Hi,
The second prerelease of JPluck 0.9 is available: This is basically a refresh with many bugfixes and some new features. Thanks to everyone for their feedback. COOKIES Cookie support was seriously broken in pre1 because of several reasons. It has been improved greatly in this release. - In pre1 you could not select a Mozilla/Netscape cookies.txt file as the cookie source. This has been fixed. - Fixed bugs with parsing cookies.txt - Maintaining session cookies now actually works. Some web applications that offer personalized content insist on maintaining a session cookie to function. (One example is http://www.telerama.fr/pda/pda/index.asp , you need to append a special query string for this to work.) - It is now possible to specify "None" as the cookie source. This allows you to handle session cookies only, without sending existing cookies from the browser. A nice side-effect of the cookie support is that most sites that use form-based logins can now be plucked. This is possible because these sites usually offer the option to store a cookie on your machine after a succesful login, so that you don't have to log in the next time. JPluck simply sends the cookie from the browser and mimics this behavior. I only tried this on NYTimes.com, but I'm pretty sure it will work on other sites, like My Yahoo, as well. IMAGES - Image handling has been improved greatly. JPluck now queries ImageIO if an appropriate codec has been installed for the given image MIME type (image/gif, image/jpeg, image/png as standard). Determining it this way also lets you install decoders to read additional image formats. Sun has released an ImageIO extension with several new readers, among which BMP, WBMP, TIFF, JPEG2000. I haven't tried out these codecs yet but they should work. - Images are now resized to fill the maximum available space under the 60K limit. This algorithm honors the bit depth and the maximum dimension specified in the Conversion Settings. - Tip: use 8bpp for color hires displays. 16bpp only allows for a 173x173 area while 8bpp allows for 244x244. You can't have full-screen color images on a hires device due to the 60K limit. - Tip: enable linear resizing for better thumbnail quality. HTML/XML parsing - Anchors now start a new paragraph with 0 spacing. This improvement allows links to anchors inside paragraphs to work without changing the document visually. (Thanks to Mike for the hint.) One problematic page in this regard is http://members.aol.com/pocketpress/. However, the current issue of that site has invalid anchors to begin with. Try it out in a browser, you'll see. Anchors did work well on an older issue a few days ago. - Whitespace in <pre> tags is now retained. Previously, whitespace was collapsed into a single space(like normal HTML sections). - Content handling has been revised so that JPluck now attempts to parse content as HTML/XML by default (if it is not an image). This way a server does not necessarily have to return a correct content-type header. Many servers send a fallback content-type for XML documents because their MIME configuration hasn't been updated. XSL XSL support has been much improved under the hood. The XSL parser now scans the directory where the JXL is stored and the "xsl" subdirectory of the JPluck home directory for XSL files. The XSL files can in turn contain a special JPluck processing instruction with metadata that describes which URLs or which XML doctype this particular stylesheet applies to. This allows JPluck to discover and apply stylesheets automatically without the user having to specify them explicitly on a per-Document basis. The distribution includes a pre-configured RSS stylesheet. If you try to convert an RSS feed, this stylesheet will kick in and internally transform the RSS to an HTML document which in turn is converted to a Plucker text record. I tested the RSS stylesheet on a few feeds and they seemed to work alright. One place to find RSS feeds is http://www.syndic8.com/. The RSS stylesheet was slapped together quickly, there's room for improvement, particularly in displaying item dates. The stylesheets for Common Dreams and The Register are also bundled with the distribution. They contain processing instructions so that they are applied automatically. Both sites are now part of the showcase. A GUI for binding XSL stylesheets to documents will be included in the next prerelease. MISCELLANEOUS - Added cache tab to Options dialog. This allows you to view the current size of the cache and clear its contents. There may be more options to control cache usage in the next prerelease. - The User Agent field now allows you to select from a few common defaults (AvantGo, Netscape, MSIE). - "Stay on host" and "Stay below starting directory" fields are now configurable in the default document settings. - When starting the application. - Bugs with JPluck C should be cleared. Bill Nalen reported that JPluck C also works with his HotSync conduit but I haven't tried this yet. NOTE TO USERS OF 0.9 PRE1 If you try to open a JXL saved by 0.9 pre1, with XSL transformations specified, you will get the error: Element "document" does not allow "transform" here. You have to edit the file in a text editor and replace the parts "<transform>" and "</transform>" with "<transformation>" and "</transformation>" respectively. This error occurs because I changed the JXL DTD. This being a prerelease there will be bugs, but overall I think it's a great improvement. The next prerelease will add some more useful features like scheduling and a system tray option on Windows. With the tray icon you can close and reopen JPluck quickly without incurring the Java startup overhead. I didn't have the time to build a proper upgrade package, so you'll have to download the entire distribution again. I'll try to provide an upgrade with the next prerelease. I couldn't upload to SourceForge's FTP server, so I've made only the binaries available from the web server. I'll try uploading again tomorrow. You can get it here: http://jpluck.sourceforge.net/prerelease/ or here http://home.planet.nl/~fridael/jpluck/prerelease/ My connection to SourceForge is pretty slow right now so I mirrored it at my crappy ISP. Let me know what you think. Thanks -Laurens --- http://jpluck.sourceforge.net/ JPluck - open-source Plucker document creation toolkit. NOTE: Please report JPluck bugs and issues directly to me by e-mail. Do not discuss JPluck technical issues on the Plucker mailing lists. _______________________________________________ plucker-list mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-list
