Path: /sdk/add_on/scriptbuilder/
This class is a helper class for loading and building scripts, with a basic pre-processor that supports conditional compilation, include directives, pragma directives, and metadata declarations.
By default the script builder resolves include directives by loading the included file from the relative directory of the file it is included from. If you want to do this in another way, then you should implement the include callback which will let you process the include directive in a custom way, e.g. to load the included file from memory, or to support multiple search paths. The include callback should call the AddSectionFromFile or AddSectionFromMemory to include the section in the current build.
If the application should to support pragma directives, it must register the pragma callback to process the text provided for the pragma directive. Without the callback the script builder will give an error when encountering a pragma directive.
The script builder will also remove any lines that start with #!
as comments. This is done to support shebang interpreter directives often used on Linux and UNIX based operative systems.
If you do not want process metadata then you can compile the add-on with the define AS_PROCESS_METADATA 0, which will exclude the code for processing this. This define can be made in the project settings or directly in the header.
Example script with include directive:
#include "commonfuncs.as" void main() { // Call a function from the included file CommonFunc(); }
The builder supports conditional programming through the #if/#endif preprocessor directives. The application may define a word with a call to DefineWord(), then the scripts may check for this definition in the code in order to include/exclude parts of the code.
This is especially useful when scripts are shared between different binaries, for example, in a client/server application.
Example script with conditional compilation:
class CObject { void Process() { #if SERVER // Do some server specific processing #endif #if CLIENT // Do some client specific processing #endif // Do some common processing } }
Metadata can be added before script class, interface, function, and global variable declarations. The metadata is removed from the script by the builder class and stored for post build lookup by the type id, function id, or variable index.
Exactly what the metadata looks like is up to the application. The builder class doesn't impose any rules, except that the metadata should be added between brackets []. After the script has been built the application can obtain the metadata strings and interpret them as it sees fit. Multiple metadata blocks, i.e. [], can be defined for each entity.
Example script with metadata:
[factory func = CreateOgre] class COgre { [editable] vector3 myPosition; [editable] [range [10, 100]] int myStrength; } [factory] COgre @CreateOgre() { return @COgre(); }
Example usage: