01:16:38 GMT <BoogieMan> i was able to reproduce. even if you don't specify a database file, redis will load in the dump.rdb if it finds it
12:09:32 GMT <bleb> I'm having a tough time understanding why someone would use redisAppendCommand/redisGetReply over redisCommand from the hiredis C library
12:10:30 GMT <bleb> is it just that it's faster to write the output buffer all at once, rather than write it incrementally?
14:41:14 GMT <reynierpm> hi there, can I get some help here http://stackoverflow.com/questions/42773397/access-cache-items-created-outside-symfony-makes-them-work-sometimes-only-why ??
14:48:14 GMT <badboy_> bleb: yes, pipelining can improve things
14:56:22 GMT <bberry> does replication still apply to pubsub clients?
14:56:43 GMT <bberry> or is it slave clients only
15:01:38 GMT <bberry> seems like the latter
15:17:27 GMT <bberry> @badboy_ any idea about the problem im having with the soft limit on client-output-buffer-limit for a pubsub connection?
15:20:12 GMT <bberry> its `client-output-buffer-limit pubsub 32mb 8mb 60` in the config file.  I send a string that is 8.5mb, gets sent, no problems.  After X amount of time (where X > 60 seconds, I've waited up to an hour before) I send the same exact string again, and the client connection is severed.  Redis 3.0.7
15:40:10 GMT <badboy_> are you sure that the subscribed client fully read the data? what's the client's buffer between those two publishes? did you check if the same happens with a more recent version of Redis?
15:40:14 GMT <badboy_> bberry:
15:56:47 GMT <bberry> havent upgraded to a more recent version of redis, can do that next.  I'm pretty sure that the whole data amount was read.  The data is a tar.gz binary file (as b64) that gets untarred executed on a 3rd party device properly.  Gonna get the client info before, during,
15:56:51 GMT <bberry> er...
15:56:52 GMT <bberry> https://gist.github.com/bberry6/a34f97d73d72967a68a4369da004cde3
15:57:42 GMT <bberry> @badboy_
15:58:53 GMT <bberry> in that test i waited ~5 minutes
15:59:23 GMT <bberry> shall i update to latest redis and try it out then?
16:01:18 GMT <badboy_> yes
16:11:52 GMT <bberry> got one error from `make test`.  seems to be what people are describing here.... https://github.com/antirez/redis/issues/2715
16:12:02 GMT <bberry> i assume its okay though
16:20:00 GMT <bberry> badboy_: verbose log level is okay?
16:29:46 GMT <bberry> badboy_: same problem with 3.2.8.  Heres the output https://gist.github.com/bberry6/22c3a8ff1427a24265961d078544329f
16:45:05 GMT <bberry> badboy_: fwiw, when i set the output buffer for pubsub to 0 0 0, the client doesnt break the connection.  As it stands, it seems like I either dont correctly understand how the soft cap works, or there seems to be a bug with the way it is working.  Any other tests or log info youd like me to perform/show?
16:46:37 GMT <badboy_> no. report it upstream. i currently don't have the time to deal with it
16:47:42 GMT <bberry> gotcha
16:48:10 GMT <bberry> ty :D
17:32:38 GMT <bberry> @badboy_ should i send it to the google group or file it as an issue?
18:12:39 GMT <bberry> made an issue https://github.com/antirez/redis/issues/3893