AIProject Log

tag:

   <!--Include-->
   <xsl:if test="//include">
      <xsl:call-template name="include"/>
   </xsl:if>   

The text inside the document will be merged with all its tags. Further functionality to parse this included document as gwml is to be implemented.

Lee GWML

Extended the GWML XML transformator. Added functionality for links, lists and table of contents. It can generate TOC for a whole document from arbitrary title depth. Support for camelcase links is present. Tomorrow I will merge it in the trunk.

Wouter - validation

I have been working on actually adding the validation. This went fairly well; I had to refactor thing a bit. Instead of Strings which denote paths the validators now take InputStreams? as input. They also throw wrapper Exceptions to deal with the wide variaty of Exceptions thrown by the different libraries and to make it easy to plug in some other library which might throw an entirely different type of Exception. I also wrote a factory for the validators, this still needs to be plugged in, but that shouldn't be to much work. First I have to solve the current problem that one of the InputStreams? given to the validator is null, I hope to solve that tomorrow.

Harald & Jonathan

Made the storage available under extensionfunctions using a ThreadLocal? variabele. Besides that we started on isolating the parsers, so we can later change them into plugins and change the current parser factory by loading plugins. We found that interfaces are not allowed to have static methods. We also wanted to make static parsers, since they have only 2 methods. We thought of putting the parsers into the gw svn system and tried this. The reason we would like the parser to be more abstract is for that they can also be used in an extension include function and give the user a very powerful tool. Unfortunately we had a great deal of delay by the many groups that started to work on the trunk as well these days and by bugs to show up along with it.

Tuesday 18 October 2005

Harald & Jonathan

We searched with Martin to a solution of geting the storage within the extension functions. This seems to be difficicult to accomplish with xalan. We decided to do this by a different solution. We want to put the session in the ServletContainer? thread, so that at any time in the code the storage can be accesed.

Wouter & Lee GWML

We finished the integration of the new GWML language into the trunk. After updating our local working copy we noticed that it didn't work anymore, apparantly Gerrit moved some code that we used to another place (from URIResolver to StyleSheetLocator? ) without deleting the original location. Luckily Gerrit noticed that we were having problems with this and pointed out to the changes Now the integration is nearly complete. We still need to activate the validation and make sure we have some nicer error handling. Last but not least we have to complete the XSL transformator (now it only handles the structure of a document).

Barbara & Martijn

Held up waiting for a commit by Wouter & Lee, the "temporary" merge function is done for the time being, next step is wildcards.

Monday 17 October 2005

Wouter & Lee GWML

We tried to integrate a basic version of our XSL transformator into the current GW codebase. At first we thought it would be just two or three files that we had to replace, but apparently the GWML specific Java / XSL code is cattered all over the code base. Luckily it's mainly in the XSLT directory, but some parts are spread in the other packages. We managed to get the new GWML working for the interpretation of text entries, but it is still not working for directory listings or other applications like attachments. We expect to commit tomorrow the changes, because we can't guarantee a working trunk. Our changes will however effect the application groups, especialy the bibtex group who has already made some view and preview xsl files.

Harald & Jonathan

We started on creating an inlcude function for the xslt stylesheets, in order that sources can be included with an arbitrary specified output generated by differend kind of parsers. In this way e.g. bibtex files can be shown in fancy xhtml with at the bottom the sources published as well. Besides that we discussed with the Bibtex group on what they need and the fact that we are not going to save the to xml parsed bibtex code, but the bibtex code itself.

Barbara & Martijn

During the teammeeting Barbara expressed our issues with the merge function (side-note: becareful where you use "include function" and "merge function", including files is quite different). One could say our cries were heard (wink) and Harald & Jonathan will help out writing an extention function, which should allow us enough space and time to complete a temporary merge function based on our work so far and in such a fashion requested by Martin and Eelco.

Tuesday 11 October 2005

Wouter & Lee GWML

We started together with evaluating the current GWML language and made some design decisions. Wouter worked on finishing the GWML language while Lee started with making an XSL file for the transformations to HTML. More specific information on the XSL can be found on the personal log of Lee.

