From 237be2ed9e25e4d56deaf3935be0c217913e5a75 Mon Sep 17 00:00:00 2001
From: Anna Henningsen <anna@addaleax.net>
Date: Sat, 26 Oct 2019 16:27:51 +0200
Subject: encoding: make TextDecoder handle BOM correctly

Do not accept the BOM if it comes from a different encoding, and
only discard the BOM after it has actually been read (including
when it is spread over multiple chunks in streaming mode).

Fixes: https://github.com/nodejs/node/issues/25315

PR-URL: https://github.com/nodejs/node/pull/30132
Reviewed-By: Gus Caplan <me@gus.host>
---
 src/node_buffer.cc   |  8 ++++----
 src/node_i18n.cc     | 37 +++++++++++++++++++++++++++----------
 src/node_internals.h |  6 +++++-
 3 files changed, 36 insertions(+), 15 deletions(-)

(limited to 'src')

diff --git a/src/node_buffer.cc b/src/node_buffer.cc
index 74684110a9..3aa1ea2535 100644
--- a/src/node_buffer.cc
+++ b/src/node_buffer.cc
@@ -219,10 +219,10 @@ size_t Length(Local<Object> obj) {
 }
 
 
-inline MaybeLocal<Uint8Array> New(Environment* env,
-                                  Local<ArrayBuffer> ab,
-                                  size_t byte_offset,
-                                  size_t length) {
+MaybeLocal<Uint8Array> New(Environment* env,
+                           Local<ArrayBuffer> ab,
+                           size_t byte_offset,
+                           size_t length) {
   CHECK(!env->buffer_prototype_object().IsEmpty());
   Local<Uint8Array> ui = Uint8Array::New(ab, byte_offset, length);
   Maybe<bool> mb =
diff --git a/src/node_i18n.cc b/src/node_i18n.cc
index 162f5fda5d..ecc0528e76 100644
--- a/src/node_i18n.cc
+++ b/src/node_i18n.cc
@@ -95,6 +95,7 @@ using v8::NewStringType;
 using v8::Object;
 using v8::ObjectTemplate;
 using v8::String;
+using v8::Uint8Array;
 using v8::Value;
 
 namespace i18n {
@@ -227,14 +228,6 @@ class ConverterObject : public BaseObject, Converter {
     const char* source = input.data();
     size_t source_length = input.length();
 
-    if (converter->unicode_ && !converter->ignoreBOM_ && !converter->bomSeen_) {
-      int32_t bomOffset = 0;
-      ucnv_detectUnicodeSignature(source, source_length, &bomOffset, &status);
-      source += bomOffset;
-      source_length -= bomOffset;
-      converter->bomSeen_ = true;
-    }
-
     UChar* target = *result;
     ucnv_toUnicode(converter->conv,
                    &target, target + (limit * sizeof(UChar)),
@@ -242,10 +235,34 @@ class ConverterObject : public BaseObject, Converter {
                    nullptr, flush, &status);
 
     if (U_SUCCESS(status)) {
-      if (limit > 0)
+      bool omit_initial_bom = false;
+      if (limit > 0) {
         result.SetLength(target - &result[0]);
+        if (result.length() > 0 &&
+            converter->unicode_ &&
+            !converter->ignoreBOM_ &&
+            !converter->bomSeen_) {
+          // If the very first result in the stream is a BOM, and we are not
+          // explicitly told to ignore it, then we mark it for discarding.
+          if (result[0] == 0xFEFF) {
+            omit_initial_bom = true;
+          }
+          converter->bomSeen_ = true;
+        }
+      }
       ret = ToBufferEndian(env, &result);
-      args.GetReturnValue().Set(ret.ToLocalChecked());
+      if (omit_initial_bom && !ret.IsEmpty()) {
+        // Peform `ret = ret.slice(2)`.
+        CHECK(ret.ToLocalChecked()->IsUint8Array());
+        Local<Uint8Array> orig_ret = ret.ToLocalChecked().As<Uint8Array>();
+        ret = Buffer::New(env,
+                          orig_ret->Buffer(),
+                          orig_ret->ByteOffset() + 2,
+                          orig_ret->ByteLength() - 2)
+                              .FromMaybe(Local<Uint8Array>());
+      }
+      if (!ret.IsEmpty())
+        args.GetReturnValue().Set(ret.ToLocalChecked());
       return;
     }
 
diff --git a/src/node_internals.h b/src/node_internals.h
index 4ec883c891..2ec230d8b5 100644
--- a/src/node_internals.h
+++ b/src/node_internals.h
@@ -158,7 +158,11 @@ v8::MaybeLocal<v8::Object> New(Environment* env,
                                char* data,
                                size_t length,
                                bool uses_malloc);
-
+// Creates a Buffer instance over an existing Uint8Array.
+v8::MaybeLocal<v8::Uint8Array> New(Environment* env,
+                                   v8::Local<v8::ArrayBuffer> ab,
+                                   size_t byte_offset,
+                                   size_t length);
 // Construct a Buffer from a MaybeStackBuffer (and also its subclasses like
 // Utf8Value and TwoByteValue).
 // If |buf| is invalidated, an empty MaybeLocal is returned, and nothing is
-- 
cgit v1.2.3