April 2005 - xwiki-devs@xwiki.org

Using a Document for storing the ApplicationResource data
by Stéphane 26 Apr '05

26 Apr '05

Hi extended wiki coders, What about storing the webapp's i18n data (currently stored in ApplicationResources_xx.properties files) in a Document? That way, the i18n resources too can be edited collaboratively. It is useful when you develop an application on top of XWiki that requires many additionnal i18n words. I've started to implement this in the XWikiMessageTool class: http://www.xwiki.org/xwiki/bin/view/Dev/XWikiMessageTool I'm currently using this modified XWikiMessageTool for managing the i18n resources of my site and I'm quite happy with it for now. I've added the tag "{table}" at the beginning and at the end of the list of words, so that the resources get formatted nicely in the wiki. This has no side effects in the Properties map. Stéphane

2 1

xwiki svn repository outage
by Mathieu Peltier 25 Apr '05

25 Apr '05

Hi, I am one of the ObjectWeb webmaster. For info, It seems that the xwiki svnroot repository is currently corrupted for some reason: > svn co svn://svn.forge.objectweb.org/svnroot/xwiki/trunk svn: Berkeley DB error while opening 'nodes' table for filesystem /svnroot/xwiki/db: Cannot allocate memory I am currently trying to recover the repository using "svnadmin recover" and will try to modifiy the defaut berkeley db config (maybe it would be better to switch to fsfs backend?). See: http://subversion.tigris.org/faq.html#bdb-cannot-allocate-memory I have a backup of whole repository until revision 485 in any case. Sorry for the inconvience, regards, -- Mathieu Peltier ObjectWeb Consortium - INRIA E-mail: mathieu.peltier(a)inrialpes.fr Phone: +33 4 76 61 55 25 JabberID: mpel(a)jabber.fr

2 2

Non alphanumerics and white spaces in Wiki page Names
by Ludovic Dubost 22 Apr '05

22 Apr '05

One thing I'm planning to do, is liberate the special characters in document names. I've had too many problems not accepting accents in wiki page names, even though for european accents they are converted to non-accentued characters automatically when used in a link. The problem is that this cannot be made general because some languages have special characters that do not have a "non-special" fallback.. So I'm thinking of letting any character being chosen in Wiki page names, using the URL encoding conversion of course. Doing this is not too much a problem. One other thing is what to do with white spaces. Currently if you type [new page] the link created is newpage I've had some complaints about it because when we show page searches it shows a concatenated text string instead of a nice page name. Another problem is that it makes it very difficult to search automatically for backlinks in the database since the white space could be anywere in the database. Backlinks would be usefull for displaying and for moving pages and automatically updating links. What I propose is to now default to new%20page for this link which one exception when the page 'newpage' already exists (this treatment could be an option). This way there would be backwards compatibility. Linked with the change for non alphanumerical characters it would free page naming. What do you think ? Ludovic -- Ludovic Dubost XPertNet: http://www.xpertnet.fr/ Blog: http://www.ludovic.org/blog/ XWiki: http://www.xwiki.com Skype: ldubost AIM: nvludo Yahoo: ludovic

2 4

Document Title
by Jiri Luzny 22 Apr '05

22 Apr '05

Hi, as I'm working on the TOC stuff which will allow automatic numbering I would like to discuss an aspect that would influence how the numbering is realized. In particular, I'm interested in knowing what is the typical document structure you guys use and how you apply the headings into document sections. What I have seen on xwiki.org is a mixture of various patterns like: 1 title 1.1 chapter1 1.1.1 subchapter1 1.1 chapter2 or 1.1 title 1.1.1 chapter1 1.1.1.1 subchapter1 1.1.1 chapter2 or 1 chapter1 1.1 subchapter1 1 chapter2 or 1.1 chapter1 1.1.1 subchapter1 1.1 chapter2 In our team we used to use this pattern: 1 title 1.1 chapter1 1.1.1 subchapter1 1.1 chapter2 The problem is that generated numbered TOC than would look like: 1 title 1.1 chapter1 1.1.1 subchapter1 1.2 chapter2 which is not consistent with our company's standard Word documentation structure (a common document structure, IMHO) which has this natural TOC: title 1. chapter1 1.1 subchapter1 2. chapter2 ... My question is whether you think it would be good idea to treat a document title as a specific part of document and introduce a new tag like {title} or 0 ... and have its own CSS item. In Confluence the document title is equal to physical wiki document name and rendered separately on top of the page. Another possibility which would solve the generation of TOC issue is to add an option to the TOC tag what initial heading level it should start the generation at. I would personally prefer to treat document title specifically as I consider the mentioned TOC option as a kind of hack, IMHO. Jiri.

2 5

RE: [xwiki-commits] r469 - in xwiki/trunk/src: . main main/com main/com/xpn main/com/xpn/xwiki main/com/xpn/xwiki/api main/com/xpn/xwiki/atom main/com/xpn/xwiki/atom/lifeblog main/com/xpn/xwiki/cache main/com/xpn/xwiki/cache/api main/com/xpn/xwiki/cache/i
by Vincent Massol 21 Apr '05

21 Apr '05

Luis, You need to try using the SVN move command instead of deleting and adding. That will preserve the history of files and it'll be lighter on the commit diffs too. Thanks -Vincent > -----Original Message----- > From: Luis Arias [mailto:[email protected]] > Sent: mercredi 20 avril 2005 22:21 > To: xwiki-commits(a)objectweb.org > Subject: [xwiki-commits] r469 - in xwiki/trunk/src: . main main/com > main/com/xpn main/com/xpn/xwiki main/com/xpn/xwiki/api > main/com/xpn/xwiki/atom main/com/xpn/xwiki/atom/lifeblog > main/com/xpn/xwiki/cache main/com/xpn/xwiki/cache/api > main/com/xpn/xwiki/cache/impl > > Author: kaaloo > Date: 2005-04-20 22:20:54 +0200 (Wed, 20 Apr 2005) > New Revision: 469 > > Added: > xwiki/trunk/src/main/ > xwiki/trunk/src/main/com/ > xwiki/trunk/src/main/com/xpn/ > xwiki/trunk/src/main/com/xpn/xwiki/ > xwiki/trunk/src/main/com/xpn/xwiki/XWiki.java [snip]

2 1

Reorg of the source hierarchy and precise version identification
by Luis Arias 21 Apr '05

21 Apr '05

Hello all, I've been busy this afternoon doing some housecleaning in the code base: - seperate folders for src/main - core java src/test - unit tests src/test-cactus - integration tests I've been building against the latests stable jars I could find, including hibernate 3.0.1 for which I had to change a lot of imports and some stuff in the configuration files. I just commited the version numbered jars because we are going to be playing with Maven 2 tomorrow night at the OSSGTP and I figured it might be easier to write the dependencies that way. I also ported the RSS macro to Rome 0.6 because it broke on the latest commons-digester. Rome sure is a sweet API ! We're going to be doing some other interesting things with Rome. So all of this stuff is still on my machine while I continue unit tests (actually these are quite long because a lot of them are actually integration tests and bring up hibernate (with showSql = true :( ) etc...), but since cruise control is coming back up, I'm going to just make sure the build runs and ant release works. I've been working on tests for my lifeblog stuff, and I definitely like running them one at a time directly in eclipse, even the server tests through Jetty. So I have to see how to refactor that so the other tests benefit and for it all to be ant and ide compatible. I bet you have a lot of ideas on that one Vincent ! ;) My concern, is what to do about your diffs and how changes to the src hierarchy could affect your current work. I hope that doesn't cause you too much pain. Please let me know if you have any ideas on this issue because I'm a bit stumped. Actually maybe I can just tell you when I start to commit and when its done and then you can move your sources in the right place (easy to do) and update. Cheers, -- Luis Arias http://www.xwiki.com http://www.innover-entreprendre.net +33 6 14 20 87 93 mobile

