Upstream version 5.34.104.0
[platform/framework/web/crosswalk.git] / src / ui / gfx / text_elider.cc
index ed2a2b7..545b203 100644 (file)
 #include "base/strings/string_util.h"
 #include "base/strings/sys_string_conversions.h"
 #include "base/strings/utf_string_conversions.h"
-#include "net/base/escape.h"
-#include "net/base/net_util.h"
-#include "net/base/registry_controlled_domains/registry_controlled_domain.h"
 #include "third_party/icu/source/common/unicode/rbbi.h"
 #include "third_party/icu/source/common/unicode/uloc.h"
 #include "ui/gfx/font_list.h"
 #include "ui/gfx/text_utils.h"
-#include "url/gurl.h"
 
 using base::ASCIIToUTF16;
 using base::UTF8ToUTF16;
@@ -41,69 +37,24 @@ const char kEllipsis[] = "\xE2\x80\xA6";
 const base::char16 kEllipsisUTF16[] = { 0x2026, 0 };
 const base::char16 kForwardSlash = '/';
 
-namespace {
-
-
-// Build a path from the first |num_components| elements in |path_elements|.
-// Prepends |path_prefix|, appends |filename|, inserts ellipsis if appropriate.
-base::string16 BuildPathFromComponents(
-    const base::string16& path_prefix,
-    const std::vector<base::string16>& path_elements,
-    const base::string16& filename,
-    size_t num_components) {
-  // Add the initial elements of the path.
-  base::string16 path = path_prefix;
-
-  // Build path from first |num_components| elements.
-  for (size_t j = 0; j < num_components; ++j)
-    path += path_elements[j] + kForwardSlash;
-
-  // Add |filename|, ellipsis if necessary.
-  if (num_components != (path_elements.size() - 1))
-    path += base::string16(kEllipsisUTF16) + kForwardSlash;
-  path += filename;
-
-  return path;
-}
-
-// Takes a prefix (Domain, or Domain+subdomain) and a collection of path
-// components and elides if possible. Returns a string containing the longest
-// possible elided path, or an empty string if elision is not possible.
-base::string16 ElideComponentizedPath(
-    const base::string16& url_path_prefix,
-    const std::vector<base::string16>& url_path_elements,
-    const base::string16& url_filename,
-    const base::string16& url_query,
-    const FontList& font_list,
-    float available_pixel_width) {
-  const size_t url_path_number_of_elements = url_path_elements.size();
-
-  CHECK(url_path_number_of_elements);
-  for (size_t i = url_path_number_of_elements - 1; i > 0; --i) {
-    base::string16 elided_path = BuildPathFromComponents(url_path_prefix,
-        url_path_elements, url_filename, i);
-    if (available_pixel_width >= GetStringWidthF(elided_path, font_list))
-      return ElideText(elided_path + url_query, font_list,
-                       available_pixel_width, ELIDE_AT_END);
-  }
-
-  return base::string16();
-}
-
-}  // namespace
-
 StringSlicer::StringSlicer(const base::string16& text,
              const base::string16& ellipsis,
-             bool elide_in_middle)
+             bool elide_in_middle,
+             bool elide_at_beginning)
     : text_(text),
       ellipsis_(ellipsis),
-      elide_in_middle_(elide_in_middle) {
+      elide_in_middle_(elide_in_middle),
+      elide_at_beginning_(elide_at_beginning) {
 }
 
 base::string16 StringSlicer::CutString(size_t length, bool insert_ellipsis) {
   const base::string16 ellipsis_text = insert_ellipsis ? ellipsis_
                                                        : base::string16();
 
+  if (elide_at_beginning_)
+    return ellipsis_text +
+           text_.substr(FindValidBoundaryBefore(text_.length() - length));
+
   if (!elide_in_middle_)
     return text_.substr(0, FindValidBoundaryBefore(length)) + ellipsis_text;
 
@@ -191,202 +142,6 @@ base::string16 ElideEmail(const base::string16& email,
   return username + kAtSignUTF16 + domain;
 }
 
