{float:right|background=#eee}
{contentbylabel:title=Related Articles|showLabels=false|showSpace=false|space=@self|labels=component-templates,components}
{float} |
Under Tapestry, a component template is the file that contains the markup for that page or component.
Component templates are well formed XML documents. That means that every open tag must have a matching close tag, every attribute must be quoted, and so forth.
|
At runtime, Tapestry parses the documents and only checks for wellformedness. Even when the document has a DTD or schema, there are no validity checks. |
For the most part, these templates are standard HTML/XHTML; Tapestry extensions to ordinary markup are provided in the form of a Tapestry namespace.
We'll cover the specific content of templates shortly, first a few details about connecting a component to its template.
Component templates are stored with the component class file. The files have a ".tml" extension (i.e., Tapestry Markup Language), and are stored in the same package as corresponding component class.
Under a typical Maven directory structure, the Java class for a component might be src/main/java/org/example/myapp/components/MyComponent.java. The corresponding template will be src/main/resources/org/example/myapp/components/MyComponent.tml.
Likewise, the Java class for a page might be src/main/java/org/example/myapp/pages/MyPage.java and the corresponding template will be src/main/resources/org/example/myapp/pages/MyPage.tml.
The template and the compiled class will be packaged together in the WEB-INF/classes folder of the application WAR.
For pages (not components), a second location will be searched: in the web application context. The location is based on the logical name of the page, in the previous example, the template would be MyPage.tml in the root folder of the web application.
A template on the classpath takes precedence over a file in the web application context.
|
Allowing pages to store their template in the web context is a feature that may go away at some point. It was included as a way for HTML designers to edit template directly and live preview the template directly, without executing the Tapestry application. This comes with a large number of limitations and leads to a false sense of security that a template that previews correctly will render properly (this is not always the case). |
In certain cases, Tapestry will simplify the the logical name of a page. For example, the page class org.example.pages.address.CreateAddress will be given a logical name of "address/Create" (the redundant "Address" is removed as a suffix). However, this only affects how the page is referenced in URLs; the template file will still be CreateAddress.tml, whether on the classpath, or as address/CreateAddress.tml (in the web context).
|
Tapestry actually creates multiple names for the name page: address/Create and address/CreateAddress are both synonymous. You can user either in Java code that refers to a page by name, or as the page parameter of a PageLink. |
Main Article: Localization
Templates are in much the same way as individual files of a component's message catalog: the effective locale is inserted into the name of the file. Thus a German users will see the content generated from MyPage_de.tml and French users will see content generated from MyPage_fr.tml. When no specific localization is available, the default location (MyPage.tml) is used.
|
It is necessary to enable support for a locale before Tapestry will attempt to localize to that locale. This requires configuration in your application module; if you are using the Tapestry Quickstart archetype, only locale "en" will be enabled by default. |
As mentioned above, component templates are well-formed XML documents. This means that if you want to use any HTML entities (such as \& \ < > ©), you must use an HTML or XHTML doctype in your template (starting in 5.3, this is more-or-less automatic, see notes below). If you choose to use (X)HTML doctypes in your templates, they will be passed on to the client in the resultant (X)HTML. Note that if your pages are composed of multiple components, each with a template, and each template contains a doctype declaration, only the first doctype encountered by the template parser will be passed on to the client.
It should also be noted that even though *X*HTML DTDs are valid XML DTDs, HTML DTDs aren't. This means that HTML doctypes cannot be used by XML parsers. Tapestry works around this limitation internally by using XHTML DTDs to parse templates that use HTML DTDs. This internal mapping is possible because XHTML 1.0 is nothing more than "a reformulation of the three HTML 4 document types as applications of XML 1.0," as per the W3C. Don't worry though – the original HTML 4 doctype will still be emitted to the client!
