<html>
<body>
<h1>Validation Framework</h1>
<h3>Dependency Support</h3>
<p>Sometimes resources depend on other resources, to determine whether they are 
valid or not. An example would be that an XML resource could depend on a DTD. If at some 
point in the future the DTD changes, then the dependent XML resource should be 
revalidated to ensure that it still meets the requirements of the changed DTD 
file. In order to do this efficiently, there needs to be an association between 
the DTD and all the resources that depend on it.<p>To be as efficient as possible, the validation framework allows 
this dependency information to be returned as part of a
<a href="AbstractValidator.html#validate(org.eclipse.core.resources.IResource, int, ValidationState, org.eclipse.core.runtime.IProgressMonitor)">
validate operation</a>. Since the validator needed to figure this out anyways, 
why not pass this information back to the framework, so that it can be saved. 
This saves the cost of having to parse the file at some later point in time 
(when the file is probably sitting cold on disk). The fewer times that we need 
to open, read and/or parse the same file, will improve the performance of the 
product.
<p>In particular when the validate method returns, a
<a href="ValidationResult.html">ValidationResult</a> must be returned, inside of 
this result, all your dependencies need to be specified with the&nbsp;
<a href="ValidationResult.html#setDependsOn(org.eclipse.core.resources.IResource[])">
setDependsOn</a> method.<p>In addition to keeping the dependency information 
up-to-date through the normal course of validation, direct access is also 
provided to the dependency support through the
<a href="ValidationFramework.html#getDependencyIndex()">getDependencyIndex</a> 
method. The <a href="IDependencyIndex.html">IDependencyIndex</a> allows you to 
directly manipulate the dependency information for a validator.
</body>
</html>