components/url_formatter/elide_url.cc

   1 // Copyright 2014 The Chromium Authors. All rights reserved.
   2 // Use of this source code is governed by a BSD-style license that can be
   3 // found in the LICENSE file.
   4
   5 #include "components/url_formatter/elide_url.h"
   6
   7 #include "base/logging.h"
   8 #include "base/strings/string_split.h"
   9 #include "base/strings/utf_string_conversions.h"
  10 #include "components/url_formatter/url_formatter.h"
  11 #include "net/base/escape.h"
  12 #include "net/base/registry_controlled_domains/registry_controlled_domain.h"
  13 #include "url/gurl.h"
  14 #include "url/url_constants.h"
  15
  16 #if !defined(OS_ANDROID)
  17 #include "ui/gfx/text_elider.h"  // nogncheck
  18 #include "ui/gfx/text_utils.h"  // nogncheck
  19 #endif
  20
  21 namespace {
  22
  23 #if !defined(OS_ANDROID)
  24 const base::char16 kDot = '.';
  25
  26 // Build a path from the first |num_components| elements in |path_elements|.
  27 // Prepends |path_prefix|, appends |filename|, inserts ellipsis if appropriate.
  28 base::string16 BuildPathFromComponents(
  29     const base::string16& path_prefix,
  30     const std::vector<base::string16>& path_elements,
  31     const base::string16& filename,
  32     size_t num_components) {
  33   // Add the initial elements of the path.
  34   base::string16 path = path_prefix;
  35
  36   // Build path from first |num_components| elements.
  37   for (size_t j = 0; j < num_components; ++j)
  38     path += path_elements[j] + gfx::kForwardSlash;
  39
  40   // Add |filename|, ellipsis if necessary.
  41   if (num_components != (path_elements.size() - 1))
  42     path += base::string16(gfx::kEllipsisUTF16) + gfx::kForwardSlash;
  43   path += filename;
  44
  45   return path;
  46 }
  47
  48 // Takes a prefix (Domain, or Domain+subdomain) and a collection of path
  49 // components and elides if possible. Returns a string containing the longest
  50 // possible elided path, or an empty string if elision is not possible.
  51 base::string16 ElideComponentizedPath(
  52     const base::string16& url_path_prefix,
  53     const std::vector<base::string16>& url_path_elements,
  54     const base::string16& url_filename,
  55     const base::string16& url_query,
  56     const gfx::FontList& font_list,
  57     float available_pixel_width) {
  58   const size_t url_path_number_of_elements = url_path_elements.size();
  59
  60   CHECK(url_path_number_of_elements);
  61   for (size_t i = url_path_number_of_elements - 1; i > 0; --i) {
  62     base::string16 elided_path = BuildPathFromComponents(
  63         url_path_prefix, url_path_elements, url_filename, i);
  64     if (available_pixel_width >= gfx::GetStringWidthF(elided_path, font_list))
  65       return gfx::ElideText(elided_path + url_query, font_list,
  66                        available_pixel_width, gfx::ELIDE_TAIL);
  67   }
  68
  69   return base::string16();
  70 }
  71
  72 // Splits the hostname in the |url| into sub-strings for the full hostname,
  73 // the domain (TLD+1), and the subdomain (everything leading the domain).
  74 void SplitHost(const GURL& url,
  75                base::string16* url_host,
  76                base::string16* url_domain,
  77                base::string16* url_subdomain) {
  78   // Get Host.
  79   *url_host = base::UTF8ToUTF16(url.host());
  80
  81   // Get domain and registry information from the URL.
  82   *url_domain =
  83       base::UTF8ToUTF16(net::registry_controlled_domains::GetDomainAndRegistry(
  84           url, net::registry_controlled_domains::EXCLUDE_PRIVATE_REGISTRIES));
  85   if (url_domain->empty())
  86     *url_domain = *url_host;
  87
  88   // Add port if required.
  89   if (!url.port().empty()) {
  90     *url_host += base::UTF8ToUTF16(":" + url.port());
  91     *url_domain += base::UTF8ToUTF16(":" + url.port());
  92   }
  93
  94   // Get sub domain.
  95   const size_t domain_start_index = url_host->find(*url_domain);
  96   base::string16 kWwwPrefix = base::UTF8ToUTF16("www.");
  97   if (domain_start_index != base::string16::npos)
  98     *url_subdomain = url_host->substr(0, domain_start_index);
  99   if ((*url_subdomain == kWwwPrefix || url_subdomain->empty() ||
 100        url.SchemeIsFile())) {
 101     url_subdomain->clear();
 102   }
 103 }
 104
 105 #endif  // !defined(OS_ANDROID)
 106 }  // namespace
 107
 108 namespace url_formatter {
 109
 110 #if !defined(OS_ANDROID)
 111
 112 // TODO(pkasting): http://crbug.com/77883 This whole function gets
 113 // kerning/ligatures/etc. issues potentially wrong by assuming that the width of
 114 // a rendered string is always the sum of the widths of its substrings.  Also I
 115 // suspect it could be made simpler.
 116 base::string16 ElideUrl(const GURL& url,
 117                         const gfx::FontList& font_list,
 118                         float available_pixel_width,
 119                         const std::string& languages) {
 120   // Get a formatted string and corresponding parsing of the url.
 121   url::Parsed parsed;
 122   const base::string16 url_string = url_formatter::FormatUrl(
 123       url, languages, url_formatter::kFormatUrlOmitAll,
 124       net::UnescapeRule::SPACES, &parsed, nullptr, nullptr);
 125   if (available_pixel_width <= 0)
 126     return url_string;
 127
 128   // If non-standard, return plain eliding.
 129   if (!url.IsStandard())
 130     return gfx::ElideText(url_string, font_list, available_pixel_width,
 131                           gfx::ELIDE_TAIL);
 132
 133   // Now start eliding url_string to fit within available pixel width.
 134   // Fist pass - check to see whether entire url_string fits.
 135   const float pixel_width_url_string =
 136       gfx::GetStringWidthF(url_string, font_list);
 137   if (available_pixel_width >= pixel_width_url_string)
 138     return url_string;
 139
 140   // Get the path substring, including query and reference.
 141   const size_t path_start_index = parsed.path.begin;
 142   const size_t path_len = parsed.path.len;
 143   base::string16 url_path_query_etc = url_string.substr(path_start_index);
 144   base::string16 url_path = url_string.substr(path_start_index, path_len);
 145
 146   // Return general elided text if url minus the query fits.
 147   const base::string16 url_minus_query =
 148       url_string.substr(0, path_start_index + path_len);
 149   if (available_pixel_width >= gfx::GetStringWidthF(url_minus_query, font_list))
 150     return gfx::ElideText(url_string, font_list, available_pixel_width,
 151                           gfx::ELIDE_TAIL);
 152
 153   base::string16 url_host;
 154   base::string16 url_domain;
 155   base::string16 url_subdomain;
 156   SplitHost(url, &url_host, &url_domain, &url_subdomain);
 157
 158   // If this is a file type, the path is now defined as everything after ":".
 159   // For example, "C:/aa/aa/bb", the path is "/aa/bb/cc". Interesting, the
 160   // domain is now C: - this is a nice hack for eliding to work pleasantly.
 161   if (url.SchemeIsFile()) {
 162     // Split the path string using ":"
 163     const base::string16 kColon(1, ':');
 164     std::vector<base::string16> file_path_split = base::SplitString(
 165         url_path, kColon, base::TRIM_WHITESPACE, base::SPLIT_WANT_ALL);
 166     if (file_path_split.size() > 1) {  // File is of type "file:///C:/.."
 167       url_host.clear();
 168       url_domain.clear();
 169       url_subdomain.clear();
 170
 171       url_host = url_domain = file_path_split.at(0).substr(1) + kColon;
 172       url_path_query_etc = url_path = file_path_split.at(1);
 173     }
 174   }
 175
 176   // Second Pass - remove scheme - the rest fits.
 177   const float pixel_width_url_host = gfx::GetStringWidthF(url_host, font_list);
 178   const float pixel_width_url_path =
 179       gfx::GetStringWidthF(url_path_query_etc, font_list);
 180   if (available_pixel_width >= pixel_width_url_host + pixel_width_url_path)
 181     return url_host + url_path_query_etc;
 182
 183   // Third Pass: Subdomain, domain and entire path fits.
