Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP

Exceptions

kyrylo edited this page · 22 revisions

Exception handling

Quick Menu:

Back to Main Menu

Overview

Pry catches most of the exceptions that your code raises so that you have time to work out what went wrong and fix the problem.

The _ex_ local variable

Pry gives you access to the most recently caught exception in a local variable called _ex_. Unlike Ruby's builtin $!, this persists until another exception is raised, so can be used for detailed digging.

The _ex_ variable is a 'special local', read more about those here.

For example:

pry(main)> 4/0
ZeroDivisionError: divided by 0
from (pry):1:in `/'

pry(main)> _ex_.message
 => "divided by 0"

pry(main)> puts _ex_.backtrace
(pry):1:in `/'
(pry):1

Back to top

Using the wtf? command

As an alternative to using _ex_.backtrace you can use the wtf? command to display a few lines of the backtrace for the most recent exception. If you want to see more lines, add more question marks or exclamation marks.

Example: Using wtf to investigate the very top of the backtrace

pry(main)> sdf
NameError: undefined local variable or method `sdf' for main:Object
from (pry):17:in `__pry__'
pry(main)> wtf
Exception: NameError: undefined local variable or method `sdf' for main:Object
--
 0: (pry):17:in `<main>'
 1: /Users/john/.rvm/gems/ruby-1.9.3-p125/gems/pry-0.9.8.3/lib/pry/pry_instance.rb:251:in `eval'
 2: /Users/john/.rvm/gems/ruby-1.9.3-p125/gems/pry-0.9.8.3/lib/pry/pry_instance.rb:251:in `re'
 3: /Users/john/.rvm/gems/ruby-1.9.3-p125/gems/pry-0.9.8.3/lib/pry/pry_instance.rb:218:in `rep'
 4: /Users/john/.rvm/gems/ruby-1.9.3-p125/gems/pry-0.9.8.3/lib/pry/pry_instance.rb:202:in `block (2 levels) in repl'

Example: Append ? to see more of the backtrace

pry(main)> sdf
NameError: undefined local variable or method `sdf' for main:Object
from (pry):17:in `__pry__'
pry(main)> wtf?
 0: (pry):17:in `<main>'
 1: /Users/john/.rvm/gems/ruby-1.9.3-p125/gems/pry-0.9.8.3/lib/pry/pry_instance.rb:251:in `eval'
 2: /Users/john/.rvm/gems/ruby-1.9.3-p125/gems/pry-0.9.8.3/lib/pry/pry_instance.rb:251:in `re'
 3: /Users/john/.rvm/gems/ruby-1.9.3-p125/gems/pry-0.9.8.3/lib/pry/pry_instance.rb:218:in `rep'
 4: /Users/john/.rvm/gems/ruby-1.9.3-p125/gems/pry-0.9.8.3/lib/pry/pry_instance.rb:202:in `block (2 levels) in repl'
 5: /Users/john/.rvm/gems/ruby-1.9.3-p125/gems/pry-0.9.8.3/lib/pry/pry_instance.rb:201:in `loop'
 6: /Users/john/.rvm/gems/ruby-1.9.3-p125/gems/pry-0.9.8.3/lib/pry/pry_instance.rb:201:in `block in repl'
 7: /Users/john/.rvm/gems/ruby-1.9.3-p125/gems/pry-0.9.8.3/lib/pry/pry_instance.rb:200:in `catch'
 8: /Users/john/.rvm/gems/ruby-1.9.3-p125/gems/pry-0.9.8.3/lib/pry/pry_instance.rb:200:in `repl'
 9: /Users/john/.rvm/gems/ruby-1.9.3-p125/gems/pry-0.9.8.3/lib/pry/pry_class.rb:136:in `start'

