Reference 8

From PowerMops
Revision as of 04:42, 1 January 2009 by Jtittsler (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search


Modules

Mops provides a facility for creating separately compilable, relocatable modules. Modules encourage you to separate your program into well-defined, independent units, which makes your code easier to understand and maintain. You must explicitly state which definitions from a given module will be available (exported) to callers outside the module. Any other definitions become unavailable to the rest of the program after the module is compiled. Modules are loaded automatically on the heap whenever one of the exported definitions is referenced. The application must manage and release any modules that are no longer needed.

Module Guidelines

Certain guidelines must be observed when dividing your application into modules. Exported names are taken to be executable words. Therefore you mustn't export other things such as objects or Values. If you need to export an object, you will need to define a word in the module which gets the address of the object (just naming it will do that), then export that word name. You can then use late binding to send a message to the object.

You may export class names. Thus, you can define a class in a module, include the class name in the imports list, and instantiate objects of that class anywhere. In fact in the Mops system itself, we handle windows, menus and dialogs this way. All the methods of the class are in effect exported along with the class name; you don't have to take any special action, apart from putting the class name in the imports list, and taking care with action handlers (see below). Whenever you send a message to an exported class, the module will be invoked automatically.

Naturally this has some performance implications. There is a fair amount of overhead involved in invoking a module. You can reduce this to some extent by locking the module over a number of calls (this is true for ordinary exported words as well). But if message execution for a particular class is really time-critical, it would really best to leave the class in the main dictionary. Where exported classes are most useful is for those classes that depend heavily on Toolbox calls for most of their methods (such as windows, menus and dialogs). Toolbox calls are generally much slower than Mops module invocations, so the extra time penalty of putting the class into a module won't be significant.

In Mops you can enter modules only through the exported words (otherwise a machine base address register will not be set correctly, and you'll crash). Therefore a module should not store the address of one of its internal words in a vector in the main dictionary. Even if the module is locked in memory, you would not be able to execute this vector. We have included a check on stores to vectors, that a module address isn't being stored outside the module, so as to give an error message if this is attempted. This check can be turned off, but don't do it unless you are absolutely sure you know what you are doing!

How to Use Modules

Each module has a corresponding object in the main dictionary, which is used for interaction with the module, and which conceptually is the module. These module objects have appropriate methods to load them, purge them, query their status etc.

Creating modules in Mops is a three-stage process:

  1. Create a definition for the module and the entry points that are to be available to the rest of the application. This must exist in the resident portion of the application, and has the following format:
    FROM MyMod IMPORT{ word1 word2 word3 }
    This statement declares a module, MyMod, from which will be imported three definitions, word1, word2 and word3. These two names will exist only in the disk image of the module until one of them is referenced, at which time the entire module will be loaded into the heap. On the disk, the binary image of the module will have the name MyMod.bin (68k Mops) or MyMod.pbin (PowerMops).
  2. Write the source code for the module in a separate file, called whatever is the name of your module, followed by '.txt'. Thus in the above example the source file would be called MyMod.txt.
  3. The module must be compiled and saved in its binary format before it will be available to callers. To compile a module, you must send a compile: message to the module, e.g.
    compile: MyMod
    When the module has been compiled, a message will appear stating that the module has been saved. You must have room in your dictionary to load the module source file.
    When we compile a module, we temporarily hide all of the main dictionary above the module object itself. This means you don't have to worry about what you might have loaded in the dictionary when you want to recompile a module.

If you need to call a module several times in succession, you can save some runtime overhead by locking it in memory while you are making the calls. Do it thus:

lock: MyMod

You will also need to lock the module if you get the address of somewhere in the module — an object, say — and need to use the address again later. If the module isn't locked, it may move in the heap in the meantime or be removed altogether, so that the address won't be valid any longer. When you are finished with the module, remember to unlock it thus:

unlock: MyMod

When you are completely finished with a module, you can release its heap memory thus:

release: MyMod

You do not have to do this, however, since loaded modules which are not actually being executed will be released automatically if more heap space is needed. If you want a module not to be released from memory, you can send the message

keep: MyMod

This will flag a module as needing to be kept in memory. Unlike the situation with lock:, the module may be moved within the heap by the Memory Manager, if it isn't actually being executed. To undo a keep:, send

drop: MyMod

This will not release the module from memory, but will once again allow it to be released if more heap space is needed for something else.

To include a module in an installed application, mark it as installable any time before calling install. Do it like this:

true setInstall: MyMod

Reference 7 Reference Reference 9
Documentation