Next: 3. Signed and binary Up: 2. Principles of conversion Previous: 2.2 Text to XML Contents Index

2.3 XML to text conversion

This conversion is a little bit more difficult and even in some cases impossible. Remember, TXT2XML implements only a subset of XML and COBOL. For example, this XML file :

**Figure:** Nonconvertible XML file
$\begin{figure}\par\par \begin{list}{}{ \setlength{\rightmargin}{\leftmargin... ...st} \textbf{</city>} \end{list} \textbf{</weather>}\end{list}\par \end{figure}$

In this file, the XML element city is a complex one. It includes the simple element sky but also a content (Hot and sunny). The corresponding COBOL copybook would be something like :

**Figure:** Corresponding COBOL copybook
$\begin{figure}\par\par \begin{list}{}{ \setlength{\rightmargin}{\leftmargin... ...pic x(20).} \par \textbf{03 sky pic x(8).}\end{list} \end{list}\par \end{figure}$

This is impossible in COBOL : a grouping item can't have a picture clause.^2.5

So, not all the XML files can be converted. But, with avalid COBOL copybook, the XML file is read one line at the time and data is extracted using the following rules :

Mixed XML content is not supported,
All the XML tags and contents may expand on more than one line,
XML headers starting with <?xml and ending with ?> are skipped,
XML entities starting with <!ENTITY and ending with > are not supported. If such a declaration is found, the conversion process is aborted,
XML internal and external document type declaration starting with <!DOCTYPE and ending with either > or ]> are skipped,
XML comments starting with <!- and ending with -> are skipped,
The first child element is then searched. Each time an XML first child closing tag is found, a text record is written and a new text record is initialized with default values from the COBOL copybook,
Attributes of the first child element are supported. But, attributes of the root element are ignored.
Each XML element, even without a content, must have an equivalent in the COBOL copybook. To find it, the XML element name is capitalized and underscores are translated to dashes. For example : <cobol_item_01> will become COBOL-ITEM-01 and is searched in the COBOL copybook,
Each XML attribute may have an equivalent in the COBOL copybook,
XML attributes like xml:space are ignored,
Each time a content is found, it's length and data type are checked against the characteristics of the corresponding COBOL item. Numeric items are right justified and left zero filled. If necessary characters are escaped.^2.6
If some XML elements or attributes are missing, conversion continues with the new element found. Missing COBOL items are left to their default value (see section ).
Empty XML elements, even with attributes, are supported.

Next: 3. Signed and binary Up: 2. Principles of conversion Previous: 2.2 Text to XML Contents Index

sunuraxi@users.sourceforge.net