I'm doing some trickery with a bunch of Rake tasks for a complex project, gradually refactoring away some of the complexity in chunks at a time. This has exposed the bizarre web of dependencies left behind by the previous project maintainer.
What I'd like to be able to do is to add a specific path in the project to require's list of paths to be searched, aka $:. However, I only want that path to be searched in the context of one particular method. Right now I'm doing something like this:
def foo()
# Look up old paths, add new special path.
paths = $:
$: << special_path
# Do work ...
bar()
baz()
quux()
# Reset.
$:.clear
$: << paths
end
def bar()
require '...' # If called from within foo(), will also search special_path.
...
end
This is clearly a monstrous hack. Is there a better way?
Since $: is an Array, you have to be careful about what you are doing. You need to take a copy (via dup) and replace it later. It' simpler to simply remove what you have added, though:
def foo
$: << special_path
# Do work ...
bar()
ensure
# Reset.
$:.delete(special_path)
end
Without more info, it's difficult to know if there is a better way.
require is actually a method, it's Kernel#require (which calls rb_require_safe) so you could at least perform your hackery in a monkey-patched version. If you like that kind of thing.
Alias the orignal require out of the way
If passed an absolute path, call the original require method
Else iterate over load path by creating an absolute path and calling the original require method.
Just for fun I had a quick bash at that, prototype is below. This isn't fully tested, I haven't checked the semantics of rb_require_safe, and you probably would also need to look at #load and #include for completeness -- and this remains a monkey-patch of the Kernel module. It's perhaps not entirely monstrous, but it's certainly a hack. Your call if it's better or worse than messing with the global $: variable.
module Kernel
alias original_require require
# Just like standard require but takes an
# optional second argument (a string or an
# array of strings) for additional directories
# to search.
def require(file, more_dirs=[])
if file =~ /^\// # absolute path
original_require(file)
else
($: + [ more_dirs ].flatten).each do |dir|
path = File.join(dir, file)
begin
return original_require(path)
rescue LoadError
end
end
raise LoadError,
"no such file to load -- #{file}"
end
end
end
Examples:
require 'mymod'
require 'mymod', '/home/me/lib'
require 'mymod', [ '/home/me/lib', '/home/you/lib' ]
Related
I want to be able to statically analyze my code. That is, to know from the plain text of the file where every function and variable comes from. IDEs and text editor plugins work better when they can trace the origin of every symbol as well.
So for example, if I have application code like this:
#...
Y.some_method()
#...
Then I want to see Y in an include/import/require/extend/def statement somewhere on the page.
In other languages I use, one can explicitly choose which sub-parts of a namespace to bring in to the current context.
Python:
from X import Y
Haskell:
import X (Y)
Elixir:
alias X.Y, as: Y
And while it's possible to import all contained names in Python, the "wildcard import" is frowned upon:
from X import *
". . . they make it unclear which names are present in the namespace, confusing both readers and many automated tools."
In Ruby, it seems that this fully implicit "wildcard" way is the only way to bring in a contained name:
include X
This makes Y available, but is there some way to make this explicit? The docs for Ruby include don't show any options.
What I'd really like to do in Ruby is something like one of these:
from X include Y
include X::Y as Y
The best I've come up with so far is:
require 'x/y' ; Y = X::Y
Here's a crazy hack in the answer to another question which would enable this.
Try this. But I agree with #tadman that you should consider doing it in the Ruby way.
Object.define_singleton_method(:include) do |*mths, from: nil|
mod = from || mths.first
mod = mod.dup
if from
all_mths = mod.instance_methods
(all_mths - mths).each { |mth| mod.send :undef_method, mth }
end
super(mod)
end
module Foobar
def foo
puts :foo
end
def bar
puts :bar
end
end
class Abc
include Foobar
end
Abc.new.foo # => 'foo'
Abc.new.bar # => 'foo'
class AbcWithoutBar
include :foo, from: Foobar
end
AbcWithoutBar.new.foo # => 'foo'
AbcWithoutBar.new.bar # => NoMethodError
Ruby always executes the code that you require
And since there is no partial execution of a file there cannot be partial require.
When you require a feature Ruby locates the corresponding file using the load paths in $: and then double checks against the list of loaded files in $" and if the file has not yet been loaded executes the file.
Ruby is a dynamic language, the best way to reason about its source code is halting a running program rather than statically. In fact even class and def are not declarations but just method calls that are executed at runtime. Consider for example this contrived example
class Surprise < [Array, Hash, Fixnum, Object].sample
end
If you want to know where a method or class has been defined best use pry. You can require pry and then use binding.pry to stop anywhere in your source code and spelunk around to inspect objects and source code. Two of the most useful commands are ls and $
ls prints all methods of an object or class
$ prints the file location and source code of a method
Yes, I know I can just use load instead of require. But that is not a good solution for my use case:
When the app boots, it requires a config file. Each environment has its own config. The config sets constants.
When the app boots, only one environment is required. However, during testing, it loads config files multiple times to make sure there are no syntax errors.
In the testing environment, the same config file may be loaded more than once. But I don't want to change the require to load because every time the a spec runs, it reloads the config. This should be done via require, because if the config has already been loaded, it raises already initialized constant warnings.
The cleanest solution I can see is to manually reset the require flag for the config file after any config spec.
Is there a way to do that in Ruby?
Edit: adding code.
When the app boots it calls the init file:
init.rb:
require "./config/environments/#{ ENV[ 'RACK_ENV' ]}.rb"
config/environments/test.rb:
APP_SETTING = :foo
config/environments/production.rb:
APP_SETTING = :bar
spec/models/config.rb: # It's not a model spec...
describe 'Config' do
specify do
load './config/environments/test.rb'
end
specify do
load './config/environments/production.rb'
end
Yes it can be done. You must know the path to the files that you want to reload. There is a special variable $LOADED_FEATURES which stores what has been loaded, and is used by require to decide whether to load a file when it is requested again.
Here I am assuming that the files you want to re-require all have the unique path /myapp/config/ in their name. But hopefully you can see that this would work for any rule about the path name you can code.
$LOADED_FEATURES.reject! { |path| path =~ /\/myapp\/config\// }
And that's it . . .
Some caveats:
require does not store or follow any kind of dependency tree, to know what it "should" have loaded. So you need to ensure the full chain of requires starting with the require command you run in the spec to re-load the config, and including everything you need to be loaded, is covered by the removed paths.
This will not unload class definitions or constants, but simply re-load the files. In fact that is literally what require does, it just calls load internally. So all the warning messages about re-defining constants will also need to be handled by un-defining the constants you expect to see defined in the files.
There is probably a design of your config and specs that avoids the need to do this.
if you really want to do this, here's one approach that doesn't leak into your test process. Fork a process for every config file you want to test, communicate the status back to the test process via IO.pipe and fail/succeed the test based on the result.
You can go as crazy as you want with the stuff you send down the pipe...
Here's some quick and dirty example to show you what I mean.
a config
# foo.rb
FOO = "from foo"
another config
# bar.rb
FOO = "from bar"
some faulty config
# witherror.rb
asdf
and your "test"
# yourtest.rb
def load_config(writer, config_file)
fork do
begin
require_relative config_file
writer.write "success: #{FOO}\n"
rescue
writer.write "fail: #{$!.message}\n"
end
writer.close
exit # maybe this is even enough to NOT make it run your other tests...
end
end
rd, writer = IO.pipe
load_config(writer, "foo.rb")
load_config(writer, "bar.rb")
load_config(writer, "witherror.rb")
writer.close
puts rd.read
puts rd.read
puts rd.read
puts FOO
The output is:
success: from foo
success: from bar
fail: undefined local variable or method `asdf' for main:Object
yourtest.rb:24:in `<main>': uninitialized constant FOO (NameError)
as you can see, the FOO constant doesn't leak into your test process etc.
Of course you're only through half way because there's more to it like, making sure only one process runs the test etc.
Frankly, I don't think this is a good idea, no matter what approach you chose because you'll open a can of worms and imho there's no really clean way to do this.
I'm trying to avoid ever adding a redundant path to ruby's LOAD_PATH. It's not a remarkably complicated task, I'm just wondering if there is a cleaner method then what I've come up with.
