Reading XML files

For ISO Module-2

Tested with XDS and Stony Brook Modula-2

By Frank Schoonjans (frank.schoonjans@medcalc.net) - 7 December 2004

I am working on a library to read XML files. When programming the parser I wanted to avoid (a) the use of recursion, and (b) any limitations on string length.

Error reporting is limited.

I have tested with Stony Brook and XDS. In XDS, you must select the Project option "Enable language extensions" and "Default memory management".

I hope to receive your comments or suggestions.

The XMLRead library has the following interface:

DEFINITION MODULE XMLReadLib;

FROM SYSTEM IMPORT CAST;

TYPE StringPointer = POINTER TO ARRAY OF CHAR;

TYPE XMLHandle;

CONST INVALID_XMLHandle=CAST(XMLHandle,NIL);

PROCEDURE XMLOpenDocument(filename : ARRAY OF CHAR) : XMLHandle;
PROCEDURE XMLCloseDocument(VAR handle: XMLHandle);

TYPE AttrPair   = RECORD
                    IdPtr  : StringPointer;
                    ValPtr : StringPointer;
                  END;

     PairRecPtr = POINTER TO ARRAY OF AttrPair;

     XMLElementRec = RECORD
                       NamePtr    : StringPointer;
                       DataPtr    : StringPointer;
                       AttribSz   : CARDINAL; (* number of Attributes *)
                       Attributes : PairRecPtr;
                     END;

     Tags = (opentag,closetag,emptyelementtag,processinginstruction,comment);

TYPE XMLCallbackProc = PROCEDURE(CARDINAL, Tags, XMLElementRec) : BOOLEAN;

PROCEDURE XMLSetCallback(handle   : XMLHandle;
                         userdata : CARDINAL;
                         proc     : XMLCallbackProc) : BOOLEAN;

PROCEDURE XMLParseDocument(handle : XMLHandle) : INTEGER;
(* returns  -1 = Invalid file handle
             0 = OK
             1 = Found '<' but expected '>'
             2 = Found '>' but expected '<'
             3 = Wrong end tag
             4 = Not all elements are closed
             5 = File contains trailing text
             6 = Found '>' but expected '?>'
             7 = Found '<!' but expected '<!--'
             8 = Found '>' but expected '-->'
             9 = Expected closing "
            10 = Expected =
            15 = Out of memory
            *)

PROCEDURE XMLGetErrorString(handle : XMLHandle; VAR strptr : StringPointer);

END XMLReadLib.

Here is a simple test program

MODULE ReadTest;

FROM XMLReadLib IMPORT XMLHandle,XMLOpenDocument,XMLSetCallback,XMLParseDocument,
                       XMLCloseDocument,XMLElementRec,Tags,XMLGetErrorString,
                       StringPointer;

FROM STextIO    IMPORT WriteString,WriteLn,ReadChar;

PROCEDURE ShowElement(element : XMLElementRec);
VAR c : CARDINAL;
BEGIN
    WriteString(element.NamePtr^); WriteString("'"); WriteLn;
    IF element.DataPtr#NIL THEN
       WriteString("       data : '");
       WriteString(element.DataPtr^);
       WriteString("'");
       WriteLn;
    END;
    IF element.AttribSz>0 THEN
       FOR c:=0 TO element.AttribSz-1 DO
          WriteString("              '");
          WriteString(element.Attributes^[c].IdPtr^);
          WriteString("' : '");
          WriteString(element.Attributes^[c].ValPtr^);
          WriteString("'");
          WriteLn;
       END;
    END;
END ShowElement;

PROCEDURE CallbackFunction(userdata : CARDINAL;
                           tag      : Tags;
                           element  : XMLElementRec) : BOOLEAN;
BEGIN
    CASE tag OF
    | processinginstruction : WriteString("Proc instr. : '"); ShowElement(element);
    | opentag               : WriteString("Open tag    : '"); ShowElement(element);
    | closetag              : WriteString("Close tag   : '"); ShowElement(element);
    | emptyelementtag       : WriteString("Empty elem. : '"); ShowElement(element);
    | comment               : WriteString("Comment     : '"); ShowElement(element);
    END;
    RETURN TRUE;
END CallbackFunction;

VAR handle : XMLHandle;
    ch     : CHAR;
    str    : StringPointer;
BEGIN
    handle:=XMLOpenDocument("test.xml");
    IF XMLSetCallback(handle,1,CallbackFunction) THEN END;
    IF XMLParseDocument(handle)#0 THEN
         WriteString("ERROR : ");
         XMLGetErrorString(handle,str);
         WriteString(str^);
         WriteLn;
    END;
    XMLCloseDocument(handle);
    ReadChar(ch);
END ReadTest.

After you open the file using XMLOpenDocument, you assign a callback procedure to it with XMLSetCallback.

Next call XMLParseDocument and finally close the file using XMLCloseDocument.

The XMLParseDocument procedure will call the Callback function you have defined when it encounters a processing instruction (enclosed in <? ?>), an open tag (enclosed in < >), a close tag (enclosed in </ >), an empty-element-tag (enclosed in < />) or a comment (enclosed in ).

The first parameter of the callback producure is the CARDINAL number (of your choice) that you have specified in the XMLSetCallback function.
The second parameter of the callback producure is the type of tag found in the XML file (opentag, closetag, emptyelementtag, processinginstruction, comment).
The third parameter is a record with tag or element information. E.g. for the element
<document author="frank"> data </document>
The members of the XMLElementRec record are:
NamePtr : a pointer to the string "document"
DataPtr : a pointer to the string "data"
AttribSz : 1, and
Attributes[0].NamePtr : a pointer to the string "author"
Attributes[0].ValuePtr : a pointer to the string "frank"

For the following XML file:

<?xml version="1.0"?>
<document author="frank">
     <empty_element/>
     <field1>
          <!--comment 123-->data for field 1
          <sub1>sub stuff 1</sub1>
          <sub2>sub stuff 2</sub2>
     </field1>
     <field2>data for field2</field2>
</document>

the output of our test program is:

Proc instr. : 'xml'
              'version' : '1.0'
Open tag    : 'document'
              'author' : 'frank'
Empty elem. : 'empty_element'
Open tag    : 'field1'
Comment     : 'comment 123'
Open tag    : 'sub1'
Close tag   : 'sub1'
       data : 'sub stuff 1'
Open tag    : 'sub2'
Close tag   : 'sub2'
       data : 'sub stuff 2'
Close tag   : 'field1'
       data : 'data for field 1'
Open tag    : 'field2'
Close tag   : 'field2'
       data : 'data for field2'
Close tag   : 'document'
              'author' : 'frank'

Download

XMLReadLib.rar