I am currently working on an application for Catalyst. One set of pages (the meat of the application) will direct users differently depending on a site wide state, i.e., everyone on the site starts at state 1 (roughly simultaneously), and when a certain amount of time or a certain event has taken place, they are moved onto state 2 (again, as simultaneously as possible). The issue is that state 2 depends on a somewhat high load set of calculations and database hits, so I would only like to perform that calculation once. How would one typically deal with this problem?

The best answer that I came up with is that I should use a cache, such as Catalyst::Plugin::Cache, to store the results of the state change. The problem is that I expect the change from state 1 to 2 to be triggered by a user. If user #1 triggers the state change, I may still be performing the calculations necessary when user #2 starts looking at the cache. Of course, since the calculations are still being performed, the cache is still empty, so process dealing with user #2 will start performing the calculations, even though the process dealing with user #1 is already performing these calculations.

In HTML::Mason, the caching mechanism has a "busy_lock" to prevent this problem. However, I have not been able to find any standardized way of doing this in Catalyst. At the moment, I'm thinking I'll have to use something like a semaphore to indicate when the cache is currently being processed (and to avoid race conditions while checking if the cache is being processed). The thing is, I feel like I can hardly be the first person to come up against this problem, so I'm wondering if there is a standard or semi-standard way to deal with these issues, either as a general module or a Catalyst specific one.