(?) Multiple concurrent requests for the same anchor.
Thank you, Markku, for your answer.
> > My question is whether there is some standard way to find out if the
> > given anchor is already being loaded (there is a request for it)?
> I *am* using the document (HTAnchor_setDocument) for this purpose.
As you can see in HTAnchor.c, it just does: me->document = document .
> Maybe I am just lucky and my application works with current library
> (5.0a) and breaks with the next version, or perhaps the following
> piece of code (copied from some examples) helps me?
> HTRequest_setReloadMode(obj->request, HT_CACHE_FLUSH);
> Hopefully, that CACHE_FLUSH disables the check for anchor document in
> memory cache filter when I really want to start a new load operation.
This is the main problem.
May be you haven't understood my poor crippled English, I'll try again.
I do NOT want to start a _new_load_ each time. I want make it efficiently.
I do want to use the anchor from memory cache, if it is _already_ there,
i.e. has already been loaded, get the document from persistent cache,
if the anchor isn't present in memory, but is present in persistent cache,
and that entry hasn't expired, and so on... and request the object from
network if it necessary (I think this scheme is used as main policy
in libwww, isn't it?).
And I want to get informed in some way, if the anchor is currently being
loaded (a request for the anchor has been issued, but not finished yet).
If this is the case, starting a new load (and even reload) operation
would be unreasonable --- the document which is currently being loaded
is fresh and that's why it's the best variant.