Update EscapeExternalHandlerValue to keep '#', '[', and ']'

chrome/browser/android/url_sanitize_utils.cc

+// Copyright 2014 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#include "chrome/browser/android/url_sanitize_utils.h"
+
+#include "base/logging.h"
+
+namespace chrome {
+namespace android {
+
+namespace {
+
+static const char kHexString[] = "0123456789ABCDEF";
+inline char IntToHex(int i) {
+  DCHECK_GE(i, 0) << i << " not a hex value";
+  DCHECK_LE(i, 15) << i << " not a hex value";
+  return kHexString[i];
+}
+
+inline bool IsHexChar(unsigned char c) {
+  if ((c >= 'a' && c <= 'f') || (c >= 'A' && c <= 'F') ||
+      (c >= '0' && c <= '9'))
+    return true;
+  return false;
+}
+
+// A fast bit-vector map for ascii characters.
+//
+// Internally stores 256 bits in an array of 8 ints.
+// Does quick bit-flicking to lookup needed characters.
+struct Charmap {
+  bool Contains(unsigned char c) const {
+    return ((map[c >> 5] & (1 << (c & 31))) != 0);
+  }
+
+  uint32 map[8];
+};
+
+// Everything except alphanumerics, and some reserved characters
+// (;/?:@&=+$,-_.!~*'()). See RFC 2396 for the list of reserved characters.
+static const Charmap kQueryOrRefPartCharmap = {{
+  0xffffffffL, 0x5000002dL, 0x78000000L, 0xb8000001L,
+  0xffffffffL, 0xffffffffL, 0xffffffffL, 0xffffffffL
+}};
+
+// Everything except alphanumerics, and some reserved characters
+// (;/:@&=+$,-_.!~*'()). See RFC 2396 for the list of reserved characters.
+static const Charmap kPathPartCharmap = {{
+  0xffffffffL, 0xd000002dL, 0x78000000L, 0xb8000001L,
+  0xffffffffL, 0xffffffffL, 0xffffffffL, 0xffffffffL
+}};

[asanka, 2014/10/14 16:46:03]

Are these tables different from net/base/escape.cc? If not, we should use the
implementation there to escape individual components of the URL.

[Jaekyun Seok (inactive), 2014/10/15 08:02:25]

Yes, both are different from net/base/escape.cc.

Moreover, functions in net/base/escape.cc escape already escaped characters
(%XX). So I can't use them anyway.

[asanka, 2014/10/17 00:07:28]

Doesn't net::EscapeExternalHandlerValue() do what you want?

+
+// Given text to escape and a Charmap defining which values to escape,
+// return an escaped string. If characters are already escaped, they aren't
+// converted.
+static std::string Escape(const std::string& text, const Charmap& charmap) {
+  std::string escaped;
+  escaped.reserve(text.length() * 3);
+  for (unsigned int i = 0; i < text.length(); ++i) {
+    unsigned char c = static_cast<unsigned char>(text[i]);
+    if ('%' == c && i + 2 < text.length() &&
+        IsHexChar(static_cast<unsigned char>(text[i + 1])) &&
+        IsHexChar(static_cast<unsigned char>(text[i + 2]))) {
+      escaped.push_back('%');
+    } else if (charmap.Contains(c)) {
+      escaped.push_back('%');
+      escaped.push_back(IntToHex(c >> 4));
+      escaped.push_back(IntToHex(c & 0xf));
+    } else {
+      escaped.push_back(c);
+    }
+  }
+  return escaped;
+}
+
+}  // namespace
+
+GURL SanitizeUrl(const GURL& url) {
+  std::string query_input = url.query();
+  std::string ref_input = url.ref();
+  std::string path_input = url.path();
+  std::string query_output = Escape(query_input, kQueryOrRefPartCharmap);
+  std::string ref_output = Escape(ref_input, kQueryOrRefPartCharmap);
+  std::string path_output = Escape(path_input, kPathPartCharmap);
+
+  GURL::Replacements replacements;
+  bool modified = false;
+  if (!query_output.empty() && query_output != query_input) {
+    replacements.SetQueryStr(query_output);
+    modified = true;
+  }
+  if (!ref_output.empty() && ref_output != ref_input) {
+    replacements.SetRefStr(ref_output);
+    modified = true;
+  }
+  if (!path_output.empty() && path_output != path_input) {
+    replacements.SetPathStr(path_output);
+    modified = true;
+  }
+
+  if (modified)
+    return url.ReplaceComponents(replacements);
+  return url;
+}
+
+}  // namespace android
+}  // namespace chrome