+
 # test AbstractString functions at beginning of string.jl
 struct tstStringType <: AbstractString
     data::Array{UInt8,1}
 end
+
 @testset "AbstractString functions" begin
     tstr = tstStringType(Vector{UInt8}("12"))
     @test_throws ErrorException endof(tstr)