Enable PCRE UTF-8 validity string checks

Strings are no guaranteed to contain valid UTF-8, and PCRE documentation says
that the behavior is undefined in that case.

Author: nalimilan

base/regex.jl

@@ -4,8 +4,8 @@
 
 include("pcre.jl")
 
-const DEFAULT_COMPILER_OPTS = PCRE.UTF | PCRE.NO_UTF_CHECK | PCRE.ALT_BSUX
-const DEFAULT_MATCH_OPTS = PCRE.NO_UTF_CHECK
+const DEFAULT_COMPILER_OPTS = PCRE.UTF | PCRE.ALT_BSUX
+const DEFAULT_MATCH_OPTS = zero(UInt32)
 
 mutable struct Regex
     pattern::String

test/regex.jl

@@ -49,3 +49,17 @@ end
 
 # Proper unicode handling
 @test  match(r"∀∀", "∀x∀∀∀").match == "∀∀"
+
+@test_throws ErrorException match(r"a", "\xe2\x88") # 1 byte missing at end
+@test_throws ErrorException match(r"a", "\xe2\x08\x80") # byte 2 top bits not 0x80
+@test_throws ErrorException match(r"a", "\xf8\x89\x89\x80\x80") # 5-byte character is not allowed (RFC 3629)
+@test_throws ErrorException match(r"a", "\UDFFF") # code points 0xd800-0xdfff are not defined
+@test_throws ErrorException match(r"a", "\xc0\x80") #  overlong 2-byte sequence
+@test_throws ErrorException match(r"a", "\xff") # illegal byte (0xfe or 0xff)
+
+@test_throws ErrorException Regex("\xe2\x88") # 1 byte missing at end
+@test_throws ErrorException Regex("\xe2\x08\x80") # byte 2 top bits not 0x80
+@test_throws ErrorException Regex("\xf8\x89\x89\x80\x80") # 5-byte character is not allowed (RFC 3629)
+@test_throws ErrorException Regex("\UDFFF") # code points 0xd800-0xdfff are not defined
+@test_throws ErrorException Regex("\xc0\x80") #  overlong 2-byte sequence
+@test_throws ErrorException Regex("\xff") # illegal byte (0xfe or 0xff)