Remove "<HR>" tags to import Firefox bookmarks correctly.

chrome/utility/importer/bookmark_html_reader.cc

Index: chrome/utility/importer/bookmark_html_reader.cc
diff --git a/chrome/utility/importer/bookmark_html_reader.cc b/chrome/utility/importer/bookmark_html_reader.cc
index 5d34d7b191a817ce63ad6cf27df6cf68d6a56ce1..e3c687f17423ffc03d85da2642b16429421601b5 100644
--- a/chrome/utility/importer/bookmark_html_reader.cc
+++ b/chrome/utility/importer/bookmark_html_reader.cc
@@ -112,6 +112,13 @@ void ImportBookmarksFile(
     std::string line;
     TrimString(lines[i], " ", &line);
 
+    // Remove "<HR>" if |line| starts with it. See http://crbug.com/257474.
+    static const char kHrTag[] = "<HR>";
+    while (StartsWithASCII(line, kHrTag, false)) {

[zhchbin, 2013/08/10 06:06:02]

Note this, I have change it from "if" to "while". Actually speaking, "<HR>" is
the "separator" in the firefox. I have upload a video[0] to show it. My test
case file "firefox23.html" is also exported with these bookmarks.

[0]:
https://docs.google.com/file/d/0B_7w63Sy3AuycWlfTmhvanZLUEE/edit?usp=sharing 

[tfarina, 2013/08/10 14:30:51]

Could also note in the comments that we are removing/skipping the <HR> tag
because it's the separator in Firefox that Chrome does not support?

[zhchbin, 2013/08/10 14:45:39]

Done.

+      line.erase(0, arraysize(kHrTag) - 1);
+      TrimString(line, " ", &line);
+    }
+
     // Get the encoding of the bookmark file.
     if (internal::ParseCharsetFromLine(line, &charset))
       continue;