06:38:38 GMT <Ashutosh> HELP
06:58:59 GMT <solrize> uh?
16:26:44 GMT <mustmodify> are there pros and cons other than the normal "you get to compile your own stuff" for using apt-get vs downloading and compiling Redis?
17:34:00 GMT <marcv> Hello. The only difference Redis doc mentions about EVAL and EVALSHA is bandwidth consumption. But apart from that, is one faster than the other or are they completely identical performance-wise?
19:07:20 GMT <rouzier> Hey guys I was wondering if it would be a good idea to have a [RL]PUSHNX command
19:10:20 GMT <hahuang61> hey guys, is there a way to set an entire list?
19:10:33 GMT <hahuang61> and overwrite whatever is there if there already exists a list
19:10:59 GMT <rouzier> You can do that in two command
19:11:04 GMT <minus> rouzier: sounds like a case where using SETNX would make sense
19:11:22 GMT <rouzier> DEL list_name
19:11:46 GMT <rouzier> RPUSH list_name v1 v2 v3 v4 v5
19:11:52 GMT <minus> hahuang61: may be better to just serialize your data and drop that into a normal key
19:11:57 GMT <rouzier> Ok I will look into that
19:12:35 GMT <hahuang61> minus: i'd like to be able to pop from it though
19:14:00 GMT <minus> if it doesn't have to be atomic, just do what rouzier suggested
19:14:28 GMT <minus> if it has to be you could wrap it in a watch/multi/exec
19:14:41 GMT <minus> but that could fail
19:15:15 GMT <rouzier> minus is there anything equivalent to a blocking command for keys?
19:15:48 GMT <minus> rouzier: i don't think so. there's keyspace notifications though
19:16:14 GMT <minus> i.e. you can subscribe to keys appearing/getting deleted
19:17:00 GMT <hahuang61> it doesn't have to be atomic. thanks for the suggest guys
19:17:35 GMT <rouzier> The only problem is notification subcriber is not watch at the moment.
19:17:37 GMT <minus> yeah, then just do a DEL first, doesn't hurt if the key doesn't exist
19:18:13 GMT <rouzier> I know I could wrap the LPUSHNX logic in a lua script.
19:18:20 GMT <minus> yeah
19:18:33 GMT <rouzier> I was just wondering does it make sense as a command
19:18:54 GMT <rouzier> I already have a working implementation
19:19:02 GMT <minus> i'd think it is an edge case and few people would benefit from it
19:19:25 GMT <minus> and personally i'd avoid cluttering the command space with more commands unless very useful
19:20:24 GMT <minus> but for both your cases, rouzier, hahuang61, it may be worth taking a step back and thinking if there isn't a different solution that's simpler and already works prefectly as is
19:22:09 GMT <rouzier> Having a [LR]PUSHNX simplies the logic in my case.
19:22:35 GMT <rouzier> Like I said I will just wrap that logic in a lua script.
19:23:53 GMT <rouzier> My main issue is that I can have only a simple type of job running at a time.
19:25:05 GMT <rouzier> minux you are right
19:25:29 GMT <minus> of course i am ;D
19:26:05 GMT <minus> if you're processing jobs you can maybe have a queue per job type instead
19:32:53 GMT <rouzier> The problem is making sure that certain jobs types should be ignore if another work is currently working on them. I think watch multi setx exec will work in this case
19:37:21 GMT <rouzier> setnx
19:38:05 GMT <minus> kinda sounds like a priority queue
19:41:54 GMT <rouzier> I am processing snmp traps in a redis queue. If a certain trap type is currently being worked on another worker should should be handling that trap.
19:42:11 GMT <rouzier> should not be handling that trap.
19:43:06 GMT <rouzier> I am just trying to think of a good pattern to handle this scenario. I was think I using Design pattern: Locking with SETNX
19:43:38 GMT <rouzier> Since I only have a single redis server
19:45:58 GMT <rouzier>  /names
19:46:38 GMT <rouzier> minus thanks for the help
21:36:53 GMT <slact> Are there any plans to add '}' escaping for cluster key hash tags?
21:38:01 GMT <slact> so that, say, '{foo:\}bar}' would be hashed as key "foo:}bar", rather than "foo:\"