The following doctypes are the most common (X)HTML doctypes:
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"> <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN" "http://www.w3.org/TR/html4/strict.dtd"> <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd"> <!DOCTYPE html> |
That last one is for HTML5 and only works in Tapestry 5.2.5 and later versions. In earlier versions, Tapestry didn't support the HTML5 doctype directly (see TAP5-1040), but there's a partial work-around: just add the following to your page class (or your layout class, if you use one) instead of adding the doctype to your template (.tml) files:
/**
* Output the HTML5 doctype, as a work-around to https://issues.apache.org/jira/browse/TAP5-1040
*/
@SetupRender
final void renderDocType(final MarkupWriter writer) {
writer.getDocument().raw("<!DOCTYPE html>");
}
|
|
Tapestry 5.3 introduces two significant improvements to template Doctypes. A template without a <!DOCTYPE> is parsed as if it had the HTML Doctype ( A template with the HTML Doctype ( |
Component templates should include the Tapestry namespace, defining it in the root element of the template.
<html xmlns:t="http://tapestry.apache.org/schema/tapestry_5_3.xsd">
<head>
<title>Hello World Page</title>
</head>
<body>
<h1>Hello World</h1>
</body>
</html>
|
This defines the namespace using the standard prefix, "t:". The examples on this page all assume the use of the standard prefix.
For backwards compatibility, you may continue to use the old namespace URIs: http://tapestry.apache.org/schema/tapestry_5_0_0.xsd or http://tapestry.apache.org/schema/tapestry_5_1_0.xsd. However, the following elements added, as part of Tapestry 5.1, will not work with the 5_0_0.xsd:
The 5_3.xsd fixes some minor bugs in the 5_1_0.xsd, but is functionally equivalent.
Tapestry elements are elements defined using the Tapestry namespace prefix.
All other elements should be in the default namespace, with no prefix.
There are a certain number of Tapestry elements that act as template directives; beyond that, any element in the Tapestry namespace will be a Tapestry component.
In many cases, a component is designed to integrate its template with its container's template.
The <t:body> element is used to identify where, within a component's template, its body (from the container's template) is to be rendered.
Components have control over if, and even how often, their body is rendered.
The following example is a Layout component, that adds basic HTML elements around the page specific content:
<html xmlns:t="http://tapestry.apache.org/schema/tapestry_5_3.xsd">
<head>
<title>My Tapestry Application</title>
</head>
<body>
<t:body/>
</body>
</html>
|
A page would use this component as follow:
<t:layout xmlns:t="http://tapestry.apache.org/schema/tapestry_5_3.xsd"> My Page Specific Content </t:layout> |
|
In this example, t:layout creates a Layout component (more on this shortly). |
When the page renders, the page's template and the Layout component's template are merged together:
<html>
<head>
<title>My Tapestry Application</title>
</head>
<body>
My Page Specific Content
</body>
</html>
|
|
Tapestry 4 users will recognize the <t:body> element as a replacement for the RenderBody component. |
A <t:container> element contains markup without being considered part of the template. This is useful for components that render several top level tags, for example, a component that renders several columns within a table row:
<t:container xmlns:t="http://tapestry.apache.org/schema/tapestry_5_3.xsd">
<td>${label}</td>
<td>${value}</td>
</t:container>
|
This component only makes sense when used inside a <tr> element of its container's template.
Without the <t:container> element, there would be no way to create a valid XML document as the template, because XML documents must always have a single root element.
A <t:block> is a container of a portion of the component template. A block does not normally render; any component or contents you put inside a block will not ordinarily be rendered. However, by injecting the block you have precise control over when and if the content renders.
A block may be anonymous, or it may have an id (specified with the id attribute). Non-anonymous blocks may be injected into the component.
|
A <t:block> must be in the Tapestry namespace, but the id attribute should not be. This is different from components in the template, where the t:id attribute that defines the component id must be in the Tapestry namespace. |
Ids must be valid Java identifiers: start with a letter, and contain only letters, numbers and underscores.
|
This element has been deprecated starting in Tapestry 5.1 (but is still fully supported, for backwards compatibility). Use parameter namespaces instead. |
A <parameter> element is a special kind of block. It is placed inside the body of an embedded component. The block defined by the <parameter> is passed to the component. <parameter> includes a mandatory name attribute to identify which parameter of the component to bind.