Example: Append even more ? or ! marks to see everything

pry(main)> sdf
NameError: undefined local variable or method `sdf' for main:Object
from (pry):17:in `__pry__'
pry(main)> wtf?!?!!!
NameError: undefined local variable or method `sdf' for main:Object
from (pry):17:in `__pry__'
 0: (pry):17:in `<main>'
 1: /Users/john/.rvm/gems/ruby-1.9.3-p125/gems/pry-0.9.8.3/lib/pry/pry_instance.rb:251:in `eval'
 2: /Users/john/.rvm/gems/ruby-1.9.3-p125/gems/pry-0.9.8.3/lib/pry/pry_instance.rb:251:in `re'
 3: /Users/john/.rvm/gems/ruby-1.9.3-p125/gems/pry-0.9.8.3/lib/pry/pry_instance.rb:218:in `rep'
 4: /Users/john/.rvm/gems/ruby-1.9.3-p125/gems/pry-0.9.8.3/lib/pry/pry_instance.rb:202:in `block (2 levels) in repl'
 5: /Users/john/.rvm/gems/ruby-1.9.3-p125/gems/pry-0.9.8.3/lib/pry/pry_instance.rb:201:in `loop'
 6: /Users/john/.rvm/gems/ruby-1.9.3-p125/gems/pry-0.9.8.3/lib/pry/pry_instance.rb:201:in `block in repl'
 7: /Users/john/.rvm/gems/ruby-1.9.3-p125/gems/pry-0.9.8.3/lib/pry/pry_instance.rb:200:in `catch'
 8: /Users/john/.rvm/gems/ruby-1.9.3-p125/gems/pry-0.9.8.3/lib/pry/pry_instance.rb:200:in `repl'
 9: /Users/john/.rvm/gems/ruby-1.9.3-p125/gems/pry-0.9.8.3/lib/pry/pry_class.rb:136:in `start'
10: /Users/john/.rvm/gems/ruby-1.9.3-p125/gems/pry-0.9.8.3/lib/pry/cli.rb:150:in `block in <top (required)>'
11: /Users/john/.rvm/gems/ruby-1.9.3-p125/gems/pry-0.9.8.3/lib/pry/cli.rb:59:in `call'
12: /Users/john/.rvm/gems/ruby-1.9.3-p125/gems/pry-0.9.8.3/lib/pry/cli.rb:59:in `block in parse_options'
13: /Users/john/.rvm/gems/ruby-1.9.3-p125/gems/pry-0.9.8.3/lib/pry/cli.rb:59:in `each'
14: /Users/john/.rvm/gems/ruby-1.9.3-p125/gems/pry-0.9.8.3/lib/pry/cli.rb:59:in `parse_options'
15: /Users/john/.rvm/gems/ruby-1.9.3-p125/gems/pry-0.9.8.3/bin/pry:16:in `<top (required)>'
16: /Users/john/.rvm/gems/ruby-1.9.3-p125/bin/pry:19:in `load'
17: /Users/john/.rvm/gems/ruby-1.9.3-p125/bin/pry:19:in `<main>'

To see the entire backtrace, pass the -v (--verbose) flag, e.g: wtf -v

Back to top

Using the cat --ex command

In addition to giving you easy access to the backtrace, Pry can also show you the actual code that caused problems. This is done with the --ex argument to the cat command.

For example:

[1] pry(main)> require 'set'
=> true
[2] pry(main)> Set.new(1)
ArgumentError: value must be enumerable
from /home/pryer/ruby/lib/ruby/1.8/set.rb:263:in `merge'
[3] pry(main)> cat --ex

Exception: ArgumentError: value must be enumerable
--
From: /home/pryer/ruby/lib/ruby/1.8/set.rb @ line 263 @ level:  0 of backtrace (of 3).

   258:   # returns self.
   259:   def merge(enum)
   260:     if enum.is_a?(Set)
   261:       @hash.update(enum.instance_eval { @hash })
   262:     else
 =>263:       enum.is_a?(Enumerable) or raise ArgumentError, "value must be enumerable"
   264:       enum.each { |o| add(o) }
   265:     end
   266: 
   267:     self
   268:   end

This is partly useful, I know now that the code I'm calling is explicitly expecting an Enumerable, but I didn't give it one. What's not so clear is where it's getting that Enumerable from — after all, I didn't call Set#merge! Luckily cat --ex also supports an optional argument, cat --ex N jumps N lines up the backtrace. To find out how my code ended up in Set#merge I can use cat --ex 1:

[4] pry(main)> cat --ex 1

Exception: ArgumentError: value must be enumerable
--
From: /home/pryer/ruby/lib/ruby/1.8/set.rb @ line 76 @ level:  1 of backtrace (of 3).

