chrome/browser/ui/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 "chrome/browser/ui/elide_url.h"
   6
   7 #include "base/strings/string_split.h"
   8 #include "base/strings/utf_string_conversions.h"
   9 #include "net/base/escape.h"
  10 #include "net/base/net_util.h"
  11 #include "net/base/registry_controlled_domains/registry_controlled_domain.h"
  12 #include "ui/gfx/text_elider.h"
  13 #include "ui/gfx/text_utils.h"
  14 #include "url/gurl.h"
  15
  16 using base::UTF8ToUTF16;
  17 using gfx::ElideText;
  18 using gfx::GetStringWidthF;
  19 using gfx::kEllipsisUTF16;
  20 using gfx::kForwardSlash;
  21
  22 namespace {
  23
  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] + kForwardSlash;
  39
  40   // Add |filename|, ellipsis if necessary.
  41   if (num_components != (path_elements.size() - 1))
  42     path += base::string16(kEllipsisUTF16) + 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(url_path_prefix,
  63         url_path_elements, url_filename, i);
  64     if (available_pixel_width >= GetStringWidthF(elided_path, font_list))
  65       return ElideText(elided_path + url_query, font_list,
  66                        available_pixel_width, gfx::ELIDE_AT_END);
  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 = UTF8ToUTF16(url.host());
  80
  81   // Get domain and registry information from the URL.
  82   *url_domain = UTF8ToUTF16(
  83       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 += UTF8ToUTF16(":" + url.port());
  91     *url_domain += 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 = 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 }  // namespace
 106
 107 // TODO(pkasting): http://crbug.com/77883 This whole function gets
 108 // kerning/ligatures/etc. issues potentially wrong by assuming that the width of
 109 // a rendered string is always the sum of the widths of its substrings.  Also I
 110 // suspect it could be made simpler.
 111 base::string16 ElideUrl(const GURL& url,
 112                         const gfx::FontList& font_list,
 113                         float available_pixel_width,
 114                         const std::string& languages) {
 115   // Get a formatted string and corresponding parsing of the url.
 116   url_parse::Parsed parsed;
 117   const base::string16 url_string =
 118       net::FormatUrl(url, languages, net::kFormatUrlOmitAll,
 119                      net::UnescapeRule::SPACES, &parsed, NULL, NULL);
 120   if (available_pixel_width <= 0)
 121     return url_string;
 122
 123   // If non-standard, return plain eliding.
 124   if (!url.IsStandard())
 125     return ElideText(url_string, font_list, available_pixel_width,
 126                      gfx::ELIDE_AT_END);
 127
 128   // Now start eliding url_string to fit within available pixel width.
 129   // Fist pass - check to see whether entire url_string fits.
 130   const float pixel_width_url_string = GetStringWidthF(url_string, font_list);
 131   if (available_pixel_width >= pixel_width_url_string)
 132     return url_string;
 133
 134   // Get the path substring, including query and reference.
 135   const size_t path_start_index = parsed.path.begin;
 136   const size_t path_len = parsed.path.len;
 137   base::string16 url_path_query_etc = url_string.substr(path_start_index);
 138   base::string16 url_path = url_string.substr(path_start_index, path_len);
 139
 140   // Return general elided text if url minus the query fits.
 141   const base::string16 url_minus_query =
 142       url_string.substr(0, path_start_index + path_len);
 143   if (available_pixel_width >= GetStringWidthF(url_minus_query, font_list))
 144     return ElideText(url_string, font_list, available_pixel_width,
 145                      gfx::ELIDE_AT_END);
 146
 147   base::string16 url_host;
 148   base::string16 url_domain;
 149   base::string16 url_subdomain;
 150   SplitHost(url, &url_host, &url_domain, &url_subdomain);
 151
 152   // If this is a file type, the path is now defined as everything after ":".
 153   // For example, "C:/aa/aa/bb", the path is "/aa/bb/cc". Interesting, the
 154   // domain is now C: - this is a nice hack for eliding to work pleasantly.
 155   if (url.SchemeIsFile()) {
 156     // Split the path string using ":"
 157     std::vector<base::string16> file_path_split;
 158     base::SplitString(url_path, ':', &file_path_split);
 159     if (file_path_split.size() > 1) {  // File is of type "file:///C:/.."
 160       url_host.clear();
 161       url_domain.clear();
 162       url_subdomain.clear();
 163
 164       const base::string16 kColon = UTF8ToUTF16(":");
 165       url_host = url_domain = file_path_split.at(0).substr(1) + kColon;
 166       url_path_query_etc = url_path = file_path_split.at(1);
 167     }
 168   }
 169
 170   // Second Pass - remove scheme - the rest fits.
 171   const float pixel_width_url_host = GetStringWidthF(url_host, font_list);
 172   const float pixel_width_url_path = GetStringWidthF(url_path_query_etc,
 173                                                      font_list);
 174   if (available_pixel_width >=
 175       pixel_width_url_host + pixel_width_url_path)
 176     return url_host + url_path_query_etc;
 177
 178   // Third Pass: Subdomain, domain and entire path fits.
 179   const float pixel_width_url_domain = GetStringWidthF(url_domain, font_list);
 180   const float pixel_width_url_subdomain =
 181       GetStringWidthF(url_subdomain, font_list);
 182   if (available_pixel_width >=
 183       pixel_width_url_subdomain + pixel_width_url_domain +
 184       pixel_width_url_path)
 185     return url_subdomain + url_domain + url_path_query_etc;
 186
 187   // Query element.