-// TODO(pkasting): http://crbug.com/77883 This whole function gets
-// kerning/ligatures/etc. issues potentially wrong by assuming that the width of
-// a rendered string is always the sum of the widths of its substrings.  Also I
-// suspect it could be made simpler.
-base::string16 ElideUrl(const GURL& url,
-                        const FontList& font_list,
-                        float available_pixel_width,
-                        const std::string& languages) {
-  // Get a formatted string and corresponding parsing of the url.
-  url_parse::Parsed parsed;
-  const base::string16 url_string =
-      net::FormatUrl(url, languages, net::kFormatUrlOmitAll,
-                     net::UnescapeRule::SPACES, &parsed, NULL, NULL);
-  if (available_pixel_width <= 0)
-    return url_string;
-
-  // If non-standard, return plain eliding.
-  if (!url.IsStandard())
-    return ElideText(url_string, font_list, available_pixel_width,
-                     ELIDE_AT_END);
-
-  // Now start eliding url_string to fit within available pixel width.
-  // Fist pass - check to see whether entire url_string fits.
-  const float pixel_width_url_string = GetStringWidthF(url_string, font_list);
-  if (available_pixel_width >= pixel_width_url_string)
-    return url_string;
-
-  // Get the path substring, including query and reference.
-  const size_t path_start_index = parsed.path.begin;
-  const size_t path_len = parsed.path.len;
-  base::string16 url_path_query_etc = url_string.substr(path_start_index);
-  base::string16 url_path = url_string.substr(path_start_index, path_len);
-
-  // Return general elided text if url minus the query fits.
-  const base::string16 url_minus_query =
-      url_string.substr(0, path_start_index + path_len);
-  if (available_pixel_width >= GetStringWidthF(url_minus_query, font_list))
-    return ElideText(url_string, font_list, available_pixel_width,
-                     ELIDE_AT_END);
-
-  // Get Host.
-  base::string16 url_host = UTF8ToUTF16(url.host());
-
-  // Get domain and registry information from the URL.
-  base::string16 url_domain = UTF8ToUTF16(
-      net::registry_controlled_domains::GetDomainAndRegistry(
-          url, net::registry_controlled_domains::EXCLUDE_PRIVATE_REGISTRIES));
-  if (url_domain.empty())
-    url_domain = url_host;
-
-  // Add port if required.
-  if (!url.port().empty()) {
-    url_host += UTF8ToUTF16(":" + url.port());
-    url_domain += UTF8ToUTF16(":" + url.port());
-  }
-
-  // Get sub domain.
-  base::string16 url_subdomain;
-  const size_t domain_start_index = url_host.find(url_domain);
-  if (domain_start_index != base::string16::npos)
-    url_subdomain = url_host.substr(0, domain_start_index);
-  const base::string16 kWwwPrefix = UTF8ToUTF16("www.");
-  if ((url_subdomain == kWwwPrefix || url_subdomain.empty() ||
-      url.SchemeIsFile())) {
-    url_subdomain.clear();
-  }
-
-  // If this is a file type, the path is now defined as everything after ":".
-  // For example, "C:/aa/aa/bb", the path is "/aa/bb/cc". Interesting, the
-  // domain is now C: - this is a nice hack for eliding to work pleasantly.
-  if (url.SchemeIsFile()) {
-    // Split the path string using ":"
-    std::vector<base::string16> file_path_split;
-    base::SplitString(url_path, ':', &file_path_split);
-    if (file_path_split.size() > 1) {  // File is of type "file:///C:/.."
-      url_host.clear();
-      url_domain.clear();
-      url_subdomain.clear();
-
-      const base::string16 kColon = UTF8ToUTF16(":");
-      url_host = url_domain = file_path_split.at(0).substr(1) + kColon;
-      url_path_query_etc = url_path = file_path_split.at(1);
-    }
-  }
-
-  // Second Pass - remove scheme - the rest fits.
-  const float pixel_width_url_host = GetStringWidthF(url_host, font_list);
-  const float pixel_width_url_path = GetStringWidthF(url_path_query_etc,
-                                                     font_list);
-  if (available_pixel_width >=
-      pixel_width_url_host + pixel_width_url_path)
-    return url_host + url_path_query_etc;
-
-  // Third Pass: Subdomain, domain and entire path fits.
-  const float pixel_width_url_domain = GetStringWidthF(url_domain, font_list);
-  const float pixel_width_url_subdomain =
-      GetStringWidthF(url_subdomain, font_list);
-  if (available_pixel_width >=
-      pixel_width_url_subdomain + pixel_width_url_domain +
-      pixel_width_url_path)
