Fetching contributors…
Cannot retrieve contributors at this time
64 lines (56 sloc) 1.72 KB
* do we want to specify the maximum depth of nested namespaces?
* expire time may be affected by timezone and system time change (NOTE)
* verify return 0/1 everywhere
* Add disk database file checksum to verify if its broken or valid file
* ht->ls is wrong! dups appear
* support plain text databases
* implement mmap for w32
* implement sdb_ainc/sdb_adec
* implement io filters for load/save databases (io wrap)
* support json 2.0 ({foo:"bar"}) <- no need for quotes in name
* create string pool for a list of sizes
newpool (sizeof (char), 10, 100, 1000);
* Add api to store serialized in memory
* Add api for raw bytes
* Add api and syntax for json_length()
* Add comparision stuff
* Add support for socket file for memcached api
* add api to create db without caching it before
* remove buffer
- each key/value have a timestamp of creation/modification
- sdb_diff() must check both databases and look for removed entries
- syncronization between hosts - realtime storage
- run sync every N seconds? every N keys in cache?
- sdb_log to dump to a file descriptor all sets.. as a stream
* remove endian conversions
* use slices instead of mallocs
* use strstore
* accelerate get ops with cached results (const char* instead of strdup)
* implement udp protocol for memcache
* rewrite mcsdb in luvit
* use 16 bit key and 16 bit value // dont think so
* use kernel's linked list? it should be faster..
* hashtable is not cleaned up when sync
- must honor expire time entries
* key and value boundaries are probably not properly handled
string types:
char *hugestring;
int hugestring_size;