Redis auto_reconnect woes

[jabyrd3]

Hey, I noticed some confounding behavior in the redis driver.

If you've passed auto_reconnect in either 1.1 or 1.2 it mostly works! this is great! but there's an exception handler in the madeConnection method (fired when tcp handshake is complete via twisted itself) that sets RedisFactory.continueTrying to false. I'm sure there was some reasoning behind this originally, but shouldn't the driver keep trying to connect for subsequent requests even if madeConnection throws the exception the first time around?

Its line 224 in 1.1, but the same issue exists in 1.2.

If i'm missing something or being dumb, im sorry for bugging whoever's still subscribed to this repo, but this behavior is confusing to me and i'm curious about why it does what it does.

[fiorix]

The Redis driver is far behind the one in the original repo https://github.com/fiorix/txredisapi. It'd be good to sync it in cyclone.