-    return url_subdomain + url_domain + url_path_query_etc;
-
-  // Query element.
-  base::string16 url_query;
-  const float kPixelWidthDotsTrailer = GetStringWidthF(
-      base::string16(kEllipsisUTF16), font_list);
-  if (parsed.query.is_nonempty()) {
-    url_query = UTF8ToUTF16("?") + url_string.substr(parsed.query.begin);
-    if (available_pixel_width >=
-        (pixel_width_url_subdomain + pixel_width_url_domain +
-         pixel_width_url_path - GetStringWidthF(url_query, font_list))) {
-      return ElideText(url_subdomain + url_domain + url_path_query_etc,
-                       font_list, available_pixel_width, ELIDE_AT_END);
-    }
-  }
-
-  // Parse url_path using '/'.
-  std::vector<base::string16> url_path_elements;
-  base::SplitString(url_path, kForwardSlash, &url_path_elements);
-
-  // Get filename - note that for a path ending with /
-  // such as www.google.com/intl/ads/, the file name is ads/.
-  size_t url_path_number_of_elements = url_path_elements.size();
-  DCHECK(url_path_number_of_elements != 0);
-  base::string16 url_filename;
-  if ((url_path_elements.at(url_path_number_of_elements - 1)).length() > 0) {
-    url_filename = *(url_path_elements.end() - 1);
-  } else if (url_path_number_of_elements > 1) {  // Path ends with a '/'.
-    url_filename = url_path_elements.at(url_path_number_of_elements - 2) +
-        kForwardSlash;
-    url_path_number_of_elements--;
-  }
-  DCHECK(url_path_number_of_elements != 0);
-
-  const size_t kMaxNumberOfUrlPathElementsAllowed = 1024;
-  if (url_path_number_of_elements <= 1 ||
-      url_path_number_of_elements > kMaxNumberOfUrlPathElementsAllowed) {
-    // No path to elide, or too long of a path (could overflow in loop below)
-    // Just elide this as a text string.
-    return ElideText(url_subdomain + url_domain + url_path_query_etc, font_list,
-                     available_pixel_width, ELIDE_AT_END);
-  }
-
-  // Start eliding the path and replacing elements by ".../".
-  const base::string16 kEllipsisAndSlash =
-      base::string16(kEllipsisUTF16) + kForwardSlash;
-  const float pixel_width_ellipsis_slash =
-      GetStringWidthF(kEllipsisAndSlash, font_list);
-
-  // Check with both subdomain and domain.
-  base::string16 elided_path =
-      ElideComponentizedPath(url_subdomain + url_domain, url_path_elements,
-                             url_filename, url_query, font_list,
-                             available_pixel_width);
-  if (!elided_path.empty())
-    return elided_path;
-
-  // Check with only domain.
-  // If a subdomain is present, add an ellipsis before domain.
-  // This is added only if the subdomain pixel width is larger than
-  // the pixel width of kEllipsis. Otherwise, subdomain remains,
-  // which means that this case has been resolved earlier.
-  base::string16 url_elided_domain = url_subdomain + url_domain;
-  if (pixel_width_url_subdomain > kPixelWidthDotsTrailer) {
-    if (!url_subdomain.empty())
-      url_elided_domain = kEllipsisAndSlash[0] + url_domain;
-    else
-      url_elided_domain = url_domain;
-
-    elided_path = ElideComponentizedPath(url_elided_domain, url_path_elements,
-                                         url_filename, url_query, font_list,
-                                         available_pixel_width);
-
-    if (!elided_path.empty())
-      return elided_path;
-  }
-
-  // Return elided domain/.../filename anyway.
-  base::string16 final_elided_url_string(url_elided_domain);
-  const float url_elided_domain_width = GetStringWidthF(url_elided_domain,
-                                                        font_list);
-
-  // A hack to prevent trailing ".../...".
-  if ((available_pixel_width - url_elided_domain_width) >
-      pixel_width_ellipsis_slash + kPixelWidthDotsTrailer +
-      GetStringWidthF(ASCIIToUTF16("UV"), font_list)) {
-    final_elided_url_string += BuildPathFromComponents(base::string16(),
-        url_path_elements, url_filename, 1);
-  } else {
-    final_elided_url_string += url_path;
-  }
-
-  return ElideText(final_elided_url_string, font_list, available_pixel_width,
-                   ELIDE_AT_END);
-}
-
 base::string16 ElideFilename(const base::FilePath& filename,
                              const FontList& font_list,
                              float available_pixel_width) {
@@ -445,10 +200,11 @@ base::string16 ElideText(const base::string16& text,
 
