W2ML Specification (Draft)

This element performs no processing by itself. It is mostly useful with generic W2ML attributes.

Hello <w2:g>world</w2:g>!

Hello world!

Configuring the base with this attribute overrides base deduction from document URI, notably in case of inclusion.

The URI must be absolute. An empty attribute value means that the base is deduced from the document URI.

 <div w2:base="site:///test/">
  ...
 </div>

I is an integer number.

If I is not negative, then only the first I characters are kept. If I is negative, then the first -I characters are deleted (no character deleted for tfirst="-0").

<w2:g tfirst="3">12345</w2:g>
<w2:g tfirst="-3">12345</w2:g>

123
45

I is an integer number.

If I is not negative, then only the last I characters are kept. If I is negative, then the last -I characters are deleted (no character deleted for tlast="-0").

<em w2:tlast="3">12345</em>
<em w2:tlast="-3">12345</em>

<em>345</em>
<em>12</em>

I is an integer number.

The string configured by tfiller is repeated until the number of characters in the outDOM text is |I|. If I is positive, then the text is filled at its beginning, otherwise at its end. If |I| is smaller or equal to the number of characters already present, then nothing is done.

<w2:g tfiller="x"
      twidth="5">123</w2:g>
<w2:g tfiller="ab"
      twidth="-6">123</w2:g>
<w2:g tfiller="x"
      twidth="-2">123</w2:g>

xx123
123aba
123

S must be a non-empty string of characters.

This attribute configures S as the filler string used by the twidth attribute. The default filler string is a single non-breaking space.

URI is a base URI.

URI gives the base location of image resources (typically a directory containing PNG or GIF files), each resource containing a glyph. Resources must be named after the Unicode code point of the character in uppercase hexadecimal digits, prefixed with 0 if necessary to be four digits long: 0041 for 'A', 00E9 for 'é', 20AC for '€'… A suffix can be configured with the timg2 attribute.

It should be noted that the URI is not used by the W2ML processor. The URI is copied as is in the outDOM, concatenated with the code point and the suffix, to give the resource location to the HTTP client. Typically, it is implemented with the HTML img element, and the URI goes into the src attribute.

<w2:g timg="fonts/"
      timg2=".png">Hi !</w2:g>

<img alt="H" src="fonts/0048.png"
><img alt="i" src="fonts/0069.png"
> <img alt="!" src="fonts/0021.png"
>

S is the end of an URI.

The default suffix is empty.

Supported zone values:

• local: local time zone of the processor;
• utc, UTC, gmt or GMT: universal time (UTC, also known as GMT).

<w2:g tz="utc">
 <w2:hour/>:<w2:minute/>
</w2:g>

 3:15

Supported source values:

• request: time at request reception;
• modification: last modification time, i.e. resource modification time of last modified resource included and processed in the inDOM, or request time if the inDOM has been modified by processing;
• doc modification: last modification time of the document containing this attribute, or request time if the document has been modified by processing;
• clock: time when the timeof attribute is processed, using a high-resolution clock if available.

<w2:g timeof="modification">
 Last modified:
 <w2:hour/>:<w2:minute/>.
</w2:g>

 Last modified:
 13:4.

<w2:year/>

2005

<w2:month/>

3

• digit: Months go from 1 to 12.
• abbr: Months go from Jan to Dec (locale dependent).
• name: Months go from January to December (locale dependent).

<w2:day fmt="abbr"/>

Sun

• digit: Days go from 1 to 31.
• abbr: Days go from Sun to Sat (locale dependent).
• name: Days go from Sunday to Saturday (locale dependent).

<w2:ampm/>

AM

<w2:hour/>
<w2:hour fmt="12"/>

13
1

• 24: Hours go from 0 to 23.
• 12: Hours go from 1 to 12.

<w2:minute/>

3

<w2:second/>

3

<w2:tz/>
<w2:tz fmt="num"/>

CET
+0100

• abbr: Output time zone name or abbreviation.
• num: Output time zone as hour offset from GMT.

<w2:time/><br/>
<w2:time fmt="rfc822"/>

Wed Jan 31 01:33:06 2007<br />
Wed, 31 Jan 2007 01:33:06 +0100

• local: Output date and time in local format (locale dependent).
• rfc822: Output time in an RFC 822 compliant format (as amended by RFC 1123).

