Create a new, custom, recovery middleware for Negroni so that:
we can present a pretty error page to the user if the application
panics
we can output the stack trace using Logrus' logger for consistency
with the rest of the application
I'm a little concerned that my making this recovery code more complex
(by rendering a pretty error page), we're increasing the risk of
panicking while recovering from the original panic.
This might be a good argument for dealing with this higher up the stack
(i.e. by configuring Nginx to output a pretty page when a HTTP 5XX
occurs), but for now I'm not using Nginx so this will do.
Create a new, custom, recovery middleware for Negroni so that:
I'm a little concerned that my making this recovery code more complex (by rendering a pretty error page), we're increasing the risk of panicking while recovering from the original panic.
This might be a good argument for dealing with this higher up the stack (i.e. by configuring Nginx to output a pretty page when a HTTP 5XX occurs), but for now I'm not using Nginx so this will do.
Fixes #16.