ribit 0.3.0

data/bin/ribit-setup

@@ -0,0 +1,153 @@
+#!/usr/bin/ruby
+
+#--
+#    Copyright (C) 2004-2005 Tero Vuorelaorela
+#
+#    This program is free software; you can redistribute it and/or modify it
+#    under the terms of the GNU General Public License as published by the Free
+#    Software Foundation; either version 2 of the License, or (at your option)
+#    any later version.
+#
+#    This program is distributed in the hope that it will be useful, but WITHOUT
+#    ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+#    FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+#    more details.
+#
+#    You should have received a copy of the GNU General Public License along with
+#    this program; if not, write to the Free Software Foundation, Inc., 59 Temple
+#    Place - Suite 330, Boston, MA 02111-1307, USA.
+#++
+require 'ftools'
+require 'fileutils'
+require 'optparse'
+require 'optparse/time'
+require 'ostruct'
+require 'common/cvs'
+
+
+# The options specified on the command line will be collected in *options*.
+# We set default values here.
+options = OpenStruct.new
+options.all = false
+options.path = []
+options.cvsroot = []
+  
+opts = OptionParser.new do |opts|
+  opts.banner = "Usage: ribit-setup [options]"
+
+  opts.separator ""
+  opts.separator "Options:"
+      
+  opts.on("-p", "--path PATH",
+          "Define the data path") do |path|
+    options.path << path
+  end
+    
+  # NOTE: description is split to several lines    
+  opts.on("-a", "--copy-all",
+          "If specified then all possible files are copied.",
+          "If not defined then ribitconfig.xml and categories.xml is skipped") do |lib|
+    options.all = true
+  end     
+
+  opts.on("-c", "--from-cvs CVSROOTDIR",
+          "Instead of saving default files takes files from specified CVSROOT directory.") do |dir|
+    options.cvsroot << dir
+  end 
+  
+  # No argument, shows at tail.  This will print an options summary.
+  opts.on_tail("-h", "--help", "Show this message") do
+    puts opts
+    exit
+  end
+    
+end
+
+begin
+  opts.parse!(ARGV)
+rescue OptionParser::InvalidOption => e
+  # show help and exit
+  puts "Invalid option(s)!"
+  puts opts
+  exit
+end
+  
+if ( options.path.empty? )
+  puts "ERROR: No destination path defined!"
+  puts opts
+  exit
+end
+
+# if no cvsroot => copy default files
+if ( options.cvsroot.empty? )
+  # we know this scripts is located in <ribit-installation>/bin
+  # and we'd like find out what is that <ribit-installation>
+  ribitHome = File.join( File.dirname( __FILE__ ), '..' )
+  configDir = File.join( ribitHome, 'share', 'config' )
+  destDir = options.path
+
+  if ( options.all )
+    puts 'Copying ribitconfig.xml'
+    File.cp( File.join( configDir, 'ribitconfig.xml' ), File.join( destDir, 'ribitconfig.xml' ) )
+  end
+
+  puts 'Copying content'
+  sourceContentDir = File.join( configDir, 'content' )
+  destContentDir = File.join( destDir, 'content' )
+  FileUtils.mkdir( destContentDir ) unless File.exists?( destContentDir )
+
+  # TODO: delete all only if specially requested
+  #if ( File.exist?( destContentDir ) )
+  #  FileUtils.rm_rf( destContentDir )
+  #end
+  #d = Dir.new( destDir )
+  #d.each  {|file| puts ("Got " + file) }
+  #puts ( "source  " + sourceContentDir )
+
+  Dir.glob( sourceContentDir + "/**/*" ).each do |item|
+    #  )
+    if ( /\/CVS\//.match( item ) == nil and /\/CVS$/.match( item ) == nil )
+      # the path contains "sourceContentDir" prefix and we want get rid of it
+      # (+1 comes from '/' after "sourceContentDir")
+      endPath = item[sourceContentDir.length + 1 , item.length]
+    
+      if ( File.file?( item ) )
+    
+        # if all flag is false then we might skip some files (=categories.xml)
+        if ( options.all or !/categories.xml$/.match( endPath ) )
+		  puts( "Copying " + File.join( sourceContentDir, endPath ).to_s )
+          FileUtils.cp( 
+            File.join( sourceContentDir, endPath ), 
+            File.join( destContentDir, endPath ) )
+            
+          FileUtils.touch( File.join( destContentDir, endPath ) )
+        end
+        
+      elsif( !File.exists?( File.join( destContentDir, endPath ) )  )
+        FileUtils.mkdir( File.join( destContentDir, endPath ) )
+      end
+
+    end
+  end
+else
+  # source is cvsroot
+  cvs = CVS::CVS.new
+  cvs.cvsroot = options.cvsroot[0]
+  # checkout guarantees that basic structure will be there
+  cvs.checkout( "content", options.path[0] )
+  
+  # next update command requires changing current dir
+  oldDir = FileUtils.pwd
+  cd( options.path[0] )
+  begin
+    cvs.run_cvs( "update", ['-R', '-f', '-A', '-C', 'content'] )
+  ensure
+    cd( oldDir )
+  end
+end
+
+puts 'DONE'
+
+
+
+

data/bin/ribitd

@@ -0,0 +1,18 @@
+#!/sbin/runscript
+
+depend() {
+	need net
+}
+
+start() {
+	ebegin "Starting ribit"
+	start-stop-daemon --start --quiet --background -make-pidfile --pidfile /var/run/ribit.pid --exec ribit /opt/ribit/ribitconfig.xml
+	eend $?