Harald & Jonathan

We tested to add xslt stylesheets in the gw svn and use them instead of styelsheets that are present in the source. The Stylesheet Locator needed some alterations for it as well. Stylesheets that attached using the attach servlet are not editable, because they are treated as a binary. It is possible to use them by attaching them to the stylesheets/ directory.

Barbara & Martijn

Still working on what is almost turning into bending the system to allow the proper XML files known to the XSLT written for merging them.

Monday 10 October 2005

Wouter & Lee GWML

Finally a break-through! We decided to get rid of the jing/jarv/jdk5 integration crap and go for our own custom solution! the W3C? schema was already working by default using the new javax.xml.validation libraries, but the relax ng should've worked also, but apparently not. We managed to get the jing library working separately and then wouter did a nice job with integrating it with a wrapper facade on top of W3C? and Relax NG. Lee looked for the DTD integration (source: http://www.geocities.com/herong_yang/xml/dtd_validation.html). We are going to integrate the whole stuff tomorrow The GWML language is also almost finished. We will soon issue the GWML language to the other teams and start writing XSLT files to transform it to html. We are getting finally back on the right track

Harald & Jonathan

We encoutered some bugs with the default stylesheets, which are invoked after the default parser. Together with the group meeting, koen's experiment and some bugs created by a merge with the trunk from the bibtex group we had a great deal of delay. It felt like a day putting out fires all the time instead of being productive.

Barbara & Martijn

Stumbled across issues with how parameters can be sent from the java engine to an XSL, the current fashion in which params (mainly a path) are sent along isn't flexible enough to allow for a list of files which we desire to be merged. The XSLT is mainly done and no longer the problem, working out how to make it properly accesible by the applications using the system is someting we're strugling with.

Tuesday 04 October 2005

Wouter & Lee GWML

We are working on the implementation of W3C? Schema, Relax NG and DTD validation. This is however more difficult than we expected. The problem is apparently that the support for Relax NG is not very extensive, especially in the Java 1.5 we are using. The solution we want to achieve is using the javax.xml.validation library from JDK 5 as a uniform interface for the different kind of validations. This works perfectly for W3C? schema, but refuses to work with Relax NG. Although the Sun documentation (if there is any) states that there is support for Relax NG, but that is not present. So we found out that we could use some "connector" made for the JARV library, but that was experimental and not guaranteed to work in the final 1.5 Java. It's not working in our case, although we are pretty sure we are doing the right thing.

The plan is to test some stuff on JDK 1.4 and see how things are going over there

The GWML definition however is working fine. We have almost a draft from our language. Next group meeting we will discuss this with Eelco and Martin.

Harald & Jonathan

Finishing up on the default parser.

Barbara & Martijn

Attempting to push a simple XSLT for merging XML files forward for commit, while also working on the URI resolver (which isn't in any shape to be commited currently sadly).

Monday 03 October 2005

Group meeting

We talked about the progress of the group members as well as some specific details. Everyone is still working on the issues they were working on last time, but there is progress.

We will probably integrate our repository with the trunk today or tomorrow so the other groups can merge their changes.

• Jonathan & Harald: Started on intergrating a default parser in case a unknown content type is specified. Default stylesheets have to be added to it. We discussed with Gerrit & Huib about how to implent possible parser extensions. We might re-implement the parsers as plugins, so that different parsers belonging to different contenttypes can be added.
• Barbara & Martijn: Working on the URIresolver. It certainly is a bit of a nuisance and rather inflexible at that, refactoring and completing some of the TODO's we hope will make this class less of a headache.

Tuesday 27 September 2005

• Wouter GWML: Been busy creating a draft for the GWML definition, using some ideas from DocBook? and HTML itself. Based on this we can create the Relax NG schema.
• Jonathan & Harald: Refactoring and fixing minor bugs and deficiencies that came along the mime/content-type alteration.
• Barbara & Martijn: Optimizing current XSLTs, writing and delivering a short XSLT to include a so called left bar to help further the work on how the pages are displayed.

