Provided by: tclxml_3.3~svn11-3_amd64 
NAME
TclXML - XML parser support for Tcl
SYNOPSIS
package require xml
package require parserclass
::xml::parserclass option ? arg arg ... ?
::xml::parser ? name? ? -option value ... ?
parser option arg
DESCRIPTION
TclXML provides event-based parsing of XML documents. The application may register callback scripts for
certain document features, and when the parser encounters those features while parsing the document the
callback is evaluated.
The parser may also perform other functions, such as normalisation, validation and/or entity expansion.
Generally, these functions are under the control of configuration options. Whether these functions can
be performed at all depends on the parser implementation.
The TclXML package provides a generic interface for use by a Tcl application, along with a low-level
interface for use by a parser implementation. Each implementation provides a class of XML parser, and
these register themselves using the ::xml::parserclass create command. One of the registered parser
classes will be the default parser class.
Loading the package with the generic package require xml command allows the package to automatically
determine the default parser class. In order to select a particular parser class as the default, that
class' package may be loaded directly, eg. package require xml::libxml2. In all cases, all available
parser classes are registered with the TclXML package, the difference is simply in which one becomes the
default.
COMMANDS
::xml::parserclass
The ::xml::parserclass command is used to manage XML parser classes.
Command Options
The following command options may be used:
create
create name ? -createcommand script? ? -createentityparsercommand script? ? -parsecommand
script? ? -configurecommand script? ? -getcommand script? ? -deletecommand script?
Creates an XML parser class with the given name.
destroy
destroy name
Destroys an XML parser class.
info
info names default
Returns information about registered XML parser classes.
::xml::parser
The ::xml::parser command creates an XML parser object. The return value of the command is the name of
the newly created parser.
The parser scans an XML document's syntactical structure, evaluating callback scripts for each feature
found. At the very least the parser will normalise the document and check the document for well-
formedness. If the document is not well-formed then the -errorcommand option will be evaluated. Some
parser classes may perform additional functions, such as validation. Additional features provided by the
various parser classes are described in the section Parser Classes
Parsing is performed synchronously. The command blocks until the entire document has been parsed.
Parsing may be terminated by an application callback, see the section Callback Return Codes. Incremental
parsing is also supported by using the -final configuration option.
Configuration Options
The ::xml::parser command accepts the following configuration options:
-attlistdeclcommand
-attlistdeclcommand script
Specifies the prefix of a Tcl command to be evaluated whenever an attribute list declaration is
encountered in the DTD subset of an XML document. The command evaluated is: script name attrname
type default value
where:
name
Element type name
attrname
Attribute name being declared
type
Attribute type
default
Attribute default, such as #IMPLIED
value
Default attribute value. Empty string if none given.
-baseuri -baseurl
-baseuri URI
-baseurl URI
Specifies the base URI for resolving relative URIs that may be used in the XML document to refer
to external entities.
-baseurl is deprecated in favour of -baseuri.
-characterdatacommand
-characterdatacommand script
Specifies the prefix of a Tcl command to be evaluated whenever character data is encountered in
the XML document being parsed. The command evaluated is: script data
where:
data
Character data in the document
-commentcommand
-commentcommand script
Specifies the prefix of a Tcl command to be evaluated whenever a comment is encountered in the XML
document being parsed. The command evaluated is: script data
where:
data
Comment data
-defaultcommand
-defaultcommand script
Specifies the prefix of a Tcl command to be evaluated when no other callback has been defined for
a document feature which has been encountered. The command evaluated is: script data
where:
data
Document data
-defaultexpandinternalentities
-defaultexpandinternalentities boolean
Specifies whether entities declared in the internal DTD subset are expanded with their replacement
text. If entities are not expanded then the entity references will be reported with no expansion.
