Fix encoding of high surrogate in UTF-16

CHANGES

@@ -4,7 +4,7 @@ Next version
   causing https://caml.inria.fr/mantis/view.php?id=7603
 - Support for passing argument through external files (-arg/-arg0)
   (Bernhard Schommer)
-
+- Fix encoding of high surrogate for U+10000-U+10FFFF in UTF-16 response files (David Allsopp)
 
 Version 0.36
 - Add Unicode support (patch by Nicolás Ojeda Bär)

Compat403.ml

@@ -38,3 +38,8 @@ module String = struct
       (map Char.lowercase_ascii, map Char.uppercase_ascii)
 end
 
+module Uchar = struct
+  let unsafe_of_int c = c
+
+  let to_int c = c
+end

Compat406.ml

@@ -0,0 +1,27 @@
+(************************************************************************)
+(*   FlexDLL                                                            *)
+(*   Alain Frisch                                                       *)
+(*                                                                      *)
+(*   Copyright 2007 Institut National de Recherche en Informatique et   *)
+(*   en Automatique.                                                    *)
+(************************************************************************)
+
+module Buffer = struct
+  include Buffer
+
+  (* Taken from 4.06.0 *)
+  let add_utf_16le_uchar b u = match Uchar.to_int u with
+  | u when u < 0 -> assert false
+  | u when u <= 0xFFFF ->
+      Buffer.add_char b (Char.unsafe_chr (u land 0xFF));
+      Buffer.add_char b (Char.unsafe_chr (u lsr 8))
+  | u when u <= 0x10FFFF ->
+      let u' = u - 0x10000 in
+      let hi = 0xD800 lor (u' lsr 10) in
+      let lo = 0xDC00 lor (u' land 0x3FF) in
+      Buffer.add_char b (Char.unsafe_chr (hi land 0xFF));
+      Buffer.add_char b (Char.unsafe_chr (hi lsr 8));
+      Buffer.add_char b (Char.unsafe_chr (lo land 0xFF));
+      Buffer.add_char b (Char.unsafe_chr (lo lsr 8))
+  | _ -> assert false
+end

Makefile

@@ -131,7 +131,7 @@ COMPILER-$(OCAML_VERSION):
 
 test_ver = $(shell if [ $(OCAML_VERSION) -lt $(1) ] ; then echo lt ; fi)
 
-Compat.ml: COMPILER-$(OCAML_VERSION) $(if $(call test_ver,4050),Compat405.ml) $(if $(call test_ver,4030),Compat403.ml) $(if $(call test_ver,4020),Compat402.ml)
+Compat.ml: COMPILER-$(OCAML_VERSION) $(if $(call test_ver,4020),Compat402.ml) $(if $(call test_ver,4030),Compat403.ml) $(if $(call test_ver,4050),Compat405.ml) $(if $(call test_ver,4060),Compat406.ml)
 	cat $^ > $@
 
 flexlink.exe: $(OBJS) $(RES)

appveyor_build.sh

@@ -61,6 +61,10 @@ case $OCAMLBRANCH in
         ;;
 esac
 
+if [ $OCAMLBRANCH = "4.03" ] ; then
+  sed -i -e "s/:=.*/:=/" config/Makefile.msvc64
+fi
+
 configure_ocaml
 
 if [ ! -f $OCAMLROOT/STAMP ] || [ "$(git rev-parse HEAD)" != "$(cat $OCAMLROOT/STAMP)" ]; then

reloc.ml

@@ -147,11 +147,9 @@ let utf8_next s i =
 
 let toutf16 s =
   let i = ref 0 in
-  let b = Buffer.create (String.length s) in
-  let cp n = Buffer.add_char b (Char.chr (n land 0xFF)); Buffer.add_char b (Char.chr ((n lsr 8) land 0xFF)) in
+  let b = Buffer.create (String.length s * 2) in
   while !i < String.length s do
-    let n = utf8_next s i in
-    if n <= 0xFFFF then cp n else (cp (0xD7C0 + (n lsl 10)); cp (0xDC00 + (n land 0x3FF)))
+    Buffer.add_utf_16le_uchar b (Uchar.unsafe_of_int (utf8_next s i))
   done;
   Buffer.contents b