Monday 26 September 2005

Group meeting

We evaluated the group member's work:

• Barbara & Martijn: They are both investigating ways to transform multiple XML files in one transformation. We agreed that they both work out a proof of concept with their ideas and during the next group meeting we decide which alternative is the most feasible.
• Jacob: He took the css file from the old wiki and made it work with the new code base. A thorough rework and refactoring of the old ugly coded css file was necessary. Now he is working on applying the layout to the current code base and commit it in subversion.
• Jonathan & Harald: Mime-type & content-type used for selecting the stylesheet and for setting the http header is seperated. The mime type for setting the http header is currently defined by the xslt stylesheet instead of by the user entry. Stylesheet names changed along the new name convention of stylesheets, so that they include the content-type.
• Wouter & Lee: Trying to get validation of RELAX NG working. Problem is that the current libraries from JING and JARV are quite old and not fully compatible with Java 5 Tiger. The new java.xml.validation libraries look promising, but for a strange reason there is no support for RELAX NG, although the home page does mention it. Eelco Dolstra shared some ideas for refactoring the GWML to a "docbook" style syntax.
• Gerrit & Huib: Writing documentation to make the processes more clear. Cleaning up html out of java code. Assisting other team members and teams with explanation of the code.
• Damir: Still not present. He should be working on the authentication part, but if he won't join development sessions anymore, then we'll skip that

Agreements made on 26-09-2005 with the different teams regarding XSLT/XML

BlogServerRequirements

This group will further discuss what their wishes may include: A new xml-file for each entry and comment or not? After they negotiate such, they'll let us know how they hope the filestructure for the blog will end up. Whether this is one big xml-file or many smaller xml-files sorted in a directory-structure. Should multiple xml-files be highly desirable we will need to support the merger of many files into one.

Update: Atleast for now, the Blog group has stated to prefer a single xml-file solution.

ConferenceManagement

The conference management group promised to support a transition to GWML. Lee and Wouter in the Application Interface group are currently further developing this MIME-type, though it is currently supported by the engine. Their request is Lee and Wouter keep them informed on changes in the GWML MIME-type so they can guarentee vallid GWML is sent.

BibtexServer

This group will take care of their own XSL-transformations to XML, after which they'll deliver ready-made HTML to us. We'll have to import their MIME-type and transformation into the engine. They'll deal with the transformation, we have to make the engine aware of which transformation to call.

Wednesday 21 september 2005

discussion XSLT solution of Barbara Pieters

I have written a short discussion including some pro's and cons on my solution to let an xsl tranformation transform more than 1 xmlfile at a time. To take a look at my solution, see the link to the jarfile at the bottom of the page (this one also contains a readme-file). To read this discussion see the following link: DiscussionSolutionXsltBP

I hope this will help you to decide which solution (Martijn's or mine) we are going to use.

Tuesday 20 september 2005

Wouter & Lee GWML

Several links leading to documents and tutorials we encountered

RelaxNG? (alternative for DTD and XML Schema):

• http://www.relaxng.org
• http://www.relaxng.org/tutorial-20011203.html#IDAGSZR
• http://www-128.ibm.com/developerworks/library/x-matters25.html
• http://www-128.ibm.com/developerworks/library/x-matters26.html
• http://www-128.ibm.com/developerworks/library/x-matters27.html

Validators

• https://bali.dev.java.net/
• http://www.thaiopensource.com/relaxng/jing.html
• http://iso-relax.sourceforge.net/JARV/JARV.html
• http://java.sun.com/j2se/1.5.0/docs/api/javax/xml/validation/SchemaFactory.html newInstance methode: beschrijving hoe implementaties te koppelen

DocBook? (markup language for technical documentation):

• http://doc-book.sourceforge.net/homepage/
• http://www.oreilly.com/catalog/docbook/chapter/book/ch02.html
• http://xml.web.cern.ch/XML/goossens/dbatcern/

Monday 19 september 2005

First project meeting with all team members and Eelco Visser and Martin Bravenboer.

Devided the subjects among the members

• parsing more than 1 xml file using xslt to html (Barbara & Martijn)
• the servlets part (Harold, Jonathan, Gerrit, Huib)
• gwml parsing (Lee, Wouter)
• css creation (Jacob)
• authorization (Damir)

From the servlets part:
Got an idea about how the servlets perform their operations. We suggest a change for the content-type handling. Further we started thinking about converting the fixed html into xml/xslt.