Print some useful info when hitting some cases of invalid UTF-8 (#24311)

JuliaLang · Oct 26, 2017 · 29fcb37 · 29fcb37
1 parent 7ae9955
commit 29fcb37
Show file tree

Hide file tree

Showing 2 changed files with 16 additions and 1 deletion.
diff --git a/base/strings/errors.jl b/base/strings/errors.jl
@@ -3,7 +3,7 @@
 ##    Error messages for Unicode / UTF support
 
 const UTF_ERR_SHORT             = "invalid UTF-8 sequence starting at index <<1>> (0x<<2>> missing one or more continuation bytes)"
-const UTF_ERR_INVALID_INDEX     = "invalid character index"
+const UTF_ERR_INVALID_INDEX     = "invalid character index <<1>> (0x<<2>> is a continuation byte)"
 
 struct UnicodeError <: Exception
     errmsg::AbstractString   ##< A UTF_ERR_ message

diff --git a/test/strings/basic.jl b/test/strings/basic.jl
@@ -655,3 +655,18 @@ end
     @test last(s, length(s)) == s
     @test_throws BoundsError last(s, length(s)+1)
 end
+
+@testset "invalid code point" begin
+    s = String([0x61, 0xba, 0x41])
+    @test !isvalid(s)
+    @test_throws UnicodeError s[2]
+    e = try
+        s[2]
+    catch e
+        e
+    end
+    b = IOBuffer()
+    show(b, e)
+    @test String(take!(b)) == "UnicodeError: invalid character index 2 (0xba is a continuation byte)"
+end
+