+}
+
+stop() {
+	ebegin "Stopping ribit"
+	start-stop-daemon --stop --quiet --pidfile /var/run/ribit.pid
+    rm /var/run/ribit.pid
+	eend $?
+}

data/doc/TODO.txt

@@ -0,0 +1,207 @@
+
+====>Ribit Quick Memo and TODO
+
+
+KAMPELA:
+(DONE) - add all old content
+- change main page
+- test with mobile (new instance)
+
+NEXT VERSION 0.3:
+
+- bug: when creating a new category => it can't be edited from categories page
+- bug: creating new toplevel category =>  undefined method `get_next_free_page_id' for nil:NilClass
+- bug: <,> inside pre
+- bug: & char inside page name (test also ; > < #)
+
+- when name is changed => delete everything from cache
+
+(DONE) - on top bar would be nice to have categories link
+- on categories page would nice to have first list of categories => link to page listing
+
+* version + build number (including date) to releases an startup logging
+
+* do data reloading in startup, not in first request
+  
+* standalone version
+* putting to rubyforge (www pages, some fine tuning)
+* unit test html
+* coverage results 
+	- coverage ok
+	* filter out test classes
+	
+* print argo docs
+* think: www application structure
+* think: how to test www application; checking html structure
+* catching all from standard ruby loggers
+  - configuring log
+  
+* test ruby mock 
+  - simple unit tests
+    * call order
+    * exceptions
+    * default value 
+
+* haven't found good basic xhtml reference => maybe I should collect my own (www)
+* split classes to several files + take Ribit namespace into use
+* all inside Ribit module
+* split css to view, edit and edit category
+* refactor code to more readable form (more files?)
+* QUESTION: css could be used to set background image, can we that to get proper
+  section divider lines ?
+    * other option is to recogninize which one is first and which last paragraph
+      and set class attribute properly
+
+* think about default ribit content
+
+* think about category default page
+  * if some kind of index page then
+     * creating new pages without links
+     * how to move existing data to new system
+     * how write links to this default page
+  * idea is to avoid ::Work::Work style links
+  
+* www instructions
+  - cache
+
+================================================================================
+
+- how to make table of contents: example http://eclipsedarcs.org/doku.php
+
+NEXT:
+* sisällysluettelo categories sivulle: hypyt kategorian luo
+* blogimaiset entryt
+* blokkien editointi sivun sisältä
+
+* delete
+  - category
+   - page (and also moving from category to another)
+   
+* creating a new category just from through link (right away to edit page)
+* move link outlook to style sheet ([link] and link1 - link2)
+* cgi interface; how caching should work => disable ?
+* should startup main page just contain category => category main page then used ?
+* speed
+* startup: if file is missing but in CVS => restore
+* go through log messages, is it enough or too precise ?
+* tweak ribit-setup so that it can be run in dev
+* changing baseurl, some rules, ext url and internal url
+* better localization
+* XML validation
+* QUESTION: is startin space good indication to continue list item ?
+* check JSP Wiki and use similar stuff (+ www.snipsnap.org)
+
+* separate ribitconfig.xml and content, now they has to be (?) under same dir
+  - could ribit config point directly to content dir ? I.e. can name change
+
+* tables (?? howto implement headers)
+* pre-tag
+* running a single test
+* valid basic html (lists are not inside <p> ?)
