ResKnife/NuTemplateEditor/TEMPLATE EDITOR.txt

73 lines
3.7 KiB
Plaintext

ABOUT THE CODE
--------------
WHAT HAPPENS WHEN A TEMPLATE EDITOR IS OPENED FOR A RESOURCE
The template editor uses a dedicated class for each field type (or family of field types).
These classes are all subclasses of the common base class NuTemplateElement. For things like
lists, there is a subclass NuTemplateGroupElement from which you can instead subclass to
have some of the work done for you.
When opening a resource, the template editor first creates a NuTemplateStream for the template
resource and calls readOneObject: on it until it runs out of data, instantiating a hierarchy
of NuTemplateElement subclasses for based on the template.
After that, it creates a NuTemplateStream for the resource and loops over the template object
hierarchy, creating a copy of each item, and then calling readDataFrom:containingArray: on
the copy, which in turn reads data from the NuTemplateStream and stores it in its instance
variables. For this to work, subclasses of NuTemplateElement *must* implement the NSCopying
protocol.
This results in an object hierarchy that describes the entire resource using template fields
containing the data, which can now be edited, displayed in lists, etc.
WHAT HAPPENS WHEN THE TEMPLATE EDITOR IS CLOSED
As soon as the template editor is closed, another NuTemplateStream is created for the
resource and passed to all template fields (the copies with the resource data, not the
pre-parsed template hierarchy), via writeDataTo:, which is where they should write their
data to the stream. Before that, sizeOnDisk is called on each NuTemplateElement to
calculate the new size of the resource before the actual process of writing them out.
This size must include the size of all of their sub-items, and writeDataTo: must call
these sub items if they are to be written to disk.
SPECIAL CASE: LISTS
When the editor encounters an LSTB element, the LSTB element is called upon to parse its
"sub-elements" (the items that make up one list item). The LSTB element reads all elements
from the template, up to the terminating LSTE element, and keeps the array of parsed
sub-elements around in one of its data member variables. Parsing then resumes after the
"LSTE" element.
When the LSTB element is later called upon to read its data from the resource displayed in
the template, it will loop over its list of sub-elements and will generate a copy of itself
and all sub-elements except the "LSTE" for each list item. After all items have been read,
a single copy of the "LSTE" element is added at the end of the list. If the list is empty,
the "LSTB" will simply create an "LSTE" element and delete itself.
The "LSTE" element contains a copy of the sub-elements list of the LSTB, but doesn't read
any data from the resource. When either an "LSTE" or "LSTB" element are selected, the user
can choose "New List Item..." from the "Resource" menu to create a copy of the selected
element's sub-elements, along with an "LSTB" element containing them. Thus, the user can
create new items before any existing item (i.e. "LSTB" element), as well as at the end of
the list or when the list is empty, by selecting the "LSTE". "LSTB" items can also be
deleted using the "Edit" menu.
In short:
-> New list items are created by copying an existing "LSTB" and all its sub-elementss.
-> A single "LSTE" element serves as an "end marker" for full and as a placeholder for
empty lists, and can also be used to create new "LSTB"s with sub-elements.
Analogous behaviour is employed for other kinds of lists, except that additional
features in LSTEs are activated to allow writing terminating zeroes to the file, and
that the equivalent to the "LSTB" item may also update a counter field.
REVISIONS:
2003-08-13 UK Finished chapter on lists, added revision history.
2003-08-08 UK Created.