 188   base::string16 url_query;
 189   const float kPixelWidthDotsTrailer = GetStringWidthF(
 190       base::string16(kEllipsisUTF16), font_list);
 191   if (parsed.query.is_nonempty()) {
 192     url_query = UTF8ToUTF16("?") + url_string.substr(parsed.query.begin);
 193     if (available_pixel_width >=
 194         (pixel_width_url_subdomain + pixel_width_url_domain +
 195          pixel_width_url_path - GetStringWidthF(url_query, font_list))) {
 196       return ElideText(url_subdomain + url_domain + url_path_query_etc,
 197                        font_list, available_pixel_width, gfx::ELIDE_AT_END);
 198     }
 199   }
 200
 201   // Parse url_path using '/'.
 202   std::vector<base::string16> url_path_elements;
 203   base::SplitString(url_path, kForwardSlash, &url_path_elements);
 204
 205   // Get filename - note that for a path ending with /
 206   // such as www.google.com/intl/ads/, the file name is ads/.
 207   size_t url_path_number_of_elements = url_path_elements.size();
 208   DCHECK(url_path_number_of_elements != 0);
 209   base::string16 url_filename;
 210   if ((url_path_elements.at(url_path_number_of_elements - 1)).length() > 0) {
 211     url_filename = *(url_path_elements.end() - 1);
 212   } else if (url_path_number_of_elements > 1) {  // Path ends with a '/'.
 213     url_filename = url_path_elements.at(url_path_number_of_elements - 2) +
 214         kForwardSlash;
 215     url_path_number_of_elements--;
 216   }
 217   DCHECK(url_path_number_of_elements != 0);
 218
 219   const size_t kMaxNumberOfUrlPathElementsAllowed = 1024;
 220   if (url_path_number_of_elements <= 1 ||
 221       url_path_number_of_elements > kMaxNumberOfUrlPathElementsAllowed) {
 222     // No path to elide, or too long of a path (could overflow in loop below)
 223     // Just elide this as a text string.
 224     return ElideText(url_subdomain + url_domain + url_path_query_etc, font_list,
 225                      available_pixel_width, gfx::ELIDE_AT_END);
 226   }
 227
 228   // Start eliding the path and replacing elements by ".../".
 229   const base::string16 kEllipsisAndSlash =
 230       base::string16(kEllipsisUTF16) + kForwardSlash;
 231   const float pixel_width_ellipsis_slash =
 232       GetStringWidthF(kEllipsisAndSlash, font_list);
 233
 234   // Check with both subdomain and domain.
 235   base::string16 elided_path =
 236       ElideComponentizedPath(url_subdomain + url_domain, url_path_elements,
 237                              url_filename, url_query, font_list,
 238                              available_pixel_width);
 239   if (!elided_path.empty())
 240     return elided_path;
 241
 242   // Check with only domain.
 243   // If a subdomain is present, add an ellipsis before domain.
 244   // This is added only if the subdomain pixel width is larger than
 245   // the pixel width of kEllipsis. Otherwise, subdomain remains,
 246   // which means that this case has been resolved earlier.
 247   base::string16 url_elided_domain = url_subdomain + url_domain;
 248   if (pixel_width_url_subdomain > kPixelWidthDotsTrailer) {
 249     if (!url_subdomain.empty())
 250       url_elided_domain = kEllipsisAndSlash[0] + url_domain;
 251     else
 252       url_elided_domain = url_domain;
 253
 254     elided_path = ElideComponentizedPath(url_elided_domain, url_path_elements,
 255                                          url_filename, url_query, font_list,
 256                                          available_pixel_width);
 257
 258     if (!elided_path.empty())
 259       return elided_path;
 260   }
 261
 262   // Return elided domain/.../filename anyway.
 263   base::string16 final_elided_url_string(url_elided_domain);
 264   const float url_elided_domain_width = GetStringWidthF(url_elided_domain,
 265                                                         font_list);
 266
 267   // A hack to prevent trailing ".../...".
 268   if ((available_pixel_width - url_elided_domain_width) >
 269       pixel_width_ellipsis_slash + kPixelWidthDotsTrailer +
 270       GetStringWidthF(base::ASCIIToUTF16("UV"), font_list)) {
 271     final_elided_url_string += BuildPathFromComponents(base::string16(),
 272         url_path_elements, url_filename, 1);
 273   } else {
 274     final_elided_url_string += url_path;
 275   }
 276
 277   return ElideText(final_elided_url_string, font_list, available_pixel_width,
 278                    gfx::ELIDE_AT_END);
 279 }
 280
 281 base::string16 ElideHost(const GURL& url,
 282                          const gfx::FontList& font_list,
 283                          float available_pixel_width) {
 284   base::string16 url_host;
 285   base::string16 url_domain;
 286   base::string16 url_subdomain;
 287   SplitHost(url, &url_host, &url_domain, &url_subdomain);
 288
 289   const float pixel_width_url_host = GetStringWidthF(url_host, font_list);
 290   if (available_pixel_width >= pixel_width_url_host)
 291     return url_host;
 292
 293   if (url_subdomain.empty())
 294     return url_domain;
 295
 296   const float pixel_width_url_domain = GetStringWidthF(url_domain, font_list);
 297   float subdomain_width = available_pixel_width - pixel_width_url_domain;
 298   if (subdomain_width <= 0)
 299     return base::string16(kEllipsisUTF16) + kDot + url_domain;
 300
 301   base::string16 elided_subdomain = ElideText(
 302       url_subdomain, font_list, subdomain_width, gfx::ELIDE_AT_BEGINNING);
 303   return elided_subdomain + url_domain;
 304 }