-doctypecommand
-doctypecommand script
Specifies the prefix of a Tcl command to be evaluated when the document type declaration is
encountered. The command evaluated is: script name public system dtd
where:
name
The name of the document element
public
Public identifier for the external DTD subset
system
System identifier for the external DTD subset. Usually a URI.
dtd
The internal DTD subset
See also -startdoctypedeclcommand and -enddoctypedeclcommand.
-elementdeclcommand
-elementdeclcommand script
Specifies the prefix of a Tcl command to be evaluated when an element markup declaration is
encountered. The command evaluated is: script name model
where:
name
The element type name
model
Content model specification
-elementendcommand
-elementendcommand script
Specifies the prefix of a Tcl command to be evaluated when an element end tag is encountered. The
command evaluated is: script name args
where:
name
The element type name that has ended
args
Additional information about this element
Additional information about the element takes the form of configuration options. Possible
options are:
-empty
boolean
The empty element syntax was used for this element
-namespace
uri
The element is in the XML namespace associated with the given URI
-elementstartcommand
-elementstartcommand script
Specifies the prefix of a Tcl command to be evaluated when an element start tag is encountered.
The command evaluated is: script name attlist args
where:
name
The element type name that has started
attlist
A Tcl list containing the attributes for this element. The list
of attributes is formatted as pairs of attribute names and their values.
args
Additional information about this element
Additional information about the element takes the form of configuration options. Possible
options are:
-empty
boolean
The empty element syntax was used for this element
-namespace
uri
The element is in the XML namespace associated with the given URI
-namespacedecls
list
The start tag included one or more XML Namespace declarations.
list is a Tcl list giving the namespaces declared. The list is formatted as pairs
of values, the first value is the namespace URI and the second value is the prefix
used for the namespace in this document. A default XML namespace declaration will
have an empty string for the prefix.
-encoding
-encoding value
Gives the character encoding of the document. This option only has an effect before a document is
parsed. The default character encoding is utf-8. If the value unknown is given, or any value
other than utf-8, then the document text is treated as binary data. If the value is given as
unknown then the parser will attempt to automatically determine the character encoding of the
document (using byte-order-marks, etc). If any value other than utf-8 or unknown is given then
the parser will read the document text using that character encoding.
-endcdatasectioncommand
-endcdatasectioncommand script
Specifies the prefix of a Tcl command to be evaluated when end of a CDATA section is encountered.
The command is evaluated with no further arguments.
-enddoctypedeclcommand
-enddoctypedeclcommand script
Specifies the prefix of a Tcl command to be evaluated when end of the document type declaration is
encountered. The command is evaluated with no further arguments.
-entitydeclcommand
-entitydeclcommand script
Specifies the prefix of a Tcl command to be evaluated when an entity declaration is encountered.
The command evaluated is: script name args
where:
name
The name of the entity being declared
args
Additional information about the entity declaration. An internal
entity shall have a single argument, the replacement text. An external parsed
entity shall have two additional arguments, the public and system indentifiers of
the external resource. An external unparsed entity shall have three additional
arguments, the public and system identifiers followed by the notation name.
-entityreferencecommand
-entityreferencecommand script
Specifies the prefix of a Tcl command to be evaluated when an entity reference is encountered.
The command evaluated is: script name
where:
name
The name of the entity being referenced
-errorcommand
-errorcommand script
Specifies the prefix of a Tcl command to be evaluated when a fatal error is detected. The error
may be due to the XML document not being well-formed. In the case of a validating parser class,
the error may also be due to the XML document not obeying validity constraints. By default, a
callback script is provided which causes an error return code, but an application may supply a
script which attempts to continue parsing. The command evaluated is: script errorcode errormsg
where:
errorcode
A single word description of the error, intended for use by an
application
errormsg
A human-readable description of the error
-externalentitycommand
-externalentitycommand script
Specifies the prefix of a Tcl command to be evaluated to resolve an external entity reference. If
the parser has been configured to validate the XML document, a default script is supplied that
resolves the URI given as the system identifier of the external entity and recursively parses the
entity's data. If the parser has been configured as a non-validating parser, then by default
external entities are not resolved. This option can be used to override the default behaviour.
