NetSurf fetch, cache, and content system
There is a one-to-one mapping between URLs and content structures.
The resource at a URL may be required for two reasons:
1. The user requests a URL in the GUI by entering it or following a link.
2. A page contains an object (such as an image).
When a URL is required, call fetchcache() as follows:
c = fetchcache(url, referer, callback, p1, p2, width, height);
p1 and p2 are the callers private pointers used to identify the resource, and
they are passed to the callback. The triple (callback, p1, p2) must be unique.
The call returns immediately with a struct content. The structure may be in the
following states (c->status):
CONTENT_STATUS_TYPE_UNKNOWN -- the MIME type of the resource has not been
CONTENT_STATUS_LOADING -- the resource is being fetched or converted, and can
not be displayed.
CONTENT_STATUS_READY -- the resource is still loading, but may be displayed.
CONTENT_STATUS_DONE -- the resource has loaded completely.
States may only follow in the above order (but some may be skipped). The
callback function is called when the state changes or at other times as follows:
CONTENT_MSG_LOADING -- state has changed from CONTENT_STATUS_TYPE_UNKNOWN to
CONTENT_STATUS_LOADING. If the type is not acceptable content_remove_user()
should be called (see below).
CONTENT_MSG_READY -- state has changed to CONTENT_STATUS_READY.
CONTENT_MSG_DONE -- state has changed to CONTENT_STATUS_DONE.
CONTENT_MSG_ERROR -- a fatal error with the resource has occurred. The error
message is in the callback parameter. The content structure will be
destroyed after this message and must not be used.
CONTENT_MSG_STATUS -- the content structure's status message has changed.
If at any time the resource is no longer required, call content_remove_user():
content_remove_user(c, callback, p1, p2);
with the same callback, p1, p2 as passed to fetchcache().