http://crystalspace3d.org/tikiwiki/tiki-index.php?page=ScriptChanges
which made me start to tinker with it myself. after reading this ml post again, i'm a bit confused about what's going on. obviously someone encountered the same problems i did with csPython and this is the result. i'm coming in late in the conversation. have any of these modifications been implemented?
On 2/16/06, Amir Taaki <[EMAIL PROTECTED]> wrote:
so are we agreed on this then? -
- new iScriptModule object, supporting overloaded Call
- internal hash of iScriptModule objects in iScript, which can be looked up using FindLoadedModule(const char*)
- Do we break the current LoadModule for python? (as far as I'm aware the RunText should still work, because loading the modules should include them in global interpreter - TODO: check this out). If it does break behaviour, should I precede or make a new LoadModuleObject or something
- AddModuleLoadingPath function to update PYTHONPATH variable (takes a const char of a VFS path which is internally translated to absolute path)
Alternatively we could do away with the iScriptModule object, and store the raw python module objects (merely adding an extra modulename argument to iScript::Call)
Just thought I'd make some pointers in case we followed the 'alternative' method...
- no new iScriptModule object
- Same Call functions, but function names must have unique identifiers since they all exist in global scope
- LoadModule stays, but also get LoadScriptFile function
- no need for AddModuleLoadingPath since LoadScriptFile takes a path to python script
Amir Taaki