The command evaluated is: script name baseuri uri id
where:
name
The Tcl command name of the current parser
baseuri
An absolute URI for the current entity which is to be used to
resolve relative URIs
uri
The system identifier of the external entity, usually a URI
id
The public identifier of the external entity. If no public
identifier was given in the entity declaration then id will be an empty string.
The return result of the callback script determines the action of the parser. Note that these
codes are interpreted in a different manner to other callbacks.
TCL_OK
The return result of the callback script is used as the external entity's data. The URI
passed to the callback script is used as the entity's base URI.
This is useful to either override the normal loading of an entity's data, or to implement
new or alternative URI schemes. As an example, the script below sets an external entity
handler that intercepts "tcl:" URIs and evaluates them as inline Tcl scripts:
package require xml
proc External {name baseuri uri id} {
switch -glob -- $uri { tcl:* { regexp {^tcl:(.*)$} $uri discard script
return [uplevel #0 $script] } default { return -code continue
{} }
} }
set parser [xml::parser -externalentitycommand External] $parser parse {<!DOCTYPE example [
<!ENTITY example SYSTEM "tcl:set%20example%20HelloWorld"> ]> <example>
&example; </example> }
puts $example
This script will print "HelloWorld" to stdout.
TCL_CONTINUE
In a normal (non-safe) interpreter, the default external entity handler is used to load the
external entity data as per normal operation of the parser. If the parser is executing in
a Safe Tcl interpreter then the entity is not loaded at all.
This is useful to interpose on the loading of external entities without interfering with
the loading of entities.
TCL_BREAK
No data is returned for this entity, ie. the entity is ignored. No error is propagated.
TCL_ERROR
No data is returned for this entity, ie. the entity is ignored. A background error is
registered, using the result of the callback script.
-final
-final boolean
Specifies whether the XML document being parsed is complete. If the document is to be
incrementally parsed then this option will be set to false, and when the last fragment of document
is parsed it is set to true. For example,
set parser [::xml::parser -final 0] $parser parse $data1 $parser parse $data2 $parser configure
-final 1 $parser parse $finaldata
-ignorewhitespace
-ignorewhitespace boolean
If this option is set to true then spans of character data in the XML document which are composed
only of white-space (CR, LF, space, tab) will not be reported to the application. In other words,
the data passed to every invocation of the -characterdatacommand script will contain at least one
non-white-space character.
-notationdeclcommand
-notationdeclcommand script
Specifies the prefix of a Tcl command to be evaluated when a notation declaration is encountered.
The command evaluated is: script name uri
where:
name
The name of the notation
uri
An external identifier for the notation, usually a URI.
-notstandalonecommand
-notstandalonecommand script
Specifies the prefix of a Tcl command to be evaluated when the parser determines that the XML
document being parsed is not a standalone document.
-paramentityparsing
-paramentityparsing boolean
Controls whether external parameter entities are parsed.
-parameterentitydeclcommand
-parameterentitydeclcommand script
Specifies the prefix of a Tcl command to be evaluated when a parameter entity declaration is
encountered. The command evaluated is: script name args
where:
name
The name of the parameter entity
args
For an internal parameter entity there is only one additional
argument, the replacement text. For external parameter entities there are two
additional arguments, the system and public identifiers respectively.
-parser
-parser name
The name of the parser class to instantiate for this parser object. This option may only be
specified when the parser instance is created.
-processinginstructioncommand
-processinginstructioncommand script
Specifies the prefix of a Tcl command to be evaluated when a processing instruction is
encountered. The command evaluated is: script target data
where:
target
The name of the processing instruction target
data
Remaining data from the processing instruction
-reportempty
-reportempty boolean
If this option is enabled then when an element is encountered that uses the special empty element
syntax, additional arguments are appended to the -elementstartcommand and -elementendcommand
callbacks. The arguments -empty 1 are appended. For example: script -empty 1
-startcdatasectioncommand
-startcdatasectioncommand script
Specifies the prefix of a Tcl command to be evaluated when the start of a CDATA section section is
encountered. No arguments are appended to the script.
