Possible bug or config problem

In my previous post, I ran through a short tutorial on using nginx as a front-end web proxy for your Catalyst application running on a Starman / PSGI webserver. I just uncovered a problem, and I can’t tell if it’s a configuration problem on my end, or an actual bug…

If I have a GET request, like so : ‘/widgets//view’ or a POST request ‘widgets//update’ for a new `widget`, my double forward slash is being omitted and instead, Catalyst is instead handling my request as ‘/widgets/view’ or ‘widgets/update’ and the response is a 404 page not found error. (Note, for an already existing widget, the double forward slash would be replaced with a primary key identifier of the current `Widget` : /widgets/1542151/view’, etc).

Running Apache + FastCGI, the default catalyst HTTP server, or Starman (via localhost:5000 requests) do not yield this error. Which leads me to believe this problem is stemming from my nginx web proxy. Probably some sort of configuration problem on my part.

To test, I will start an instance of the catalyst web server in place of the starman server (should be an easy change, start the catalyst server and change the port number in my web proxy to 3000 instead of 5000)…



Basically, change the proxy_pass from http://localhost:5000/ to http://localhost:5000

  1. No comments yet.

  1. No trackbacks yet.