Example:
<t:if test="loggedIn">
Hello, ${userName}!
<t:parameter name="else">
Click <a t:type="actionlink" t:id="login">here</a> to log in.
</t:parameter>
</t:if>
|
<t:content> marks a portion of the template as the actual template content; any markup outside the <t:content> element is ignored. This is useful for eliminating portions of the template that exist to support WYSIWYG preview of the template.
<t:content> elements may not nest.
Support for the <t:content> element was adding in Tapestry release 5.1. You must use the http://tapestry.apache.org/schema/tapestry_5_1_0.xsd or http://tapestry.apache.org/schema/tapestry_5_3.xsd namespace URI for content to be recognized (otherwise you will see an error about a missing "content" component).
Marks a portion of the template for removal; it is as if the remove element and everything inside it simply was not part of the template. This is used as a kind of server-side only comment (normal HTML/XML comments are included in a page render response), or to temporarily eliminate a portion of the template. As far as Tapestry is concerned, the contents of the <remove> element do not exist (including validating consistency between components defined or injected in the Java class and the template).
Support for the <t:remove> element was adding in Tapestry release 5.1. You must use the http://tapestry.apache.org/schema/tapestry_5_1_0.xsd or http://tapestry.apache.org/schema/tapestry_5_3.xsd namespace URI for remove to be recognized (otherwise you will see an error about a missing "remove" component).
Another option when rendering output is the use of expansions. Expansions are special strings that may be embedded in template bodies, and borrow some syntax from the Ant build tool.
Welcome, ${userId}!
|
Here, ${userId} is the expansion. In this example, the userId property of the component is extracted, converted to a string, and streamed into the output.
Expansions are allowed inside text, and inside attributes of ordinary elements, and component elements. For example:
<img src="${request.contextPath}/images/catalog/product_${productId}.png"/>
|
In this hypothetical example, the component class is providing a request property and a productId property, and these are being used inside the template to assemble the src attribute of the <img> element. This is component-like behavior without actual components.
Under the covers, expansions are the same as parameter bindings. The default binding prefix for expansions is "prop:" (that is, the name of a property or a property expression), but other binding prefixes are useful, especially "message:" (to access a localized message from the component's message catalog).
|
Do no use expansions in component parameters if the parameter's default or explicit binding prefix is "prop:" or "var:". Expansions convert the value to an immutable string, resulting in a runtime exception if the component tries to update the value. Even for read-only parameters, expansions are not as desirable, since they always convert to a string, and from their to the parameter's declared type. |
|
Tapestry 4 users will note that expansions are a concise, easy replacement for the Insert component, and for the <span key="..."> directive. |
An embedded component is identified within the template as an element in the t: namespace. Example:
You have ${cartItems.size()} items in your cart.
<t:actionlink t:id="clear">Remove All</t:actionlink>.
|
The element name, "actionlink" is used to select the type of component, ActionLink.
|
As elsewhere, Tapestry is insensitive to case when mapping from a component type to a component class. |
Embedded components may have two Tapestry-specific parameters:
These attributes are specified inside the t: namespace (i.e., t:id="clear").
If the id attribute is omitted, Tapestry will assign a unique id for the element.
|
Non-trivial components should always be assigned a specific id, rather than letting Tapestry do it. You'll end up with shorter, more readable URLs and code that's easier to debug, because it will be more obvious how the request URL maps to your pages and components. This is even more strongly encouraged for form control components, where the component id will usually be the same as the query parameter that stores the value provided by the end user. |
Ids must be valid Java identifiers: start with a letter, and contain only letters, numbers and underscores.
Any other attributes are used to bind parameters of the component. These may be formal parameters or informal parameters. Formal parameters will have a default binding prefix (usually "prop:"). Informal parameters will be assumed to be literals (i.e., the "literal:" binding prefix).