This is my current solution as it stands now:
def add_loadpath(new_path)
included = $LOAD_PATH.inject(false) do |acc,path|
acc || new_path == File.expand_path(path)
end
$LOAD_PATH.unshift new_path unless included
end
Then instead of doing the usual $LOAD_PATH.unshift SOME_PATH you'd call
add_loadpath SOME_PATH
This is to avoid problems when the load path includes two paths that point to the same folder but are not the same string. For example foo/../bar and bar
I believe all paths in $LOAD_PATH are already expanded, so File.expand_path(path) is pointless. Your code can be refactored to this:
def add_loadpath(new_path)
File.expand_path(new_path)
.tap{|new_path| $LOAD_PATH.unshift(new_path) unless $LOAD_PATH.include?(new_path)}
end
or
def add_loadpath(new_path)
$LOAD_PATH.unshift(File.expand_path(new_path)).uniq!
end
I'm developing a library that provides access to gem metadata, including it's location on the file system. The idea was to let gem authors set it to a relative path from any script:
# $root/example.gemspec
Example::Gem.root '.' # => $root/
# $root/lib/example/gem.rb
Example::Gem.root '../..' # => $root/
Then, the path of the current script would be used to compute the absolute path. My implementation is currently as follows:
def root(relative_to = nil, file = __FILE__)
unless relative_to.nil?
#root = File.expand_path relative_to, File.dirname(file)
end
#root
end
I thought __FILE__ would return the path to the caller's script, but that assumption is wrong.
It worked within the library itself, but broke down when I tried to integrate it with one of my other gems; the generated path was always relative to the support library itself.
How can I implement this without having to pass the current __FILE__ on every call? Otherwise, there isn't much value to be gained; writing root('../..', __FILE__) is almost the same as writing an actual method to do the same thing.
If it's possible to figure out the path without having to specify anything, that would be even better, but I couldn't think of anything. How does Rails do it?
By the way, I'm aware of Gem::Specification#gem_dir, but it always returns paths relative to the installation directory, even if the gem is not actually there, which makes it useless in a development environment.
You can always make use of the backtrace facility provided:
caller.first
It produces an amalgam of file and line but is usually separated by :. I'd be careful to allow for filenames or paths that may contain colon for whatever reason by ignoring the line information but preserving the rest. In other words, do not split but sub:
caller.first.sub(/:\d+:in .*$/, '')
Is there a shorter way to require a file located in the same directory (as the script being executed)?
require File.expand_path(File.dirname(__FILE__) + '/some_other_script')
I read that require "my_script" and require "./my_script" will actually load the script twice (ruby will not recognize that it is actually the same script), and this is the reason why File.expand_path is recommended: if it is used every time the script is required, then it will only be loaded once.
It seems weird to me that a concise language like Ruby does not seem to have a shorter solution. For example, python simply has this:
import .some_other_module_in_the_same_directory
I guess I could monkey-patch require... but that's just evil! ;-)
Since ruby 1.9 you can use require_relative.
Check the latest doc for require_relative or another version of the Core API.
Just require filename.
Yes, it will import it twice if you specify it as filename and ./filename, so don't do that. You're not specifying the .rb, so don't specify the path. I usually put the bulk of my application logic into a file in lib, and then have a script in bin that looks something like this:
#!/usr/bin/env ruby
$: << File.join(File.dirname(__FILE__), "/../lib")
require 'app.rb'
App.new.run(ARGV)
Another advantage is that I find it easier to do unit testing if the loading the application logic doesn't automatically start executing it.
The above will work even when you're running the script from some other directory.
However, inside the same directory the shorter forms you refer to work as expected and at least for ruby 1.9 won't result in a double-require.
testa.rb
puts "start test A"
require 'testb'
require './testb'
puts "finish test A"
testb.rb
puts "start test B"
puts "finish test B"
running 'ruby testa.rb' will result in:
start test A
start test B
finish test B
finish test A
However, the longer form will work even from another directory (eg. ruby somedir/script.rb)
Put this in a standard library directory (somewhere that's already in your default loadpath $:):
# push-loadpath.rb
if caller.first
$: << File.expand_path(File.dirname(caller.first))
end
Then, this should work
% ls /path/to/
bin.rb lib1.rb lib2.rb #...
% cat /path/to/bin.rb
load 'push-loadpath.rb'
require 'lib1'
require 'lib2'
#...
caller gives you access to the current callstack, and tells you what file and where, so push-loadpath.rb uses that to add the file that load'd it to the loadpath.
Note that you should load the file, rather than require it, so the body can be invoked multiple times (once for each time you want to alter the loadpath).
Alternately, you could wrap the body in a method,
# push-loadpath.rb
def push_loadpath
$: << File.expand_path(File.dirname(caller.first))
end
This would allow you to require it, and use it this way:
% ls /path/to/
bin.rb lib1.rb lib2.rb #...
% cat /path/to/bin.rb
require 'push-loadpath'
push_loadpath
require 'lib1'
require 'lib2'
#...