+
+* titles with format '== title2 =='
+* possibility to create index of titles to beginning of the page
+* 'to top' tag to bottom
+* [ encoding like \[ (doesn't start the link) or like [[
+* unordered list
+* term list
+* preview possibility
+* favicon definition
+
+* moving pages from category to another
+* search functionality
+* better script for reverting back to old page versions
+* tag that lists of subpages
+* & is replaced by &amp; in form urls at least => no conversion
+* headers with ==header format== 
+  * tag to create table of contents
+
+* learn from other web applications: xml and css usage
+
+* distributed Ribit => local cache
+  * first it could be read only
+  * do only data move or ready rendered pages ?
+
+* removing pages and categories
+  * confirmation page
+  
+* creating new category (new page?): to whatever place in hierarchy
+  * under categories page link 'create new category'k
+  * handling of empty categories
+    - empty category main page (default page)
+
+* move most of Categories page formatting outside the class
+
+* in creation full category path should be given ? partial category ref
+
+* define css page name in config
+
+* users
+	- access rights checked for categories
+	
+* aliases for links (global) (like in EclipseWiki, WikiSpace style links) or (just short cut for [Main] for example, or tags for persons)
+
+* add XML validator to page creation (optional) but eases development
+* including UMLet and other drawing applets to pages
+   * textual input (what to draw), but applet shows that
+   * other option is that app generates dynamically gif/jpg to show
+   
+* pre tag with start and end point (spaces issue). There could small link where to load data directly
+* raw data flag
+* modifiable data flag
+* Ribit Admin API: way to force data reload
+
+* BUG?: extra ---- on bottom of page
+* BUG: if creating new page and then going 'back' and trying to create again.
+        => this could be a feature but better think through how 'Back'-button is handled
+        
+* ri to work with standard lib plus own
+
+* option: use CDATA or escaping
+* putting debugging data to page comments: request data. Option to config file
+
+* timestamp 'modified' to pages
+* windows port
+  - uid/gid changing stuff to linux only feature
+
+* generic error page (basic xhtml)
+* error page should be also XHTML! How to change it from REXML ? Make error page servlet specific
+* authorization, basic, does Webrick support ?
+* Webrick, implement possibility to have some kind of statistics
+* support for WAP
+   * new template + converter ? other options would be convert from XHTML to WML ...
+   * can we notice client type from somewhere ? => test servlet to see all HTTP-parameters
+
+* way to run one specific test (rake?)
+* debugging test runs
+* forward to tclhttpd / freeport in laajakaista   
+
+* taulukkolaskenta ominaisuuksia taulukoihin (esim summa ja keskiarvo)
+
+* howto to dynamically reload code
+
+
+QUESTIONS:
+
+* how CDATA end ]]> is escaped or is it ?
+* how to get rexml use " instead of '
+* what is 'em' in stylesheets ?
+* how to say 'wrap on' for text areas by using stylesheets ?
+
+OTHERS:
+
+- RubyEclipse
+   * TODO task tag
+   * point from stack trace
+   * the current dir
+   * CVS integration, showing other files too (filter?)
+   * bug: saving launcher properties doesn't save if I don't select some other profile
+   * debugging
+   * integrate ri
+   * integrate profiling
+   * outline: clicking => move to the method
+   * using whitespace
+   * creating new files in Ruby perspective
+   * better parsing error catching (possible?)s
+   * way to run rake from Eclipse
+   

data/doc/devdiary.txt

@@ -0,0 +1,258 @@
+13th of June, 2005
+
+It seems performance improvements were enough. Now Ribit works reasonably fast 
+on my old 200MHz Pentium. The current version is not polished (bugs found)
+but usable. There was several improvements: 
+
+1) processing template made too many loops
+2) dataload happened for every view action => needed only after saving
+3) created page cache that stores generated pages
+4) Ribit works with Last-Modified and If-Modified-Since HTTP headers
+
+When all above were implemented I realized that perhaps 3) wasn't necessary
+because 4) was implemented. Ribit is not targeted for masses. However 
+implementation of 3) resulted to easy implementation of 3). 
+
+
+3rd of June, 2005
+
+Few days ago Ribit 0.2 was finished. Unfortunately it revelead to be too slow 
+running with Pentium 200MHz. Yes, it can be used but response time is something
+like 10s. Next task is to improve performance. One bug was also found from
+category creation (fixed when writing this).
+
+Installation of new version happens pretty easily. I'm running gem_server on my
+dev machine when needed and then just updating Ribit gem from other machine. 
+I had to modify /usr/lib/ruby/gems/1.8/gems/sources-0.0.1/lib/sources.rb to add
+a new gem server server.