Use of the t: prefix is optional for all other attributes. Some users implement a build process where the Tapestry template files are validated ... in that case, any Tapestry-specific attributes, not defined by the underlying DTD or schema, should be in the Tapestry namespace, to avoid validation errors.
The open and close tags of a Tapestry component element define the body of the component. It is quite common for additional components to be enclosed in the body of another component:
<t:form> <t:errors/> <t:label for="userId"/> <t:textfield t:id="userId"/> <br/> <t:table for="password"/> <t:passwordfield t:id="password"/> <br/> <input type="submit" value="Login"/> </t:form> |
In this example, the <t:form> component's body contains the other components. Structurally, all of these components (the Form, Errors, Label, etc.) are peers: children of the page. They are all embedded within the page. The Errors, Label, TextField and so forth are enclosed within the Form's body ... meaning that the Form controls if, when, and under what circumstances those components will render.
In some cases, components require some kind of enclosure; for example, all of the field control components (such as TextField) will throw a runtime exception if not enclosed by a Form component.
It is possible to place Tapestry components in sub-packages. For example, your application may have a package org.example.myapp.components.ajax.Dialog. This component's normal type name is "ajax/dialog" (because it is in the ajax sub-folder). This name is problematic, as it is not valid to define an XML element with an element name <t:ajax/dialog>. Instead, just replace the slashes with periods: <t:ajax.dialog>. Library namespaces (described in the next section) are a preferred way of handling components in sub-folders.
|
When using a component library, the library will map its components to a virtual sub-folder of the application. The same naming mechanism works whether its is a real sub-folder or a component library sub-folder. |
If you are using many components from a common Tapestry component library, you can use a special namespace to simplify references to those components.
The special namespace URI tapestry-library:path can be defined; the path is a prefix used in conjunction with component element names.
Borrowing from the above example, all of the following are equivalent:
<html xmlns:t="http://tapestry.apache.org/schema/tapestry_5_3.xsd" xmlns:a="tapestry-library:ajax"> <t:ajax.dialog/> <span t:type="ajax/dialog"/> <a:dialog/> . . . |
In other words, the virtual folder, ajax, defined in the namespace URI takes the place of the path prefixes ajax. seen in the first element, or ajax/ seen in the second. As far as Tapestry is concerned, they are all equivalent.
A favorite feature of Tapestry 4 was invisible instrumentation, the ability to mark ordinary HTML elements as components. Invisible instrumentation leads to more concise templates that are also more readable.
For Tapestry 5, you make use of namespaced id or type attributes to mark an arbitrary element as a component, for example:
<p>
Merry Christmas:
<span t:type="Count" end="3">
Ho!
</span>
</p>
|
The id, type and mixins attributes must be placed in the Tapestry namespace. Any additional attributes may be in the Tapestry namespace or in the default namespace. Placing an attribute in the Tapestry namespace is useful when the attribute is not defined for the element being instrumented.
A component must have a type, either via the t:type attribute in the template, or by the defining the component in the Java class using the Component annotation (and using the t:id attribute on the element in the template). The Component annotation is attached to a field; either the type of the field, or the type attribute of the Component annotation, determines the type of component.
In most cases, it is an aesthetic choice between normal embedded components, and embedded components via invisible instrumentation. In a few instances, such as the Loop component, the behavior of the component is influenced by your choice. The Loop component, when included using invisible instrumentation, will render the tag and any informal parameters, around its body. Thus, for example:
<table>
<tr t:type="loop" source="items" value="item" class="prop:rowClass">
<td>${item.id}</td>
<td>${item.name}</td>
<td>${item.quantity}</td>
</tr>
</tabel>
|
Here, the loop component "merges into" the <tr> element. It will render out a <tr> for each item object in the items list. It will write a dynamic class attribute into each <tr>.
Parameter namespaces are a new feature introduced in Tapestry 5.1. They are a more concise way of passing parameter blocks to components.
