James Le Cuirot

As @kbaum pointed out in the other issue, `Time.zone=` is thread local. What @bjfish proposed above is still not safe as `Chronic.time_class` may get called by other threads during that...

Unfortunately I've completely forgotten why I looked into this. I don't believe we're using that workaround in production at the moment.

Thanks, I really needed this. However, I felt it was really hampered by the need to specify the -p all the time so I added support for a GOD_PORT environment...

I've just rebased my branch against 0.13.3. I'm keen to see this merged as deploying a non-standard version is a real PITA. I can't rely on bundler's ability to use...

The situation is quite confusing but basically I removed a previous memory leak fix. Maybe there really is a memory leak but the fix caused segfaults for me in certain...

Thanks for taking the time to look into this, it is appreciated.

I have a script that can create instances of various kinds of customer and sometimes the script cannot automatically determine which kind of customer to use so you have to...

It's because the results are returned in seemingly no particular order. I've raised the issue [here](http://forums.thegamesdb.net/showthread.php?tid=1385).

We'll probably want to track the upstream report at TheGamesDB/TheGamesDB#17. It looks like Apache Solr might be brought in to help with this.

Thanks for the quick reply. I know that Wine's DirectInput code adds the devices [here for js](https://github.com/wine-mirror/wine/blob/a4ed76b9484869d3527bcf75348ef4a03e17ec1a/dlls/dinput/joystick_linux.c#L168) and [here for event](https://github.com/wine-mirror/wine/blob/a4ed76b9484869d3527bcf75348ef4a03e17ec1a/dlls/dinput/joystick_linuxinput.c#L185) but I suspect that's not the code you meant....