Well the new edbrowse almost works. I'm encouraged that it does anything at all. I really thought it would seg fault straight away. It's simpler in so many ways. There is only one line that enters the compartment, the entry point to processing the message. jseng-moz.cpp line 2543 Remember the old system, we had no idea what functions were called when, no idea the order of execution, so we put AutoCompartment in practically every function. No need for all that confusion any more.
So it mostly works, but not quite. I have a test program http://www.eklhad.net/jf browse it and it is missing half its stuff. But, in ebjs.c, after line 1014, add this has_property(cw->docobj, "all"); Then edbrowse works properly. jf has all the properties and objects that it should have. Why should that call matter? It's almost as inocuous as a debug print. I hate interactions like that. So I'm a bit discouraged to be this close and not there. I think I need to take a break from this for a while. We've made really good progress though. Karl Dahlke _______________________________________________ Edbrowse-dev mailing list [email protected] http://lists.the-brannons.com/mailman/listinfo/edbrowse-dev