-startdoctypedeclcommand
-startdoctypedeclcommand script
Specifies the prefix of a Tcl command to be evaluated at the start of a document type declaration.
No arguments are appended to the script.
-unknownencodingcommand
-unknownencodingcommand script
Specifies the prefix of a Tcl command to be evaluated when a character is encountered with an
unknown encoding. This option has not been implemented.
-unparsedentitydeclcommand
-unparsedentitydeclcommand script
Specifies the prefix of a Tcl command to be evaluated when a declaration is encountered for an
unparsed entity. The command evaluated is: script system public notation
where:
system
The system identifier of the external entity, usually a URI
public
The public identifier of the external entity
notation
The name of the notation for the external entity
-validate
-validate boolean
Enables validation of the XML document to be parsed. Any changes to this option are ignored after
an XML document has started to be parsed, but the option may be changed after a reset.
-warningcommand
-warningcommand script
Specifies the prefix of a Tcl command to be evaluated when a warning condition is detected. A
warning condition is where the XML document has not been authored correctly, but is still well-
formed and may be valid. For example, the special empty element syntax may be used for an element
which has not been declared to have empty content. By default, a callback script is provided
which silently ignores the warning. The command evaluated is: script warningcode warningmsg
where:
warningcode
A single word description of the warning, intended for use by an
application
wanringmsg
A human-readable description of the warning
-xmldeclcommand
-xmldeclcommand script
Specifies the prefix of a Tcl command to be evaluated when the XML declaration is encountered.
The command evaluated is: script version encoding standalone
where:
version
The version number of the XML specification to which this
document purports to conform
encoding
The character encoding of the document
standalone
A boolean declaring whether the document is standalone
Parser Command
The ::xml::parser command creates a new Tcl command with the same name as the parser. This command may
be used to invoke various operations on the parser object. It has the following general form: name
option arg
option and the arg determine the exact behaviour of the command. The following commands are possible
for parser objects:
cget
cget -option
Returns the current value of the configuration option given by option. Option may have any of
the values accepted by the parser object.
configure
configure -option value
Modify the configuration options of the parser object. Option may have any of the values
accepted by the parser object.
entityparser
entityparser option value
Creates a new parser object. The new object inherits the same configuration options as the parent
parser object, but is able to parse XML data in a parsed entity. The option -dtdsubset allows
markup declarations to be treated as being in the internal or external DTD subset.
free
free name
Frees all resources associated with the parser object. The object is not usable after this
command has been invoked.
get
get name args
Returns information about the XML document being parsed. Each parser class provides different
information, see the documentation for the parser class.
parse
parse xml args
Parses the XML document. The usual desired effect is for various application callbacks to be
evaluated. Other functions will also be performed by the parser class, at the very least this
includes checking the XML document for well-formedness.
reset
reset
Initialises the parser object in preparation for parsing a new XML document.
CALLBACK RETURN CODES
Every callback script evaluated by a parser may return a return code other than TCL_OK. Return codes
are interpreted as follows:
break Suppresses invocation of all further callback scripts. The parse method returns the TCL_OK
return code.
continue Suppresses invocation of further callback scripts until the current element has finished.
error Suppresses invocation of all further callback scripts. The parse method also returns the
TCL_ERROR return code.
default Any other return code suppresses invocation of all further callback scripts. The parse
method returns the same return code.
ERROR MESSAGES
If an error or warning condition is detected then an error message is returned. These messages are
structured as a Tcl list, as described below:
{domain level code node line message int1 int2 string1 string2 string3}
domain
A code for the subsystem that detected the error.
level
Severity level of the problem.
code
A one word string describing the error.
node
If available, the token of the DOM node associated with the problem.
line
If known, the line number in the source XML document where the problem was detected.
message
A human-readable description of the problem.
int1
Additional integer data. For the parser domain, this is usually the column number where the
problem was detected.
int2
Additional integer data.
string1
Additional string data.
string2
Additional string data.
string3
Additional string data.
