XMLPublication - release notes

XMLPublication comes with ABSOLUTELY NO WARRANTY.
This is free software, and you are welcome to redistribute it
under the terms of the GNU Public License (see www.gnu.org)

by J.M. Vanel , Copyright © J.M. Vanel - 2001-2002 Back to main page

Last update: jeu 09 mai 2002 15:53:26 CEST

Changes

0.4.1 -> 0.4.2 2002-05-09

• we used <p> tags, but switched to <div> in this version, because <p> as a wrapper is not valid with respect to the HTML DTD. It should pass intact through most of the HTML editors (testers wellcome !).
• I wrote a SAX filter to replace the XSLT transform that wraps paragraphs inside HTML tags; to integrate it, I used styler , a Task for Ant; it makes useful combinations of XSLT transformations and SAX filters. This should allow to treat MUCH LARGER FILES more quickly.
• test with JDK 1.4.0 , Saxon 6.5.2, and last Tidy from CVS on Linux RedHat 7.0; it should still work on Windows.

0.4 -> 0.4.1 2002-02-18

• optimization in index generation: use of XSLT extension function saxon:tokenize(); pure XSLT mechanism was too slow
• adapt to Saxon 6.5 ( to use XSLT version 1.1 features, the stylesheet should now specify version="1.1" )
• better parameters for tidy, omit HTML warnings (use last version of tidy at sourceforge.net )

0.3 -> 0.4 2002-02-01

This is a bug-correct release with one new functionality.

• merge items (from different document sources) having one same rubric (typically the species name); in thesaurus.xml at first level add this :

  <merge-items according-to-label="my-rubric-label" group-by="rubric|source" />

  and type in the shell "ant all-merge" .

  Instead of the default, which is to create one HTML page per item, this will generate one HTML page for each different value of the specified rubric, thus merging items from different document sources.

• the standard build.xml (the one from empty_project/) now copies :
  • all *.css from USER_SETTINGS/ into html/ , html/items/ and html/items-merge/
  • all *.html from USER_SETTINGS/ into html/
• correct bug in stop-words treatment (in make-index-by-rubrics-impl.xslt ); this was a consequence of the change in structure of thesaurus.xml
• enhance skeleton of USER_SETTINGS/thesaurus.xml :
  • for keywords take in account several forms for a word:
    • word,
    • word glued with ,;. etc,
    • Word (ignore case)
  • now generate the item name rubric (<h1> tag) for paragraph-structured documents (correction in publication-magician.xslt )
• systematic use of <dependset> and failonerror in build.xml to ensure that result pages are updated when needed
• simplification in implementation: from 2 to 1 concrete structure for the master.xml file: documents with paragraphs and sub-paragraphs
• check presence of document sources with suffixes *.table.html *.parag.html *.table.htm *.parag.htm *.xls

0.2.98.2 -> 0.3 2001-11-09

• arrange utf-8 output in XHTML by changing all presentation.html files
• change all thesaurus.xml files after schema change (did a small XSLT transform helper, update-thesaurus.xslt )
• changed stop-words processing, with a new <xsl:template name="filter-words-list">
• updated user documentation in english
• [for the empty project, easy to use as a starting point] make an XSLT transform that generates out of master.xml a skeleton of USER_SETTINGS/thesaurus.xml, that works as is, and can be further customized; by the Ant target named publication.xml, a file USER_SETTINGS/publication.xml is generated from work/*.specific-tags.xml
• multiple source documents of several structures (table or paragraph) can be droped in the same project (ça c'est bien, hein, Etienne!)
• change the structure of thesaurus.xml :

  <publication>

   <rubric xpath='relative_XPath_in_item_in_master.xml'
           item-xpath='absolute_XPath_in_master.xml'
           name='unique_name_suitable_for_a_file_name'
           label='human_readable_name'
           use-stopwords='yes'
           use-keywords='yes'
           structure='row'
   >

    <keywords>
     <w>my_keyword</w> ...
    </keywords>
   </rubric>
  </publication>

  Notes:

  • xpath= is the same thing as index-by= before,
  • item-xpath= is a new attribute, default is /*/tr for table source documents, /*/*[local-name()="p"][@class="h1"] for documents with paragraphs and sub-paragraphs
  • structure='row' or 'paragraph', depending on the file type, indicated by the file suffixes : .table.htm(l) or .parag.htm(l)

  Why this change?

  • more human-readable and explicit
  • prepares the way for multiple source documents of several kinds
  • prepares the way for skeleton USER_SETTINGS/thesaurus.xml generation
• migrate all documentation in XMLPublication/doc directory, to have all in the same CVS directory, to have all in the distribution ZIP, and to have the same directory layout in the CVS, in the distribution, and in the Web site

0.2.98.1 -> 0.2.98.2 2001-11-04

• use HTML class attribute with an value of xmlpub-add-content as insertion point marker for site specific XHTML wrapper (it will survive to any HTML editor); modification in kernel/wrapper.xslt
• empty project, easy to use as a starting point

0.2.98 -> 0.2.98.1 2001-10-31

• add a sample project: XMLBibliography, thanks to T. Madillo
• simplify item_file_name.xslt in sample projects
• user doc. in french

0.2.97 -> 0.2.98 2001-10-30
- No rubric header alone in item page if rubric empty (modif. in format-HTML-impl.xslt)
- add omit-xml-declaration="yes" for XHMTL documents produced, because of problems with Internet Explorer, which didn't format as XHTML in spite of the correct namespace; it was also necessary to add encoding="utf-8" , otherwise xml-declaration occurs in case of iso-8859-1 document (modif. in make-index-by-rubrics-impl.xslt format-HTML-impl.xslt)
- remove unused xsl:processing-instruction in wrapper.xslt
- add a top-level build.xml file in the distribution
- put kernel + samples under CVS (Concurrent Versioning System)

0.2 -> 0.2.97 2001-10-15
- make-words-list.xslt completely rewriten using tokenize-string.xslt
 (x17 performance improvement)
- suppress figures in hyperlinked index
- take in account stopwords in hyperlinked index
- abandon GNU make in favour of Jakarta ant
- a new connector for Word Processor texts with paragraphs
- a new connector for MS Excel SpreadSheets

0.1 -> 0.2 2001-09-14
- changed: user_documentation.html
- changed: format-HTML.xslt
 Header generation is delegated to the template item_name
- added: item_file_name-generic.xslt: example of generic naming of header and file (template item_name)

Back to main page