Sleep duration is given by the concatenation of element's text nodes. The duration format is a natural number followed by the u (microsecond), s (second), m (minute), h (hour) or d (day) letter.

<w2:sleep>3</w2:sleep>

If inclusion fails, then the content of the include element is processed as fallback, otherwise it is not processed. Recursive inclusion and processing of the same resource fails.

If the URI scheme supports parameters, (for instance the exe scheme) then if include has got output child elements directing to arg or stdin, then output child elements are processed before inclusion. See exe scheme examples.

…bla.
<w2:include src="doc.w2ml">
 Inclusion failed
</w2:include>
Bla…

<w2:g xmlns:w2=
 "http://w2ml.org/2005/w2ml">
 <em>Included document.</em>
</w2:g>

…bla.
 <em>Included document.</em>
Bla…

URI is the URI of the resource to include. If the URI is empty, then the document includes itself (which is a forbidden recursive inclusion if it is included as W2ML).

Supported type values:

• application/x-w2ml+xml: The included resource is parsed and processed as a W2ML document;
• text/plain: The included resource is not parsed; it is used as text data.

The content of outclude is not immediately processed: the backclude element must be used in the included document (or a document included by the included document) to process it. It means that the content of outclude is normally processed as if it were the content of a backclude element in an included document. So, an outclude/backclude pair of elements includes a document "out of" or "around" the content of outclude.

A document included by outclude can itself use outclude. It works as if each outclude pushes its content on a stack and each backclude looks for content to process deeper in the stack.

Recursive inclusion of the same resource fails.

If the URI scheme supports parameters, then, if they are output child elements directing to arg or stdin, they are processed before inclusion.

…bla.
<w2:outclude
 src="border.w2ml">
 <em>Outcluded</em> content.
</w2:outclude>
Bla…

<p style="border:1ex ridge"
 xmlns:w2=
  "http://w2ml.org/2005/w2ml">
 <w2:backclude/>
</p>

…bla.
<p style="border:1ex ridge">
 <em>Outcluded</em> content.
</p>
Bla…

URI is the URI of the resource to include.

The content of the nearest outclude element ancestor that is not already included by an ancestor backclude is included and processed.

The content of the backclude element is processed when there is no content to include back, otherwise it is ignored.

Note that the recursive inclusion prohibition makes recursive backclusion impossible. But it is possible to use several sibling backclude elements in the same document, and they all include and process the content of the same outclude element.

Supported dest values:

• new: Output to new resource (new file). Require dst attribute.
• append: Output to end of resource. Require dst attribute.
• null: no output;
• stdin: Output to standard input of exe resource. To be used no more than once, only as a child of an element including a resource with the exe scheme.
• arg: Output to argument of exe resource. To be used only as a child of an element including a resource with the exe scheme. For each output to arg a new argument is added to the argument vector of the exe resource.

When the output is finished, the part of the outDOM that has been output is removed from the outDOM.

<w2:output to="new"
           dst="log">
 Text to new log file
</w2:output>

 Text to new log file

Supported output dest values are the same as for the to attribute.

<w2:output add="new"
           dst="log">
	   fmt="html">
Text to new<br/>log file
</w2:output>

Text to new<br />log file

This attribute may only be used for output destinations requiring an URI.

The use of href instead of dst is deprecated.

Supported format values:

• "": Output in XHTML or HTML syntax depending on outDOM output rules.
• "xhtml": Output tags in XHTML-compatible syntax.
• "html": Output tags in HTML-compatible syntax.
• "text": Output text only.

This attribute includes the W2ML document identified by URI and processes only users and groups elements, ignoring all other markup. Notably, generic attributes on users and groups elements are not processed. If URI is empty, then login includes its own document.

Login works like configuration: it is valid in the element and is inherited by its content, which can overrride it with other login attributes.

Once users and groups definitions are processed, login checks values of the w2mllog, w2mlusr and w2mlpwd request parameters. If a user is defined with the same user name and password, the visitor is given this user name, and the w2mlusr and w2mlpwd parameters are automatically saved in a cookie to provide them automatically to other documents of the site. If login fails, then the visitor is given a predefined user name. The user name, predefined or not, may be in one or several groups.