APPLICATION EXAMPLES
This script outputs the character data of an XML document read from stdin.
package require xml
proc cdata {data args} {
puts -nonewline $data }
set parser [::xml::parser -characterdatacommand cdata] $parser parse [read stdin]
This script counts the number of elements in an XML document read from stdin.
package require xml
proc EStart {varName name attlist args} {
upvar #0 $varName var
incr var }
set count 0 set parser [::xml::parser -elementstartcommand [list EStart count]] $parser parse [read
stdin] puts "The XML document contains $count elements"
SAFE XML
TclXML/Tcl and TclXML/libxml2 may be used in a Safe Tcl interpreter. When a document is parsed in a Safe
Tcl interpreter, any attempt by the XML document to load an external entity is handled by the
-externalentitycommand callback. This callback is evaluated in the context of the safe interpreter and
therefore is subject to the security policy in force for that interpreter. The default entity loader
will not be invoked, even if the callback script returns a TCL_CONTINUE code.
See the description of the -externalentitycommand for further details.
PARSER CLASSES
This section will discuss how a parser class is implemented.
Tcl Parser Class
The pure-Tcl parser class requires no compilation - it is a collection of Tcl scripts. This parser
implementation is non-validating, ie. it can only check well-formedness in a document. However, by
enabling the -validate option it will read the document's DTD and resolve external entities. This
parser class is referred to as TclXML/tcl.
This parser implementation aims to implement XML v1.0 and supports XML Namespaces.
Generally the parser produces XML Infoset information items. That is, it gives the application a
slightly higher-level view than the raw XML syntax. For example, it does not report CDATA Sections.
TclXML/tcl is not able to handle character encodings other than UTF-8.
libxml2 Parser Class
The libxml2 parser class provides a Tcl interface to the libxml2 XML parser library. This parser class
is referred to as TclXML/libxml2.
When the package is loaded the variable ::xml::libxml2::libxml2version is set to the version number of
the libxml2 library being used.
On MS Windows, it is necessary to load the generic XML package first, and then the TclXML/libxml2
package. For example,
package require xml package require xml::libxml2
get Method
TclXML/libxml2 provides the following arguments to the get method:
document
Returns the parsed document object. libxml2 builds an in-memory data structure of the XML
document it parses (a DOM tree). This method returns a handle (or token) for that structure.
TclXML/libxml2 manages the document object as a Tcl object. See the -keep for further
information.
Additional Options
-keep
-keep normal | implicit
Controls how the TclXML/libxml2 packages manages the document object. The default value is
implicit; the document is destroyed when the Tcl Object's internal representation is freed. If
the option is given the value normal then the document must be explicit destroyed. The only way
to explicitly destroy the document is by using the C API.
-retainpath
-retainpath xpath
The given XPath location path specifies which part of the document is to be kept after the parsing
operation has completed. By default, all document data is discard after it has been parsed.
-retainpathns
-retainpathns prefix ns ...
The value of this option is a list of pairs of XML Namespace prefixes and their corresponding
namespace URIs. These are used by the XPath location path given in the -retainpath option.
Limitations
The libxml2 parser classes has the following limitations:
* -reportempty has no effect. libxml2 does not report empty element syntax.
* Incremental (push) parsing, ie. -final 0 is not supported.
* TclXML/libxml2 does not provide (DTD) validation, (WXS) schema validation or Relax NG
validation, although the libxml2 library does provide those functions. These functions are
provided by the TclDOM/libxml2 package, but only in a "posteriori" fashion (ie. only after the
document has been parsed).
* libxml2 supports XML Namespaces. The use of XML Namespaces can be queried, but the declaration
of a XML Namespace is not reported.