   const float current_text_pixel_width = GetStringWidthF(text, font_list);
   const bool elide_in_middle = (elide_behavior == ELIDE_IN_MIDDLE);
+  const bool elide_at_beginning = (elide_behavior == ELIDE_AT_BEGINNING);
   const bool insert_ellipsis = (elide_behavior != TRUNCATE_AT_END);
 
   const base::string16 ellipsis = base::string16(kEllipsisUTF16);
-  StringSlicer slicer(text, ellipsis, elide_in_middle);
+  StringSlicer slicer(text, ellipsis, elide_in_middle, elide_at_beginning);
 
   // Pango will return 0 width for absurdly long strings. Cut the string in
   // half and try again.
@@ -498,76 +254,6 @@ base::string16 ElideText(const base::string16& text,
   return slicer.CutString(guess, insert_ellipsis);
 }
 
-SortedDisplayURL::SortedDisplayURL(const GURL& url,
-                                   const std::string& languages) {
-  net::AppendFormattedHost(url, languages, &sort_host_);
-  base::string16 host_minus_www = net::StripWWW(sort_host_);
-  url_parse::Parsed parsed;
-  display_url_ =
-      net::FormatUrl(url, languages, net::kFormatUrlOmitAll,
-                     net::UnescapeRule::SPACES, &parsed, &prefix_end_, NULL);
-  if (sort_host_.length() > host_minus_www.length()) {
-    prefix_end_ += sort_host_.length() - host_minus_www.length();
-    sort_host_.swap(host_minus_www);
-  }
-}
-
-SortedDisplayURL::SortedDisplayURL() : prefix_end_(0) {
-}
-
-SortedDisplayURL::~SortedDisplayURL() {
-}
-
-int SortedDisplayURL::Compare(const SortedDisplayURL& other,
-                              icu::Collator* collator) const {
-  // Compare on hosts first. The host won't contain 'www.'.
-  UErrorCode compare_status = U_ZERO_ERROR;
-  UCollationResult host_compare_result = collator->compare(
-      static_cast<const UChar*>(sort_host_.c_str()),
-      static_cast<int>(sort_host_.length()),
-      static_cast<const UChar*>(other.sort_host_.c_str()),
-      static_cast<int>(other.sort_host_.length()),
-      compare_status);
-  DCHECK(U_SUCCESS(compare_status));
-  if (host_compare_result != 0)
-    return host_compare_result;
-
-  // Hosts match, compare on the portion of the url after the host.
-  base::string16 path = this->AfterHost();
-  base::string16 o_path = other.AfterHost();
-  compare_status = U_ZERO_ERROR;
-  UCollationResult path_compare_result = collator->compare(
-      static_cast<const UChar*>(path.c_str()),
-      static_cast<int>(path.length()),
-      static_cast<const UChar*>(o_path.c_str()),
-      static_cast<int>(o_path.length()),
-      compare_status);
-  DCHECK(U_SUCCESS(compare_status));
-  if (path_compare_result != 0)
-    return path_compare_result;
-
-  // Hosts and paths match, compare on the complete url. This'll push the www.
-  // ones to the end.
-  compare_status = U_ZERO_ERROR;
-  UCollationResult display_url_compare_result = collator->compare(
-      static_cast<const UChar*>(display_url_.c_str()),
-      static_cast<int>(display_url_.length()),
-      static_cast<const UChar*>(other.display_url_.c_str()),
-      static_cast<int>(other.display_url_.length()),
-      compare_status);
-  DCHECK(U_SUCCESS(compare_status));
-  return display_url_compare_result;
-}
-
-base::string16 SortedDisplayURL::AfterHost() const {
-  const size_t slash_index = display_url_.find(sort_host_, prefix_end_);
-  if (slash_index == base::string16::npos) {
-    NOTREACHED();
-    return base::string16();
-  }
-  return display_url_.substr(slash_index + sort_host_.length());
-}
-
 bool ElideString(const base::string16& input, int max_len,
                  base::string16* output) {
   DCHECK_GE(max_len, 0);