+
+
+27th of May, 2005
+
+Almost finished. Still one ribit-setup bug and then everything seems to be ready
+for 0.2. I have high hopes.
+
+
+14th of May, 2005
+
+During last week I have spent quite many hours with Ribit and now it is almost
+ready. All planned basic funtionality works (incr2) just little fine tuning
+left. I still plan to make a 'crunch' so that I finally get Ribit to full usage.
+
+Between those hours I have many times thought why I'm doing this. I can't name
+one biggest reason, but several ones thought
+
+- amphitition; I'd like to finish this because I have started it
+- there is a real need for this (althought they could be open source 
+  alternatives
+- I have big plans not included to this increment
+  - distributed Ribit
+  - mobile Ribit
+- I want to learn Ruby. I already like Ruby and I smell future promise there.
+- I want to keep my skill up to date, learn new things
+- I like coding and lately I haven't had opportunities to code at work
+
+After I finish incr2 and have taken to Ruby into use I'll release it, ie. put
+to rubyforge or something.
+
+
+18th of March, 2005
+
+It has been long time when I last time developed Ribit. But I think now when 
+spring emerges I'll continue. I hope to get working version installed to my
+closet server. Now I updated and tested installation scripts and instructions
+and managed to get server running easily.
+
+There was a new version of RubyEclipse, 0.5.0. It seams it is a big improvement.
+Now I can run my unit tests from Eclipse similar way as Java tests. It required
+little trick. RubyEclipse didn't support my project structure, the code under 
+'lib' and tests under 'test'. For Test::Unit I couldn't defined ruby parameters
+(-I lib:test) as for normal Ruby application. The solution was to create 
+a wrapper script that combines -I parameter and others defined from command
+line.
+
+I also had to rewrite alltests.rb because RubyEclipse didn't support the sort
+way where test files are just 'required'. I had to code a TestSuite.
+
+I noticed also that some tests have quite ugly depencies to test data files on
+the disk. Those files were not initially in CVS but now there are.
+
+The latest Log4r and RubyEclipse wasn't exactly compatible. Log4r complained
+that its deprecated attribute was used. Situation was fixed by editing 
+RemoteTestRunner.rb from RubyEclipse.
+
+
+15th of October, 2004
+
+I haven't been working with Ribit for awhile, but still thinking required 
+features. Using categories make me still wondering is the right way.
+
+I have used Confluence 
+(http://www.atlassian.com/software/confluence/default.jsp) and it contains
+Space feature and then pages can have hierarchical tree structure and those
+seem to work pretty well. Ribit doesn't have relationships between pages
+and I wonder do I need. In contradictory categories have relationships but
+Spaces in Confluence don't. Having hierarchical page structure page
+navigation little easier.
+
+Final I came to conclusion that categories work still well. I just need
+to define a main page for each category, so that I can build navigation
+helper. This way if user wants hierarchical structure he can utilize 
+categories. Also in future categories would the access rights checking point.
+
+Still not 100% satisfied, but I realize that maybe I can't get all. But I
+have made my selection and actual user experiences tell me whether was this
+right choice.
+
+
+23th of September, 2004
+
+Multilevel lists work now finally. It wasn't easy, but the current content
+parsing architecture supported this quite well. I'm pleased that something
+is working. The idea of intergrating TestUnit to RubyEclipse starts feel 
+right way. I need to run one tests at the time and I'd like to just click
+stack trace produced by TestUnit to go to the code.
+
+
+17th of September, 2004
+
+Categories working pretty well, altough you can't yet create new categories.
+After giving few thoughts I decided to use full path reference for categories.
+For example Ribit::Development::Test Cases. This way I don't need to repeat 
+information in category name (otherwise I would need to name Development to 
+Ribit Development to make a difference from other development projects.
+
+Still open how to implement page link bar (links to closest pages). First I
+thought my decision to use categories wasn't good enough compared to hierarchy
+of pages (tree structure), but finally I got conclusion that categories 
+structure is more flexible even tough it has it negative sides (not so 
+intuitive).
+
+
+23th of August, 2004
+
+Just sketching the new file hierarchy. Some points:
+  
+  * data is split to several files so that in writing we need to write less
+    than possible
+  * splitting to several files also makes easier to work with CVS. Later old
+    versions can be easily retrieved.
+  * category properties (category.xml) is separate from category hierarchy
+    (categories.xml) so that it easy to develop predefined content (merging
+    is easy)
+  * page id is separate from page name (that is used for referring, so that 
+    name can be easily changed, but file (containing id in its name) will be 
+    still the same.
