From 319eefea3b2097b543751c63cebbb4a92011dbbf Mon Sep 17 00:00:00 2001
From: Rickard Green <rickard@erlang.org>
Date: Mon, 21 Aug 2017 15:25:00 +0200
Subject: Unicode support for erlang:display_string/1

---
 erts/emulator/beam/bif.c         | 15 +++++++------
 erts/emulator/beam/global.h      |  1 +
 erts/emulator/beam/utils.c       | 46 ++++++++++++++++++++++++++++++++++++++++
 erts/emulator/test/bif_SUITE.erl | 26 +++++++++++++++++++++--
 4 files changed, 80 insertions(+), 8 deletions(-)

(limited to 'erts')

diff --git a/erts/emulator/beam/bif.c b/erts/emulator/beam/bif.c
index 4c92b9ba2b..22805b5c02 100644
--- a/erts/emulator/beam/bif.c
+++ b/erts/emulator/beam/bif.c
@@ -3947,15 +3947,18 @@ BIF_RETTYPE display_string_1(BIF_ALIST_1)
 {
     Process* p = BIF_P;
     Eterm string = BIF_ARG_1;
-    Sint len = is_string(string);
-    char *str;
+    Sint len = erts_unicode_list_to_buf_len(string);
+    Sint written;
+    byte *str;
+    int res;
 
-    if (len <= 0) {
+    if (len < 0) {
 	BIF_ERROR(p, BADARG);
     }
-    str = (char *) erts_alloc(ERTS_ALC_T_TMP, sizeof(char)*(len + 1));
-    if (intlist_to_buf(string, str, len) != len)
-	erts_exit(ERTS_ERROR_EXIT, "%s:%d: Internal error\n", __FILE__, __LINE__);
+    str = (byte *) erts_alloc(ERTS_ALC_T_TMP, sizeof(char)*(len + 1));
+    res = erts_unicode_list_to_buf(string, str, len, &written);
+    if (res != 0 || written != len)
+	erts_exit(ERTS_ERROR_EXIT, "%s:%d: Internal error (%d)\n", __FILE__, __LINE__, res);
     str[len] = '\0';
     erts_fprintf(stderr, "%s", str);
     erts_free(ERTS_ALC_T_TMP, (void *) str);
diff --git a/erts/emulator/beam/global.h b/erts/emulator/beam/global.h
index fcb88712e9..64e543586a 100644
--- a/erts/emulator/beam/global.h
+++ b/erts/emulator/beam/global.h
@@ -1288,6 +1288,7 @@ int erts_utf8_to_latin1(byte* dest, const byte* source, int slen);
 void bin_write(fmtfn_t, void*, byte*, size_t);
 Sint intlist_to_buf(Eterm, char*, Sint); /* most callers pass plain char*'s */
 int erts_unicode_list_to_buf(Eterm list, byte *buf, Sint len, Sint* written);
+Sint erts_unicode_list_to_buf_len(Eterm list);
 
 struct Sint_buf {
 #if defined(ARCH_64)
diff --git a/erts/emulator/beam/utils.c b/erts/emulator/beam/utils.c
index ba7596cde4..16523659d9 100644
--- a/erts/emulator/beam/utils.c
+++ b/erts/emulator/beam/utils.c
@@ -3704,6 +3704,52 @@ erts_unicode_list_to_buf(Eterm list, byte *buf, Sint len, Sint* written)
     return res;
 }
 
+Sint
+erts_unicode_list_to_buf_len(Eterm list)
+{
+    Eterm* listptr;
+    Sint sz = 0;
+
+    if (is_nil(list)) {
+	return 0;
+    }
+    if (is_not_list(list)) {
+	return -1;
+    }
+    listptr = list_val(list);
+
+    while (1) {
+	Sint val;
+
+	if (is_not_small(CAR(listptr))) {
+	    return -1;
+	}
+	val = signed_val(CAR(listptr));
+	if (0 <= val && val < 0x80) {
+	    sz++;
+	} else if (val < 0x800) {
+	    sz += 2;
+	} else if (val < 0x10000UL) {
+	    if (0xD800 <= val && val <= 0xDFFF) {
+		return -1;
+	    }
+	    sz += 3;
+	} else if (val < 0x110000) {
+	    sz += 4;
+	} else {
+	    return -1;
+	}
+	list = CDR(listptr);
+	if (is_nil(list)) {
+	    return sz;
+	}
+	if (is_not_list(list)) {
+	    return -1;
+	}
+	listptr = list_val(list);
+    }
+}
+
 /*
 ** Convert an integer to a byte list
 ** return pointer to converted stuff (need not to be at start of buf!)
diff --git a/erts/emulator/test/bif_SUITE.erl b/erts/emulator/test/bif_SUITE.erl
index 4908ac410c..2320870a0e 100644
--- a/erts/emulator/test/bif_SUITE.erl
+++ b/erts/emulator/test/bif_SUITE.erl
@@ -24,7 +24,7 @@
 -include_lib("kernel/include/file.hrl").
 
 -export([all/0, suite/0,
-	 display/1, display_huge/0,
+	 display/1, display_huge/0, display_string/1,
 	 erl_bif_types/1,guard_bifs_in_erl_bif_types/1,
 	 shadow_comments/1,list_to_utf8_atom/1,
 	 specs/1,improper_bif_stubs/1,auto_imports/1,
@@ -43,7 +43,7 @@ all() ->
     [erl_bif_types, guard_bifs_in_erl_bif_types, shadow_comments,
      specs, improper_bif_stubs, auto_imports,
      t_list_to_existing_atom, os_env, otp_7526,
-     display, list_to_utf8_atom,
+     display, display_string, list_to_utf8_atom,
      atom_to_binary, binary_to_atom, binary_to_existing_atom,
      erl_crash_dump_bytes, min_max, erlang_halt, is_builtin,
      error_stacktrace, error_stacktrace_during_call_trace].
@@ -68,6 +68,28 @@ deeep(N,Acc) ->
 deeep(N) ->
     deeep(N,[hello]).
 
+display_string(Config) when is_list(Config) ->
+    true = erlang:display_string("hej"),
+    true = erlang:display_string(""),
+    true = erlang:display_string("hopp"),
+    true = erlang:display_string("\n"),
+    true = erlang:display_string(lists:seq(1100,1200)),
+    {error,badarg} = try
+                         erlang:display_string(atom),
+                         ok
+                     catch
+                         T0:E0 ->
+                             {T0, E0}
+                     end,
+    {error,badarg} = try
+                         erlang:display_string(make_ref()),
+                         ok
+                     catch
+                         T1:E1 ->
+                             {T1, E1}
+                     end,
+    ok.
+
 erl_bif_types(Config) when is_list(Config) ->
     ensure_erl_bif_types_compiled(),
 
-- 
cgit v1.2.3