#ifndef EL__ECMASCRIPT_ECMASCRIPT_H #define EL__ECMASCRIPT_ECMASCRIPT_H /* This is a trivial ECMAScript driver. All your base are belong to pasky. */ /* In the future you will get DOM, a complete ECMAScript interface and free * plasm displays for everyone. */ #include "main/module.h" #include "util/time.h" struct string; struct uri; struct view_state; struct ecmascript_interpreter { struct view_state *vs; void *backend_data; time_t exec_start; /* This is a cross-rerenderings accumulator of * @document.onload_snippets (see its description for juicy details). * They enter this list as they continue to appear there, and they * never leave it (so that we can always find from where to look for * any new snippets in document.onload_snippets). Instead, as we * go through the list we maintain a pointer to the last processed * entry. */ struct list_head onload_snippets; /* -> struct string_list_item */ struct string_list_item *current_onload_snippet; /* ID of the {struct document} where those onload_snippets belong to. * It is kept at 0 until it is definitively hard-attached to a given * final document. Then if we suddenly appear with this structure upon * a document with a different ID, we reset the state and start with a * fresh one (normally, that does not happen since reloading sets * ecmascript_fragile, but it can happen i.e. when the urrent document * is reloaded in another tab and then you just cause the current tab * to redraw. */ unsigned int onload_snippets_owner; }; /* Why is the interpreter bound to {struct view_state} instead of {struct * document}? That's easy, because the script won't raid just inside of the * document, but it will also want to generate pop-up boxes, adjust form * contents (which is doc_view-specific) etc. Of course the cons are that we * need to wait with any javascript code execution until we get bound to the * view_state through document_view - that means we are going to re-render the * document if it contains a