You must define a special namespace, usually given the prefix "p":
<html xmlns:t="http://tapestry.apache.org/schema/tapestry_5_3.xsd" xmlns:p="tapestry:parameter"> . . . |
With the "tapestry:parameter" namespace defined, you can pass block using the "p:" prefix and an element name that matches the parameter name:
<t:if test="loggedIn">
Hello, ${userName}!
<p:else>
Click <a t:type="actionlink" t:id="login">here</a> to log in.
</p:else>
</t:if>
|
This example passes a block of the template (containing the ActionLink component and some text) to the If component as parameter else. In the If component's reference page you'll see that else is type Block.
Name-spaced parameter elements are not allowed to have any attributes. The element name itself is used to identify the parameter of the component to bind.
Tapestry strips out unnecessary whitespace from templates as they are parsed. Inside any block of text, repeated whitespace is reduced to a single space character. Blocks of text that are entirely whitespace, such a line break and whitespace between two tags, is eliminated entirely.
If you do a view source on the rendered output, you'll see that the bulk of the rendered page is one long unbroken line.
This approach has certain efficiency advantages on both the server (less processing to render the page) and on the client (fewer characters to parse). Tools such as FireBug and Chrome Developer Tools are useful for allowing you to view the rendered HTML on the client properly.
In rare cases, the whitespace in a template is significant. Perhaps you are creating a <pre> (preformatted) block of text, or the whitespace interacts with your stylesheet to some desired effect.
You may use the standard XML attribute xml:space to indicate to Tapestry whether whitespace should be compressed (xml:space="default") or preserved (xml:space="preserve"). Such attributes are stripped out by the template parser; they do not appear in the rendered output.
The xml: namespace prefix is built into all XML documents, there is no special configuration (as there is with the Tapestry namespace).
For example:
<ul class="navmenu" xml:space="preserve">
<li t:type="loop" t:source="pages" t:value="var:page">
<t:pagelink page="var:page">${var:page}</t:pagelink>
</li>
</ul>
|
This will preserve the whitespace between the <ul> and <li> elements, and between the (rendered) <li> elements and the nested <a> elements. For example, the output may look something like:
<ul>
<li>
<a href="showcart>ShowCart</a>
</li>
<li>
<a href="viewaccount">ViewAccount</a>
</li>
</ul>
|
With normal whitespace compression, you would see the following rendered output:
<ul><li><a href="showcart">ShowCart</a></li><li><a href="viewaccount">ViewAccount</li></ul> |
You can even put further xml:space attributes inside nested elements to fine-tune the control over what whitespace is preserved and what is compressed.
If a component does not have a template, but extends from a component class that does have a template, then the parent class' template will be used by the child component.
This allows a component to extend from a base class but not have to duplicate the base class' template.
Tapestry 5.1 adds a significant new feature: template inheritance with extension points. Previously, a component which extended another component had to inherit the parent component's entire template, or copy-and-paste the template.
Parent template can now mark replaceable sections as <t:extension-point>s, and sub-components can extend the parent template and <t:replace> those sections.
This can work across multiple levels of inheritance.
|
Overuse of this feature is not recommended: in general use of composition, rather than inheritance, will be easier to understand and maintain. There are certain specific cases where overrides will allow a for much wider and easier reuse of a component, but the component needs to be carefully designed to support this kind of inheritance properly. |
Marks a point in a template that may be replaced. A unique id (case insensitive) is used in the template and its sub-templates to link extension points to possible overrides.
<t:extension-point id="title">
<h1>${defaultTitle}</h1>
</t:extension-point>
|
Root element of a child template that extends from its parent template. The <t:extend> attribute may only appear as the root element and may only contain <t:replace> elements.
Replaces an extension point from a parent template. <t:replace> may only appear as the immediate child of a root <t:extend> element.
<t:extend xmlns:t="http://tapestry.apache.org/schema/tapestry_5_3.xsd">
<t:replace id="title">
<h1><img src="${context:images/icon.jpg}"/>
Customer Service</h1>
</t:replace>
</t:extend>
|