> I wanted to implement cache sync policy (fetch cache dump on startup only > if upstream is newer, bigger or always) but I can't plug such logic into > current code. > > Syncing is done in uwsgi_cache_sync_from_nodes(), > uwsgi_hooked_parse() is called, which is generic function, and then calls > a > hook for cache (which is cache specific): cache_sync_hook() > But cache_sync_hook() does not return anything, so I can't signal from it > that we need to skip this node and go for next one (if we have more that > one node we can fetch from). > > Any tips where could I plug it? Could we add return value to the hook, and > if it returns != 0 then we return hooks value from uwsgi_hooked_parse() ? > > -- > Łukasz Mierzwa > _______________________________________________ > uWSGI mailing list > [email protected] > http://lists.unbit.it/cgi-bin/mailman/listinfo/uwsgi >
uwsgi_hooked_parse() takes a "data" argument (a void *) that is passed to the hook function. You can use that pointer to signal states (for example map it to a struct having a 'state' int in addition to the others neede parameters) -- Roberto De Ioris http://unbit.it _______________________________________________ uWSGI mailing list [email protected] http://lists.unbit.it/cgi-bin/mailman/listinfo/uwsgi