 184   const float pixel_width_url_domain =
 185       gfx::GetStringWidthF(url_domain, font_list);
 186   const float pixel_width_url_subdomain =
 187       gfx::GetStringWidthF(url_subdomain, font_list);
 188   if (available_pixel_width >=
 189       pixel_width_url_subdomain + pixel_width_url_domain + pixel_width_url_path)
 190     return url_subdomain + url_domain + url_path_query_etc;
 191
 192   // Query element.
 193   base::string16 url_query;
 194   const float kPixelWidthDotsTrailer =
 195       gfx::GetStringWidthF(base::string16(gfx::kEllipsisUTF16), font_list);
 196   if (parsed.query.is_nonempty()) {
 197     url_query = base::UTF8ToUTF16("?") + url_string.substr(parsed.query.begin);
 198     if (available_pixel_width >=
 199         (pixel_width_url_subdomain + pixel_width_url_domain +
 200          pixel_width_url_path - gfx::GetStringWidthF(url_query, font_list))) {
 201       return gfx::ElideText(url_subdomain + url_domain + url_path_query_etc,
 202                             font_list, available_pixel_width, gfx::ELIDE_TAIL);
 203     }
 204   }
 205
 206   // Parse url_path using '/'.
 207   std::vector<base::string16> url_path_elements =
 208       base::SplitString(url_path, base::string16(1, gfx::kForwardSlash),
 209                         base::TRIM_WHITESPACE, base::SPLIT_WANT_ALL);
 210
 211   // Get filename - note that for a path ending with /
 212   // such as www.google.com/intl/ads/, the file name is ads/.
 213   base::string16 url_filename(
 214       url_path_elements.empty() ? base::string16() : url_path_elements.back());
 215   size_t url_path_number_of_elements = url_path_elements.size();
 216   if (url_filename.empty() && (url_path_number_of_elements > 1)) {
 217     // Path ends with a '/'.
 218     --url_path_number_of_elements;
 219     url_filename =
 220         url_path_elements[url_path_number_of_elements - 1] + gfx::kForwardSlash;
 221   }
 222
 223   const size_t kMaxNumberOfUrlPathElementsAllowed = 1024;
 224   if (url_path_number_of_elements <= 1 ||
 225       url_path_number_of_elements > kMaxNumberOfUrlPathElementsAllowed) {
 226     // No path to elide, or too long of a path (could overflow in loop below)
 227     // Just elide this as a text string.
 228     return gfx::ElideText(url_subdomain + url_domain + url_path_query_etc,
 229                           font_list, available_pixel_width, gfx::ELIDE_TAIL);
 230   }
 231
 232   // Start eliding the path and replacing elements by ".../".
 233   const base::string16 kEllipsisAndSlash =
 234       base::string16(gfx::kEllipsisUTF16) + gfx::kForwardSlash;
 235   const float pixel_width_ellipsis_slash =
 236       gfx::GetStringWidthF(kEllipsisAndSlash, font_list);
 237
 238   // Check with both subdomain and domain.
 239   base::string16 elided_path = ElideComponentizedPath(
 240       url_subdomain + url_domain, url_path_elements, url_filename, url_query,
 241       font_list, available_pixel_width);
 242   if (!elided_path.empty())
 243     return elided_path;
 244
 245   // Check with only domain.
 246   // If a subdomain is present, add an ellipsis before domain.
 247   // This is added only if the subdomain pixel width is larger than
 248   // the pixel width of kEllipsis. Otherwise, subdomain remains,
 249   // which means that this case has been resolved earlier.
 250   base::string16 url_elided_domain = url_subdomain + url_domain;
 251   if (pixel_width_url_subdomain > kPixelWidthDotsTrailer) {
 252     if (!url_subdomain.empty())
 253       url_elided_domain = kEllipsisAndSlash[0] + url_domain;
 254     else
 255       url_elided_domain = url_domain;
 256
 257     elided_path = ElideComponentizedPath(url_elided_domain, url_path_elements,
 258                                          url_filename, url_query, font_list,
 259                                          available_pixel_width);
 260
 261     if (!elided_path.empty())
 262       return elided_path;
 263   }
 264
 265   // Return elided domain/.../filename anyway.