   68:   def initialize(enum = nil, &block) # :yields: o
   69:     @hash ||= Hash.new
   70: 
   71:     enum.nil? and return
   72: 
   73:     if block
   74:       enum.each { |o| add(block[o]) }
   75:     else
 =>76:       merge(enum)
   77:     end
   78:   end

It's now clear why the exception happened, and exactly what I need to do to fix it.

HINT: If you're having problems with the standard library, you should probably use the show-doc command command to read up on what you're doing wrong! In this case show-doc Set#initialize tells me it should take an Enumerable, not a number.

Use with wtf?

It can be effective to use this command in concert with the wtf? command. Use wtf? to show the backtrace and cat --ex N to show the context for a frame in that backtrace.

Back to top

Configuring cat --ex

cat --ex usually shows five lines of context before, and five lines of context after the exception. You can override this by setting Pry.config.default_window_size in your ~/.pryrc.

For example:

Pry.config.default_window_size = 10

This will show you a total of 21 lines of context: 10 before, 10 after and the actual line contained within the exception's backtrace.

Back to top

Using the edit --ex command

In the previous section, we saw how cat --ex can be used to show you how the error came about. In that case it was because we used the library wrong. Sometimes however, the problem is caused by the code being called. In that case we want to be able to fix the code as quickly as possible.

To help you do this, Pry's edit command also supports --ex in the same manner as cat. It also takes an optional number to jump up the stacktrace:

[1] pry(main)> require 'set'
=> true
[2] pry(main)> Set.new(1)
ArgumentError: value must be enumerable
from /home/pryer/ruby/lib/ruby/1.8/set.rb:263:in `merge'
[3] pry(main)> edit --ex 1 # N.B. do not try editing the standard library at home!
[4] pry(main)> Set.new(1)
=> #<Set: {1}>

Just like with cat --ex, if you provide a number to edit --ex N it will jump N levels up the stacktrace, and if you don't specify N (or set it to 0) it will edit the actual line of code that raised the exception.

Back to top

Customising the Exception handler

By default when Pry catches an exception, it doesn't show you very much information — it only includes the message, and the first line of the backtrace. If you'd prefer to see more information without having to use additional commands, you can override Pry.config.exception_handler in your ~/.pryrc.

For example, to show a few more lines of backtrace, you might want to do:

Pry.config.exception_handler = proc do |output, exception, _pry_|
  output.puts "#{exception}"
  output.puts "#{exception.backtrace.first(10)}"
end

Or, if you find that you always use cat --ex after an exception is raised, you can short-cut by getting Pry to run that for you:

Pry.config.exception_handler = proc do |output, exception, _pry_|
  _pry_.run_command 'cat --ex'
end

Because the exception_handler is designed to deal with exceptions thrown by the code that you are currently prying, it should be reasonably robust. For example, it would be silly to have an exception_handler that itself threw an exception. If you do end up throwing an exception by accident, Pry will still catch it and tell you, for example:

  (pry) output error: #<ArgumentError, too few arguments>

The other thing to remember is that you need to call output.puts, the return value of your Proc is ignored, and just using puts will send output to $stdout instead of output.

Back to top

Configuring which exceptions are caught

Pry tries to be clever about which exceptions are caught. While it tries to catch everything that your program might throw by accident, it tries not to catch exceptions that are thrown deliberately. The most obvious example of a deliberate exception is SystemExit, which is thrown when you call Kernel#exit, though Pry also allows all SignalExceptions to be passed through so that you can run the UNIX kill command on instances of Pry.

If your program has its own exceptions that should never be caught by Pry, you can add them to Pry.config.exception_whitelist in your ~/.pryrc. For example to tell Pry not to catch out-of-memory errors:

Pry.config.exception_whitelist = [SignalError, SystemExit, NoMemoryError]

If you find a class of exception that you think Pry should not catch by default, please open a bug report and let us know.

The pry-rescue plugin

For much more powerful exception handling, look into the pry-rescue plugin.

Back to top

Something went wrong with that request. Please try again.