tomato curry [company lunch]

Feb 09, 2008 14:30


I've been working with Ruby on a semi-regular basis for a while now, and there's something that's still bothering me. Well, to be honest, there are a number of things, but one of them in particular seems like it should be solvable.

Here's the last few frames of a typical Ruby on Rails traceback:

Read more... )

tracebacks, webdev, debugging, nevow, ruby, python, hacker, lazyweb, rails

Leave a comment

Comments 4

demonbane February 9 2008, 23:12:00 UTC
I know that it's possible to customize the view that Rails uses for showing the trace, but I don't know where exactly it's located. Though displaying error in the context of the code should really be handled by your IDE/editor. The Rails mode that I use for Emacs will do essentially the same thing that you show in your Nevow example above, except that I can just start typing and modifying code right then and there.

Reply

keturn February 9 2008, 23:22:49 UTC
I'm not always running rails through emacs. The tracebacks I read may not be from an instance I'm running at all, they may be passed on by another developer. If they only pass line numbers and not source excerpts, it's impossible for my emacs to figure out if their line numbers correspond to my line numbers. (Yes, bug reports should always have accompanying versions, but even so, one of us may have added a few lines of debugging information or something which would change the line count.)

And none of that lets you see what values were passed in to the method that crashed.

Reply


anonymous February 10 2008, 07:35:16 UTC
it's a sad world, with rails in twisted feed

Reply

keturn February 10 2008, 17:24:19 UTC
Just pretend I used Arrow or EventMachine as an example instead of Rails. ;)

Reply


Leave a comment

Up