Bookmark utility part 5 - Importing bookmarks

Microsoft's browser features the simplest file format, because each link is stored in a separate file. The tree structure of the favorites is realized by the actual directory structure on your disk. The name of the file, with the extension .url is the title of the bookmarked page. Inside each file, you will find that it always has "[InternetShortcut]" on the fist line. The URL of the page is in clear text after the line which start with "URL=".

Now what you have to do, is to start your program in the base directory of the favorites. In Windows 2000, this would be C:\Documents and Settings\<username>\Favorites. You read all the files with the extension .url, and collect their name and URL. Then you gather all the subdirectories of Favorites and repeat this algorithm for each of them.

Hint: For file and directory retrieving and traversing, see the last questions of Week 50 - 2001. For storing tree structures in Java, see last week's exercises.

IEBookmarks.java

Opera may not be the most famous browser, however, it is the best alternative browser on the market, available on eight different operating systems including Windows, Linux and Mac OS. If you are not familiar with this browser, you might want to download and check it out for free from their website.

Opera also features a fairly simple way to store its bookmarks. They are stored in only one file, "opera6.adr" in the preference or user directory, rather than several files and directories, but the it is quite straight forward to parse: Each entity, URL or folder, is encoded in four or six lines. The first one stating the type "#FOLDER" or "#URL" on a single line. The next line lists the name of the entity after "NAME=". If it is a folder, this is all you need, but if you it is an URL, you need to read the next line as well, starting with "URL=".

The file is encoded with pre order traversing, which means that a folder name is stated first, and then follows all the entities, including sub folders (which are also in pre order) of that folder. This means that you first need to create a root folder in you program. If a URL is read, you just add it to the folder. If a new folder is read, this would be a sub folder of the current, recursively. When you read a single "-" (hyphen) on a line, this means that you have reached the end of a folder, and you need to step up to the previous level of the tree hierarchy.

OperaBookmarks.java

The Navigator features the most complex encoding for their bookmarks. They are presented in a HTML format, so you might think it would be easy to parse, but it turns out that it is a few things to look after here.

As with the other browsers, there are two entities: URLs and folders. Folders are within <H3>...</H3> tags, but make sure you only read the text between the actual tags and not what is inside them. The sub entities of a folder is surrounded by a <DT> and matching </DT> at the bottom. Finally, an URL is encoded as a link; e.g.: <A HREF="http://...">Page title...</A>. There is only one link on each line.

NetscapeBookmarks.java

Finally, this week, you could combine all the above parser into one program. And of course your program would not only read the different formats; the user would also have the choice to save bookmarks in the desired format.