+    
+
+20th of August, 2004
+
+During this week I have managed to work out version that uses CVS to store
+content files. This required some changes to directory structures and might 
+need some more, because I'm not yet satisfied to it. I moved reading and 
+writing of content files to new class ContentStore.
+
+
+1st of August, 2004
+
+Started today working towards version 0.2 where data could be split to several
+files and somehow history of changes is recorded just be sure that data is not
+lost. These changes are also needed so that development and actual usage can
+proceed smoothle so that merging data files has minimum effort.
+
+
+25th of July, 2004
+
+Running 'gem_server' as a daemon didn't work out so easily I thought. I had to
+study start-stop-daemon man pages to find out proper parameters and in addition
+finding how to pass parameters I found way to set effective userid and writing
+pid to a file. This means I could drop my modified 'gem_server'
+
+Today I managed to fine tune rakefile and now it contains everything essential.
+Hardest part was to include generated rdoc to packages in same run. It required
+some tricks.
+
+
+24th of July, 2004
+
+I had worked to make ribit installation work using 'gem'. Perhaps because I'm 
+quite new to Ruby I had hard times to get rake properly create necessary gem files.
+
+GEM SERVER
+
+My goal was to create my own gem repository where I could install development time
+versions to my own use. There wasn't really much information how to this. Using
+'gem_server' was quite obvious but first I didn't understand how should I order
+my files. Gem repositoy dir (passes as -d parameter for gem_server) should point
+to directory with following structure:
+
+<gem-repository>
+ |
+ - cache (place your *.gem files here)
+ - specifications (place your *.gemspec files here)
+ 
+I wanted gem_server starts as a daemon, so I created init.d script similar to
+ribitd. The script starts a modified 'gem_server' that writes the pid to 
+the /var/run/gemserver.pid so that the script can be used to stop the daemon too.
+
+To make possible to install ribit from local repository I needed to add my server
+to remote server list of 'gem'. This happened by modifying 
+/usr/lib/ruby/gems/1.8/gems/sources-0.0.1/lib/sources.rb the contain also 
+localhost:8808. The simple localhost wasn't enough I needed to specify the port too.
+Now I could install ribit by typing 'gem install ribit'.
+
+I also considered you make possible to install with setup.rb/install.rb but
+because ribit already depended on 'gem' I didn't see the any point to do this.
+
+RAKE AND GEM
+
+Creating *.gem file with rake was easy (just followed instruction from internet)
+but I realized that it was not enough. I needed *.gemspec to be able install ribit 
+via gem. But now by gemspec was included to rakefile. So I needed a separate
+gemspec file that could loaded from rakefile. Yes, could almost worked but 
+there was depencies to Rake::Filelist in gemspec file, so I needed to do my own
+file including and excluding. Maybe there exists a better way to do this all, but
+I didn't figure out what it could be.
+
+
+17th of July, 2004
+
+It's been a long break since last time working with Ribit. Since last time I have
+reinstalled my development system (now Gentoo Linux and Eclipse 3.0). At the same
+time I reorganized directory structure to match better with general ruby application
+layout. Now everythings works again and I'm just about to commit to CVS. I also
+decided that I'll start using gem (http://rubygems.rubyforge.org/) because in 
+the future I'd like package Ribit as gem application. Log4r is now used via gem. Other
+depencies like webrick and testunit are statisfied by ruby 1.8 which I also have
+started to use.
+
+
+3rd of June, 2004
+
+Now I have almost finished new content parsing system that replaces HTMLConverter.
+I'm quite pleased with the system, it is quite simple and easy add to new formatters.
+All formatters are not so independent as I originally wanted (universal parser) but still
+big improvement to old. Now data is is parsed to a tree and then tree elements are
+formatted. Still some TODOs and polishing left.
+
+
+23th of May, 2004
+
+I had big plans to rewrite the HTMLConverter to be more flexible and universal.
+On the paper  my plam seemed good but when starting to implement the 
+complexity overhelmed me. Finally I realized that I was implement too general
+system: tags could be anything, tags could nested, starting and ending tag 
+could be totally different, no reserved characters.
+
+Perhaps I could have managed to get something working but most probably it
+would been very complicate to maintain. So I falled back to old implementation
+and found out that it could easily and clearly expanded. Not so beautiful, 
+efficient and encapsulated as I wished but some that worked. 
+
+The current HTMLConverter implementation relies on regular expression. Also
+the order of handling elements is meaningfull. Still wondering could data be 
+parsed to a tree and them formatted ? When I have time I have to check how 
+other Wikis has solved this problem.
+
+
+