 266   base::string16 final_elided_url_string(url_elided_domain);
 267   const float url_elided_domain_width =
 268       gfx::GetStringWidthF(url_elided_domain, font_list);
 269
 270   // A hack to prevent trailing ".../...".
 271   if ((available_pixel_width - url_elided_domain_width) >
 272       pixel_width_ellipsis_slash + kPixelWidthDotsTrailer +
 273           gfx::GetStringWidthF(base::ASCIIToUTF16("UV"), font_list)) {
 274     final_elided_url_string += BuildPathFromComponents(
 275         base::string16(), url_path_elements, url_filename, 1);
 276   } else {
 277     final_elided_url_string += url_path;
 278   }
 279
 280   return gfx::ElideText(final_elided_url_string, font_list,
 281                         available_pixel_width, gfx::ELIDE_TAIL);
 282 }
 283
 284 base::string16 ElideHost(const GURL& url,
 285                          const gfx::FontList& font_list,
 286                          float available_pixel_width) {
 287   base::string16 url_host;
 288   base::string16 url_domain;
 289   base::string16 url_subdomain;
 290   SplitHost(url, &url_host, &url_domain, &url_subdomain);
 291
 292   const float pixel_width_url_host = gfx::GetStringWidthF(url_host, font_list);
 293   if (available_pixel_width >= pixel_width_url_host)
 294     return url_host;
 295
 296   if (url_subdomain.empty())
 297     return url_domain;
 298
 299   const float pixel_width_url_domain =
 300       gfx::GetStringWidthF(url_domain, font_list);
 301   float subdomain_width = available_pixel_width - pixel_width_url_domain;
 302   if (subdomain_width <= 0)
 303     return base::string16(gfx::kEllipsisUTF16) + kDot + url_domain;
 304
 305   const base::string16 elided_subdomain = gfx::ElideText(
 306       url_subdomain, font_list, subdomain_width, gfx::ELIDE_HEAD);
 307   return elided_subdomain + url_domain;
 308 }
 309
 310 #endif  // !defined(OS_ANDROID)
 311
 312 base::string16 FormatUrlForSecurityDisplay(const GURL& url,
 313                                            const std::string& languages) {
 314   if (!url.is_valid() || url.is_empty() || !url.IsStandard())
 315     return url_formatter::FormatUrl(url, languages);
 316
 317   const base::string16 colon(base::ASCIIToUTF16(":"));
 318   const base::string16 scheme_separator(
 319       base::ASCIIToUTF16(url::kStandardSchemeSeparator));
 320
 321   if (url.SchemeIsFile()) {
 322     return base::ASCIIToUTF16(url::kFileScheme) + scheme_separator +
 323            base::UTF8ToUTF16(url.path());
 324   }
 325
 326   if (url.SchemeIsFileSystem()) {
 327     const GURL* inner_url = url.inner_url();
 328     if (inner_url->SchemeIsFile()) {
 329       return base::ASCIIToUTF16(url::kFileSystemScheme) + colon +
 330              FormatUrlForSecurityDisplay(*inner_url, languages) +
 331              base::UTF8ToUTF16(url.path());
 332     }
 333     return base::ASCIIToUTF16(url::kFileSystemScheme) + colon +
 334            FormatUrlForSecurityDisplay(*inner_url, languages);
 335   }
 336
 337   const GURL origin = url.GetOrigin();
 338   const std::string& scheme = origin.scheme();
 339   const std::string& host = origin.host();
 340
 341   base::string16 result = base::UTF8ToUTF16(scheme);
 342   result += scheme_separator;
 343   result += base::UTF8ToUTF16(host);
 344
 345   const int port = origin.IntPort();
 346   const int default_port = url::DefaultPortForScheme(
 347       scheme.c_str(), static_cast<int>(scheme.length()));
 348   if (port != url::PORT_UNSPECIFIED && port != default_port)
 349     result += colon + base::UTF8ToUTF16(origin.port());
 350
 351   return result;
 352 }
 353 }  // namespace url_formatter