<html w2:login="passwd.w2ml">
…
</html>

If it contains a name element, then it declares and defines a user, otherwise it only declares the user which name is equal to the concatenation of its inDOM text.

If this element has a name element, then it may also contains a password element. Otherwise, it may not.

If this element is in, or contains, a group element, then the user is added to the group. A user element should not contain another user element.

<w2:user>
 <w2:name>Joe</w2:name>
 <w2:password>123</w2:password>
</w2:user>

<w2:group>
 <w2:name>admin</w2:name>
 <w2:user>Joe</w2:user>
</w2:group>

<w2:user>
 <w2:name>Tom</w2:name>
 <w2:group>admin</w2:group>
</w2:user>

The inDOM text of this element is concatenated to a user or group name.

This element should not be out of user and group elements.

The inDOM text of this element is concatenated to the user password. If this element is not present, then the user can log in without password.

This element should not be out of user element.

If it contains a name element, then it declares and defines a group, otherwise it only declares the group which name is the concatenation of its inDOM text.

A group element may contain other group and user elements, but only after its name element. It means that the user or group is added to the currently defined group.

Note: a group can not contain itself, even indirectly, because it would either require a declaration of the group before it is fully defined, or require a group redefinition.

(±)c… is a string of letters. The same letter should not appear several times in (±)c…. Each letter may be prefixed with the plus or minus character, or no character.

The next prefix characters are supported:

• none: set access, i.e. the access is given if the condition is met, otherwise it is suppressed;
• +: add access, i.e. the access is given if the condition is met;
• -: suppress access, i.e. the access is suppressed if the condition is met.

The next letters are supported:

• x: access to execute: without this access no result is produced; note that login attributes and attributes with a smaller or equal rank to ac are still executed;
• d: access to delete;
• n: access to create new instances;
• e: access to edit.

This attribute must be used with exactly one of the group, user, req, or internal attribute on the same element.

<w2:g ac="x" user="Joe 2">
 execute access only for Joe 2
</w2:g>
<w2:g ac="d+e-n" group="admins">
 give delete access only to admins,
 give edit access to admins,
 suppress instance access for admins
</w2:g>

• without match or attribute nmatch: if the parameter is defined;
• otherwise: if parameter value matches the regex.

• without match or attribute nmatch: if variable is defined;
• otherwise: if variable value match the regex.

ac condition is met if the parameter/variable is defined and matches regex; for multivalues, one match is enough.

ac condition is met if the parameter/variable is undefined or does not match regex; for multivalues, no match is allowed.

IDs is a spaces separated list of element IDs.

Only the first identified descendant element is processed. Filtering does not apply to process the element content.

elements is a spaces separated list of (namespace prefixed) element names.

Only the first descendant element matching a listed element is processed. Filtering does not apply to process the element content.

elements is a spaces separated list of (namespace prefixed) element names.

Only descendant elements matching listed elements are processed. Filtering does not apply to process the content of elements.

elements is a spaces separated list of (namespace prefixed) element names.

Non-listed descendant element tags are not processed, except acces control attributes. In effect, it removes some tags and keeps text nodes.

attrs is a spaces separated list of (namespace prefixed) attribute names.

Only listed generic and non-W2ML attributes of descendant elements are processed. In effect, it disables some attributes.

IDs is a spaces separated list of element IDs.

Only the first identified descendant element of a listed ID is not processed.

elements is a spaces separated list of (namespace prefixed) element names.

Only the first descendant element matching a listed element is not processed.

elements is a spaces separated list of (namespace prefixed) element names.

Descendant elements matching listed elements are not processed.

elements is a spaces separated list of (namespace prefixed) element names.

Descendant element tags matching listed elements are not processed. Their non-listed children and text nodes are processed.

attrs is a spaces separated list of (namespace prefixed) attribute names.

Listed attributes of descendant elements are not processed.

attrs-list is a well-formed list of attributes.

The attrs-list is parsed and its attributes are added to the element (rank=-7), processed, then removed from the element (rank=9). The once attribute is also removed.

