For a change here a success story, not a question or cry for help.
 
I have successfully installed and used Cocoon 2.0.1 in batch mode for generating our static web site http://www.infodesigner.biz (I am afraid it is in German, sorry ;-).
 
Effort: one week with no prior knowledge of Cocoon and only slight knowledge of XSL. (From this two days were spent for translating the original HTML stuff to small XML pieces - boring :-((. Could be faster, I know. But see the hints below to overtake me.
 
If you are interested in our small (not as sophisticated as the one delivered with Cocoon 2.0.1) example you can download it from
  http://www.infodesigner.biz/de/about/website.zip (ca. 500 kb). You will get our website and all you need to build it. You will find two different style sheets (XSL), some XML code, some plain HTML pages, and css, gif, jpg, bmp. 
 
My platform: Windows XP.
 
Hints (see also the README in the zip - again in German, so here the translation and some more remarks):
- Use the source distribution Cocoon-2.0.1-src.zip - the binary distribution seems to be useless for us "standalonies".
- You have to build cocoon.jar with the command:
  build all
- Then you have to copy the jars so that run.bat finds them:
   copy %COCOON_HOME%\build\cocoon\cocoon.jar %COCOON_HOME%\lib\
   copy %COCOON_HOME%\lib\core\* %COCOON_HOME%\lib\
   copy %COCOON_HOME%\lib\optional\* %COCOON_HOME%\lib\
- If you do not start Cocoon from %COCOON_HOME% (and it is not on your PATH) you have to change run.bat so that appendcp.bat can be found (a simple %COCOON_HOME%\appendcp.bat is fine).
- Try run -h to get the command line parameters. Note that you have to give a LogKit.xconf configuration file via -k if you want a finer control of the logging information (since I use log4j and not LogKit here ends my advice). The redirection of log output to a file via -l does not work - even if the build.xml does assume that.
- Do not rely on the output of the LogKit - there are no ERRORs or WARNings when files could not be found or the link serializer could not find links for recursive build. You may see Broken Links - or you may not when a file has been translated but the links were not extracted. So check your web site for broken links. I could not get hold of useful information in the mass of the log output.
- I could not figure out if the link serializer can extract links from *.css stylesheets.  So files mentioned in css-stylesheets are not recognized.
- Do not change the name "link" of the view for the link serializer - it is hardwired in the code!
- Lacking a smart XML editor I used the Internet Explorer 6.0 for checking the syntax of my XML and XSL code. It shows errors at the end of the file. Dumb but useful.
 
If you have questions you can reach me via [EMAIL PROTECTED].
 
Regards,
 
Gerald
 
InfoDesign-OpenSourceDeveloper GmbH
 

Reply via email to