Config Schema

Filed under Development

Return to the index.

HTML Purifier End-User Documentation

HTML Purifier has a fairly complex system for configuration. Users interact with a HTMLPurifier_Config object to set configuration directives. The values they set are validated according to a configuration schema, HTMLPurifier_ConfigSchema.

The schema is mostly transparent to end-users, but if you're doing development work for HTML Purifier and need to define a new configuration directive, you'll need to interact with it. We'll also talk about how to define userspace configuration directives at the very end.

Write a directive file

Directive files define configuration directives to be used by HTML Purifier. They are placed in library/HTMLPurifier/ConfigSchema/schema/ in the form Namespace.Directive.txt (I couldn't think of a more descriptive file extension.) Directive files are actually what we call StringHashes, i.e. associative arrays represented in a string form reminiscent of PHPT tests. Here's a sample directive file, Test.Sample.txt:

Test.Sample
TYPE: string/null
DEFAULT: NULL
ALLOWED: 'foo', 'bar'
VALUE-ALIASES: 'baz' => 'bar'
VERSION: 3.1.0
--DESCRIPTION--
This is a sample configuration directive for the purposes of the
<code>dev-config-schema.html<code> documentation.
--ALIASES--
Test.Example

Each of these segments has a specific meaning:

Key	Example	Description
ID	Test.Sample	The name of the directive, in the form Namespace.Directive (implicitly the first line)
TYPE	string/null	The type of variable this directive accepts. See below for details. You can also add `/null` to the end of any basic type to allow null values too.
DEFAULT	NULL	A parseable PHP expression of the default value.
DESCRIPTION	This is a...	An HTML description of what this directive does.
VERSION	3.1.0	Recommended. The version of HTML Purifier this directive was added. Directives that have been around since 1.0.0 don't have this, but any new ones should.
ALIASES	Test.Example	Optional. A comma separated list of aliases for this directive. This is most useful for backwards compatibility and should not be used otherwise.
ALLOWED	'foo', 'bar'	Optional. Set of allowed value for a directive, a comma separated list of parseable PHP expressions. This is only allowed string, istring, text and itext TYPEs.
VALUE-ALIASES	'baz' => 'bar'	Optional. Mapping of one value to another, and should be a comma separated list of keypair duples. This is only allowed string, istring, text and itext TYPEs.
DEPRECATED-VERSION	3.1.0	Not shown. Indicates that the directive was deprecated this version.
DEPRECATED-USE	Test.NewDirective	Not shown. Indicates what new directive should be used instead. Note that the directives will functionally be different, although they should offer the same functionality. If they are identical, use an alias instead.
EXTERNAL	CSSTidy	Not shown. Indicates if there is an external library the user will need to download and install to use this configuration directive. As of right now, this is merely a Google-able name; future versions may also provide links and instructions.

Some notes on format and style:

Each of these keys can be expressed in the short format (KEY: Value) or the long format (--KEY-- with value beneath). You must use the long format if multiple lines are needed, or if a long format has been used already (that's why ALIASES in our example is in the long format); otherwise, it's user preference.
The HTML descriptions should be wrapped at about 80 columns; do not rely on editor word-wrapping.

Also, as promised, here is the set of possible types:

Type	Example	Description
string	'Foo'	String without newlines
istring	'foo'	Case insensitive ASCII string without newlines
text	"A\nb"	String with newlines
itext	"a\nb"	Case insensitive ASCII string without newlines
int	23	Integer
float	3.0	Floating point number
bool	true	Boolean
lookup	array('key' => true)	Lookup array, used with `isset($var[$key])`
list	array('f', 'b')	List array, with ordered numerical indexes
hash	array('key' => 'val')	Associative array of keys to values
mixed	new stdclass	Any PHP variable is fine

The examples represent what will be returned out of the configuration object; users have a little bit of leeway when setting configuration values (for example, a lookup value can be specified as a list; HTML Purifier will flip it as necessary.) These types are defined in library/HTMLPurifier/VarParser.php.

For more information on what values are allowed, and how they are parsed, consult library/HTMLPurifier/ConfigSchema/InterchangeBuilder.php, as well as library/HTMLPurifier/ConfigSchema/Interchange/Directive.php for the semantics of the parsed values.

Refreshing the cache

You may have noticed that your directive file isn't doing anything yet. That's because it hasn't been added to the runtime HTMLPurifier_ConfigSchema instance. Run maintenance/generate-schema-cache.php to fix this. If there were no errors, you're good to go! Don't forget to add some unit tests for your functionality!

If you ever make changes to your configuration directives, you will need to run this script again.

Errors

All directive files go through a rigorous validation process through library/HTMLPurifier/ConfigSchema/Validator.php, as well as some basic checks during building. While listing every error out here is out-of-scope for this document, we can give some general tips for interpreting error messages. There are two types of errors: builder errors and validation errors.

Builder errors

Exception: Expected type string, got integer in DEFAULT in directive hash 'Ns.Dir'

You can identify a builder error by the keyword "directive hash." These are the easiest to deal with, because they directly correspond with your directive file. Find the offending directive file (which is the directive hash plus the .txt extension), find the offending index ("in DEFAULT" means the DEFAULT key) and fix the error. This particular error would occur if your default value is not the same type as TYPE.

Validation errors

Exception: Alias 3 in valueAliases in directive 'Ns.Dir' must be a string

These are a little trickier, because we're not actually validating your directive file, or even the direct string hash representation. We're validating an Interchange object, and the error messages do not mention any string hash keys.

Nevertheless, it's not difficult to figure out what went wrong. Read the "context" statements in reverse:

in directive 'Ns.Dir': This means we need to look at the directive file Ns.Dir.txt
in valueAliases: There's no key actually called this, but there's one that's close: VALUE-ALIASES. Indeed, that's where to look.
Alias 3: The value alias that is equal to 3 is the culprit.

In this particular case, you're not allowed to alias integers values to strings values.

The most difficult part is translating the Interchange member variable (valueAliases) into a directive file key (VALUE-ALIASES), but there's a one-to-one correspondence currently. If the two formats diverge, any discrepancies will be described in library/HTMLPurifier/ConfigSchema/InterchangeBuilder.php.

Internals

Much of the configuration schema framework's codebase deals with shuffling data from one format to another, and doing validation on this data. The keystone of all of this is the HTMLPurifier_ConfigSchema_Interchange class, which represents the purest, parsed representation of the schema.

Hand-writing this data is unwieldy, however, so we write directive files. These directive files are parsed by HTMLPurifier_StringHashParser into HTMLPurifier_StringHashes, which then are run through HTMLPurifier_ConfigSchema_InterchangeBuilder to construct the interchange object.

From the interchange object, the data can be siphoned into other forms using HTMLPurifier_ConfigSchema_Builder subclasses. For example, HTMLPurifier_ConfigSchema_Builder_ConfigSchema generates a runtime HTMLPurifier_ConfigSchema object, which HTMLPurifier_Config uses to validate its incoming data. There is also an XML serializer, which is used to build documentation.