A list of global variable names that will be compiled into Myghty templates, and will be expected to be provided upon template execution. The variable "r", representing the Apache Request, is delivered to templates via this parameter. To add your own variables, specify their names in this list, and also specify it as a key in the RequestImpl argument global_args to provide the values. See the section Make your Own Friends for details.

Whether or not m.write calls will be automatically flushed to the final output stream as they are called, or will be buffered until the end of the request. Autoflush being on makes things a little more tricky, as you can't do any kind of page redirects (neither external nor internal) once the content starts coming out. Error messages become messy as well since they are similar to an internal redirect.

Autoflush can be set in a variety of ways and is described fully in the section The Autoflush Option.

If pointing to a Python file object, container operations performed by the caching system will be logged, allowing the viewing of how often data is being refreshed as well as how many concurrent threads and processes are hitting various containers. When running with ApacheHandler or CGIHandler, this parameter can be set to the standard Apache log via the parameter log_cache_operations.

An instance of myghty.compiler.Compiler used to generate object files. Not much reason to play with this unless you are seriously hacking your own engine. The default Compiler object will receive its optional parameters from the argument list sent to the constructor for Interpreter.

Directory to be used to store compiled object files (.py files and .pyc files), lock files used for synchronization, as well as container files which store data for the component caching system and the session object. A non-None value implies True for the "use_object_files" setting. Compiled object files are stored in the obj directory, and cache/session files are stored underneath container_XXX directories.

References a Python file object; if non-None, the Interpreter will send debugging information to this file. Note that line breaks are not sent by default; use code such as myghty.buffer.LinePrinter(sys.stderr) to wrap the file in a line-based formatter.

When running the Interpreter in an HTTP context, various implementations of HTTPHandler supply their own stream for debugging information, such as the ApacheHandler which supplies the Apache error log as a filehandle.

As of version 0.97alpha3, to actually enable debug logging requires the debug_elements parameter to be set up.

The name of the file used as a directory handler, defaults to 'dhandler'. See the section dhandlers for information on directory handlers.

If auto_flush is turned on, this option will keep auto_flush turned off for components that use <%filter> sections. This is useful if you have filters that rely upon receiving the entire block of text at once. See the section <%filter> for more information on filters.

A function object that will be passed all exceptions and given the chance to handle them before regular processing occurs. The format of the function is:

def handle_error(e, m, **params):
    # custom error handling goes here
    # ...

    # return False to continue handling error, True to disregard
    return False

Where 'e' is a myghty.exception.Error object and m is the request object. The global variable 'r' will also be passed if running in an HTTP context. The function should return True to stop all further error handling, or False to continue error handling as normal.

The request object, while it is the same instance used to handle the request initially, no longer will have any buffered content and will also not have a current component instance. It is safe to call subrequests from it, allowing the construction of custom error pages.

See also raise_error to simply raise an error outside the Request and bypass all internal and custom error handling.

An instance of myghty.objgen.ObjectGenerator that is used to generate object files. Again, for the brave hacker, you can write your own generator and output object files in whatever language you like.

Specifies the name of the Myghty interpreter that should be used. All HTTP handlers modules maintain a dictionary of HTTPHandler instances, each of which references a Myghty interpreter, keyed off of a name. When this name is explicitly specified, that particular handler will be created if it doesnt exist and then used.

This option is currently most useful for use within multiple Apache directives, so that different sets of configuration parameters can be used for different directives, without requiring the use mod_python's multiple Python interpreters feature. See the example in Advanced mod_python Configuration - Multiple ApacheHandlers.

Deprecated; use debug_elements instead.

Specifies whether or not component errors should receive a full stack trace in the Apache error log, standard error output of CGIHandler, or other HTTP-specific logging system. If false, component errors still produce a single-line description of the error in the log. See also output_errors. If raise_error is set to true, no logging of errors occurs.

This parameter is used for resolving Module Components using regular expressions which are mapped to function, class or class instance objects. See module_components for examples.

This parameter is used in combination with module_root to specify a callable that will translate an incoming URI before being resolved to a path-based module. See module_root for more detail.

Sets the initial value for the output_encoding of requests. The default value is python's system default encoding (usually ASCII.) See the section on Unicode Support for further details.

This parameter specifies the parent request when making subrequests, and is automatically provided. For more information see Programmatic Interface.

Used by module_root to specify default path tokens that should be searched in file paths. See Module Root Options for details.

References a function that will be invoked with the full text of each individual unit of Python code created during component generation (i.e., creation of .py files). The returned string will be used as the Python code that is actually placed within the generated .py file. Also see text_post_processor.

Indicates if errors should be raised when they occur, or be handled by error handling functionality. If set to True, no error logging, friendly client response, or custom internal error handler function will happen. Instead, an external error handler can be used. This parameter overrides invalidates the functionality of log_errors, output_errors and error_handler.

This parameter allows the entire interpreter.execute() or handler.handle() call to be wrapped in a try: / except: block, where the error can be handled externally to all Myghty functionality.

An instance of myghty.request.RequestImpl that will be used either as a prototype to create new RequestImpls per request, or can be passed per individual interpreter execution. RequestImpl tells Request how it should interact with its environment. Take a look at ApacheHandler.ApacheRequestImpl, CGIHandler.CGIRequestImpl, and myghty.request.DefaultRequestImpl for examples.

Indicates that callables located via module_root resolution require an attribute named 'public' set to True attached to them, in order to allow access. See Module Root Options for details.

Allows configuration of the rules used in component resolution. See the section Advanced Resolver Configuration for details.

References a function that will be invoked with the full text of each individual unit of output text created during component generation (i.e., creation of .py files). The returned string will be used as the output text that is actually placed within the generated .py file. Also see python_post_processor.

Whether or not to use directory handlers. See the section dhandlers for more information.

The additional global variable s to reference the Myghty session, or alternatively the mod_python 3.1 session, is turned on when this option is set to True. See the section Session for details.

This parameter, when set to True:

• Enables URI caching within the resolver, and/or turns on all configured URICache() resolution rules, so that visiting a URI a second time does not produce any filesystem checks, within the currently cached set of URIs.
• Disables repeated last-modification time checks on all template files and module-component files
• Disables last-modification time checks on compiled template object files against the template they were generated from, which in effect disables re-compilation of templates even across server restarts, unless the object file is removed from the obj directory underneath the configured data_dir directory