XML namespace attributes are not supported.

attrs-list is a well-formed list of attributes.

Parse the attrs-list as an attributes list. Remove the next attribute. Add or update parsed attributes to the element.

susp-time is either a timestamp in RFC 822 format, or a duration the same format than the sleep element. When the value is a duration, then the request time is added to the duration and the result is saved in the inDOM as a timestamp. The processing is disabled if the timestamp is not in the past. The attribute is destroyed when the timestamp is for the first time in the past.

Note: timesuspe on an element disables its processing until the timestamp is in the past. timesuspt on a once element destroys the element when the timestamp is in the past, ie the time acts like an expire delay. timesuspt on an element with access control attributes allows an element to introduce new accesses when a timestamp is in the past.

This element increments its inDOM content, then write it to the outDOM. The element must be empty or its content must be text and represent an integer number. If the element is empty, then the counter starts with value 0.

<w2:counter>7</w2:counter>

8

<w2:counter>8</w2:counter>

The del element removes itself (and its content) from the inDOM.

<w2:del>bla</w2:del>

This element copies the outDOM to the inDOM.

<w2:res>
 <w2:counter>7</w2:counter>
</w2:res>

8

8

This element removes itself (and its content) from the inDOM after having been processed.

<w2:once>
 <w2:counter>7</w2:counter>
</w2:once>

8

This element removes its tags from the inDOM after having been processed. Its content is not processed. Of course, it will be processed next time.

<w2:next>
 <w2:counter>7</w2:counter>
</w2:next>

 <w2:counter>7</w2:counter>

Undo any modification made in its inDOM content during processing. Modifications made in included content are not undone.

<w2:undo>
 <w2:counter>7</w2:counter>
</w2:undo>

8

<w2:undo>
 <w2:counter>7</w2:counter>
</w2:undo>

Operation success can be tested by child conditional elements.

<w2:fdel dst="foo.w2ml"/>

Operation success can be tested by child conditional elements.

<w2:fcopy src="foo" dstpath="bar/">
<w2:a n="dstfile"
 ><w2:v req="fname"/></w2:a>
</w2:fcopy>

Operation success can be tested by child conditional elements.

<w2:fren src="foo" dst="bar"/>

Operation success can be tested by child conditional elements.

<w2:fsave src="foo" dst="bar"/>

This element processes its content if the processing of its nearest fallible ancestor was unsuccessful. The following elements may be unsuccessful: fdel, fcopy, fren, and fsave.

<w2:fdel src="foo">
 file
 <w2:ok>successfully</w2:ok>
 <w2:ko>not</w2:ko>
 copied.
</w2:fdel>

 file
 successfully
 copied.

This element processes its content if the processing of its nearest conditional ancestor was unsuccessful. The following elements may be unsuccessful: fdel, fcopy, fren, and fsave.

Supported editor values:

• "" or "none": No editor.
• Element editors (may be disabled by the e access control letter):
  • "text": Text level editor. Submitted data is saved in a text node.
  • "wysiwyg": In-browser WYSIWYG editor. Submitted data may only contain stylistic and semantic XHTML markup.
  • "html": HTML source level editor.
• Templates editors:
  • "new": Create new template instance.
    

    See to attribute, from and new attributes to identify the template and its instance destination.

    It may be disabled with the n access control letter.

    Suggested UI to start edition in the browser: onclick attribute.

  • "del": Delete element.
    

    The to attribute must be used to identify the element to delete.

    It may be disabled with the d access control letter.

    Suggested UI to start edition in the browser: onclick attribute.

The ed attribute should not be used on W2ML elements.

<p w2:ed="text">bla</p>

W2ML processor and editor specific.

Supported method values:

• "": Editor dependant.
• "auto": Automatically activate editors on load.

URI is an URI.

The edited element can be backcluded, but editing is automatically disabled to avoid an infinite recursive conflict.

id is an identifier. Two templates should not share the same id.

This attribute declares a template that can be referred by the from attribute. This attribute should only be used on non-W2ML elements.