3 6

Re: Database Support
by Ludovic Dubost 21 Apr '05

21 Apr '05

Hi Bruce, Nice to meet you. The first test you need to do is start XWiki with an empty database connected to the Firebird database. In xwiki.cfg you should setup "xwiki.superadminpassword=toto" that will activate the "superadmin" user with password "toto" which will be able to setup rights. From there you should be able to access an empty wiki and create some pages to see if everythings works. Features to test are: 1/ Pages 2/ Attachments 3/ Versions One of the key features needed is the support for large text fields (without size limits). Oracle can't do this (except in the latest versions) and required some specific hibernate hacks. Converting the default Database to Firebird might be more complex. You can download the default DB on Sourceforge. It is a zipped SQL file exported from Mysql. However I think it is quite full of specific MySQL sql. I think there are some improvement in MySQL 4.1 to the export with allows for "compatible" export so you might need to import/export in MySQL to create a more compatible file. Ludovic Bruce Bacheller a écrit : >Thanks for the Quick response. > >I am on the Firebird Foundation (FirebirdSQL.org) and will get you any support needed from our THE FirebirdSQL side to get it working. > >I will host it on my site which is FirebirdXML.info or Infodes.org. >That is Linux/Apache/Tomcat. > >We are having hosting our user conference in germany next Month May 15th and I would like to have it operational for that event. I will make sure that XWiki gets good exposure to the Firebird Community as well. > >We are pretty close to Oracle in syntax. > >Can you send me a copy of the SQL code in MySQL (table creation) so I can review it and set up test. > >Looks like XWiki has developed A VERY NICE piece of Open Source Software. > >Regards, >Bruce > >-----Original Message----- >From: Ludovic Dubost [mailto:[email protected]] >Sent: Wednesday, April 20, 2005 3:54 PM >To: Bruce Bacheller; xwiki-dev(a)objectweb.org >Subject: Re: Database Support > > >Hi, > >We haven't made any tests, but XWiki uses hibernate and JDBC so there is >a good chance that it works. >You will need to try. We are open to requests in case of problems. > >Ludovic > >Bruce Bacheller a écrit : > > > >>Do you have plans to Support Firebird the Open Source DBMS ? >> >> >> >> >> > > > > -- Ludovic Dubost XPertNet: http://www.xpertnet.fr/ Blog: http://www.ludovic.org/blog/ XWiki: http://www.xwiki.com Skype: ldubost AIM: nvludo Yahoo: ludovic

1 0

Re: Database Support
by Ludovic Dubost 21 Apr '05

21 Apr '05

Hi, We haven't made any tests, but XWiki uses hibernate and JDBC so there is a good chance that it works. You will need to try. We are open to requests in case of problems. Ludovic Bruce Bacheller a écrit : > Do you have plans to Support Firebird the Open Source DBMS ? > > > -- Ludovic Dubost XPertNet: http://www.xpertnet.fr/ Blog: http://www.ludovic.org/blog/ XWiki: http://www.xwiki.com Skype: ldubost AIM: nvludo Yahoo: ludovic

1 0

Cruise Control
by Ludovic Dubost 20 Apr '05

20 Apr '05

Hi All, I'm currently setting-up cruise control 2.2.1 to setup a continious build on the latest SVN source. It will compile XWiki and run tests everytime there are changes in SVN, including the cactus based tests. Results will be published on a web site: http://build.xpertnet.biz and a notification email will be sent either to the dev list or the commit lists. Any comments about the choice: dev or commit list ? I might automatically publish the artifacts to one of the xwiki.com production machines. The build machine is running Mandrake Linux. Ludovic -- Ludovic Dubost XPertNet: http://www.xpertnet.fr/ Blog: http://www.ludovic.org/blog/ XWiki: http://www.xwiki.com Skype: ldubost AIM: nvludo Yahoo: ludovic

2 1

Chinese Character not encoded as UTF-8 in web page
by Patrick Lee 20 Apr '05

20 Apr '05

Hi Ludovic Thanks for the help. I can now compile the svn source. On 4/19/05, Ludovic Dubost <ludovic(a)pobox.com> wrote: > > Hi Patrick, > > I think it is just that you need junit.jar in your ant installation. > > Ludovic > I now have the following questions: 1. the version # is now Version 0.9.500, smaller than the previous downloadble war file ( 0.9.543). Is this correct? 2. I have attempt to save some chinese character on a new page then try to edit it again using the default (html ) editor. Both the html page source and the editable version still show 大 etc for a chinese character, just like the previous version (0.9.543). Besides, the page source shows the following: <?xml version="1.0" encoding="ISO-8859-1" ?> <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> <html xmlns="http://www.w3.org/1999/xhtml" lang="en" xml:lang="en"> <head> <meta name="language" content="en" /> <meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1" /> I suppose that the xml/html file is encoded in ISO-8859-1 instead of UTF-8. Is there anywhere I could change the default? And it is the same no matter I choose the page/document language as en or zh. 3. I have also try to export this to PDF and the result supprise me. Please check the attachment. The three encoded chinese characters have been truncated to only one code with some left over of the first and last character. I will take a look at the svn revision 452 about the uft-8 changes set soon and try to find out the problem. Thanks in advance. Patrick Lee

2 2