Eric Lindvall

Haha, you're totally right. I ended up trying to fix the wrong panic. It's actually [this panic](https://github.com/peterbourgon/diskv/blob/fc0553497cbfcf78f101d0bf8e82c6e627f4bbb0/diskv.go#L716) that we are running into regularly with errors like: > 159451 bytes still...

I've adapted it to have an option to turn all of these panics into errors and am also adding more to the message so we can diagnose what's going on.

We have this deployed on consumer devices and it only happens infrequently. We don't have any steps to reproduce it. It's much better for us to eventually stop caching values...

I've run into this where I wanted to use a specific version of sinatra, and one of the gems I depend on (cehoffman-sinatra-respond_to) depends on sinatra-sinatra, which gets required, and...

There is no current support for child process monitoring. We've been batting around ideas to solve this problem, as it's relevant for both Unicorn and Passenger as well as many...

The current idea is that god would be able to automatically monitor all of the child processes (we can get that info from the OS). It wouldn't include any mechanism...

I have not had a chance to work on it. My current thinking is that a custom kind of Watch would be created for each of the children.

This isn't something I've seen before. My best guess is the logger you're using isn't flushing to STDOUT until you exit the process.

Could you try adding `STDOUT.sync = true` to the top of your script?

I would guess that you're running into something rvm-related and I'm not going to be much help there. sorry.