<p w2:template="tpl">
 The template.
</p>

<button w2:ed="new"
        w2:to="tdest">
new
</button>

<div w2:new="tdest"
     w2:from="tpl"
     w2:to="lastChild">
...
</div>

<p w2:template="tpl">
 The template.
</p>

<button
  w2:ed="new"
  w2:from="tpl"
  w2:to="nextSibling">
 new
</button>

<p w2:del="me">
 A paragraph.
</p>

<button w2:ed="del"
        w2:to="me">
 new
</button>

id is an identifier. Two templates should not share the same id.

This attribute acts like the undo element and the instance attribute and edrun="auto" attributes. The template attribute provides a frozen template where modifications are undone. It is a classic template in the sense that is it cannot be changed. Its typical use is in an undisplayed XHTML division, so the uninstantiated template remains hidden for the visitor.

Unlike instance, template is not duplicated to the instantiated element. It means that modifications take effect when the template is instantiated.

If present on the same element than the instance attribute, then it takes precedence.

id is an identifier.

This attribute identifies destinations of the new editor. It is also added by the ed attribute on its element with a random value when the to attribute value is relative and no new attribute is present.

id is an identifier.

This attribute identifies deletable elements, i.e. destinations of the del editor.

id is an identifier.

This attribute is required by the new editor to identify the template source. The from attribute must be present on the destination element, i.e. the element with ed="new" when the to attribute value is relative, or the element identified by to when the to attribute value is absolute.

This attribute is required by the new and del editors.

The dest value can be a relative or absolute destination. An absolute destination is a reference to a new or del attribute identifier. A relative destination is one of the following dest values:

• firstChild: first child of the element.
• lastChild: last child of the element.
• previousSibling: previous sibling of the element.
• here: the element.
• nextSibling: next sibling of the element.
• lastSibling: last sibling of the element.

For ed="new", it gives the template instantiation destination. If it is relative, then the destination is relative to the current element. If it is absolute, then it identifies a destination element with a new attribute that must have another to attribute with a relative destination.

For ed="del", the to attribute destination must always be absolute.

A variable is a GET or POST HTTP request parameter, or a cookie, or an internal variables of the W2ML processor.

When a variable is multi-valued, each value is written in a span element, and elements are separated with a whitespace.

You are <w2:v req="age"/>.

You are 21.

Supported internal variable name values:

• "uri": URI of the processed W2ML document
• "parent uri", "parent parent uri"...: URIs of W2ML documents including the processed W2ML document
• "request method": the HTTP request method
• "request host": the HTTP request host (from Host header or full URI)
• "request uri": the HTTP request URI
• "request body": the HTTP request body
• "request content-type": the HTTP request Content-Type
• "request variables": parsed request parameters
• "all users": list of all defined users
• "all groups": list of all defined groups
• "user": user name of visitor
• "groups": group names of visitor
• "user ip": visitor IP address
• "server ip": server IP address
• "processor name": name of the W2ML processor; "interpreter name" is deprecated
• "processor version": version the processor; "interpreter version" is deprecated
• "processor build": build ID the processor; "interpreter build" is deprecated
• "processor full name": name, version and build of the W2ML processor; "interpreter full build" is deprecated
• "copyright": copyright
• "running time": running time of the software in seconds

Hello
<w2:v internal="user"/>
of IP
<w2:v internal="user ip"/>.

Hello
Joe
of IP
192.168.1.3.

The a element is processed by the file operations elements. It must be the first children of one of these elements. The outDOM of the a element is used as value of the attribute n of the parent element.

<w2:fcopy src="tpl.w2ml">
 <w2:a n="dst"><w2:v req="newPage"/></w2:a>
</w2:fcopy>

Supported level values:

• "": Reset to the default level.
• "emergency": Log when the system is unusable.
• "alert": Log errors that may cause emergency.
• "critical": Log errors that may affect several requests.
• "error": Log unrecovered errors that may cause unexpected behaviour.
• "warning": Log unexpected but recovered errors.
• "notice": Log interesting events.
• "information": Log information not normally interesting in logs.
• "debug": Log debugging informations.

The level values are inspired from the Apache HTTP Server LogLevel directive.

Events of the selected level and more critical levels are logged. The default level is taken from the server configuration.

Logging destination depends on the W2ML processor implementation. For instance, the HTTP server log files may be used.