pezra / rspec-mode

An RSpec minor mode for Emacs

Emacs RSpec Mode

RSpec mode provides some convenience functions for dealing with RSpec.

Installation

You can install via ELPA, or manually by downloading rspec-mode and adding the following to your init file:

(add-to-list 'load-path "/path/to/rspec-mode")
(require 'rspec-mode)

Provided you have yasnippet installed, you can load the snippets:

(eval-after-load 'rspec-mode
 '(rspec-install-snippets))

(This isn't done automatically to avoid conflicts with snippets you may already have set up.)

Usage

If rspec-mode is installed properly, it will be started automatically when ruby-mode is started.

RSpec Verifiable mode

These keybindings are available in any Ruby source file:

Keybinding	Description
C-c , v	Verify the spec file associated with the current buffer
C-c , a	Run spec for entire project
C-c , t	Toggle back and forth between a spec and its target
C-c , e	Toggle back and forth between a method and its examples in the spec file
C-c , 4 t	Find in the other window the spec or the target file
C-c , 4 e	As above, but try to navigate to the example or method corresponding to point
C-c , r	Re-run the last verification process
C-c , y	Yank the last verification command to clipboard
C-c , m	Run all specs related to the current buffer
C-c , c	Run the current spec and all after it
C-c , s	Verify the example or method defined at point
C-c , f	Re-run just the failed examples from the last run

RSpec mode

These keybindings are available in Ruby spec files:

Keybinding	Description
C-c , s	Run the specified example at point
C-c , d	Toggle the pendingness of the example at point

RSpec Dired mode

These keybindings are available in Dired buffers:

Keybinding	Description
C-c , v	Run all specs in the current directory
C-c , s	Run marked specs or spec at point (works with directories too)
C-c , a	Run the 'spec' rake task for the project of the current file
C-c , r	Re-run the last RSpec invocation

See rspec-mode.el for further usage.

Hooks

Any functions in rspec-before-verification-hook will be executed before the verification (rspec-verify) and variants.

Any functions in rspec-after-verification-hook will be executed after the verification (rspec-verify) and variants. The hook will be executed whatever the outcome of the verification.

Gotchas

Debugging

To use binding.pry or byebug, install inf-ruby and add this to your init file:

(add-hook 'after-init-hook 'inf-ruby-switch-setup)

When you've hit the breakpoint, hit C-x C-q to enable inf-ruby.

RVM

If you use RVM, you may have to set rspec-use-rvm to true to make rspec-mode function properly:

(setq rspec-use-rvm t)

Or set it using Emacs' customization system.

ZSH and RVM

If you use ZSH and RVM, you may encounter problems running the specs. It may be so that an older version of Ruby, than the one you specified in .rvmrc, is used. This is because ZSH runs a small script each time a shell is created, which modifies the $PATH. The problem is that it prepends some default paths, such as /usr/bin, which contains another ruby binary.

What you can do to solve this is to use BASH for running the specs. This piece of code does the job:

(defadvice rspec-compile (around rspec-compile-around)
  "Use BASH shell for running the specs because of ZSH issues."
  (let ((shell-file-name "/bin/bash"))
    ad-do-it))

(ad-activate 'rspec-compile)

Vagrant

You can run specs inside a Vagrant box. You can enable it through the rspec-use-vagrant-when-possible customization option. You can also set the directory where your project is inside your box through the rspec-vagrant-cwd option. This will run specs through the vagrant ssh -c 'cd <cwd>; <rspec command>'.

Docker

You can run specs inside a Docker container. This can be enabled through the rspec-use-docker-when-possible option. This enabled, rspec is executed through docker-compose run. The following customization options are available:

Option	Default value	Description
rspec-use-docker-when-possible	nil	Enable docker
rspec-docker-command	docker-compose run	Docker command to run
rspec-docker-cwd	/app/	Path rspec to run in inside of the container
rspec-docker-container	rspec-container-name	Name of the container to run rspec into

To define the options for different projects, have a look at Per-Directory Local Variables.

Auto-scrolling

Set compilation-scroll-output. For example, (setq compilation-scroll-output t) will turn on auto scrolling.

Multiple Compilation Buffer Support

RSpec Mode can support running multiple RSpec compilation buffers concurrently with the rspec-allow-multiple-compilation-buffers option. To turn this on:

(setq rspec-allow-multiple-compilation-buffers t)

Contributing

Love RSpec and Emacs? Great, help out by contributing. The easiest way to contribute is to checkout the git project, make a change and then submit a pull request.

Note on Patches/Pull Requests

• Fork the project.
• Make your feature addition or bug fix.
• Update the version and changelog in the header of rspec-mode.el to reflect the change.
• Send me a pull request. Bonus points for topic branches.

rspec-mode's Issues

binding.pry doesn't work

I get this when inserting a binding.pry:

[0GWARNING: terminal is not fully functional
-  (press RETURN)

re: top level describe blocks

The current recommended way to declare top level describe blocks is as follows:

RSpec.describe "SUT" do
  # ...
end

Could this be a feature that rspec-mode would want to support?

rspec-toggle-spec-and-target error

Hi,

I've recently cleaned out my .emacs.d and started again from scratch.

I've installed the latest elpa version of Rspec-mode (1.6). Unfortunately the toggle between spec and target (C-c , t) is not functioning and erroring with the following.

Symbols function defintion is void: first

I am in no way proficient in lisp but it appears to be an issue that affects both
rspec-spec-file-for
rspec-target-file-for

And is reproducible via
(rspec-spec-file-for "/home/singram/projects/bar/app/models/data.rb")
(rspec-target-file-for "/home/singram/projects/bar/app/models/data.rb")

The project root is correctly detected.
(rspec-project-root "/home/singram/projects/bar/app/models/data.rb")
=> "/home/singram/projects/bar/"

Tested via C-x C-e (in case that's the wrong way to evaluate)

Any insights would be appreciated if I'm doing something wrong here. Or indeed where I should look next to debug and contribute back. This has worked in previous versions but having diff'd the files cannot see anywhere obvious that would affect this functionality

Thanks

• Stuart

Respect compilation-scroll-output

I think this is a better solution than auto scrolling at the end. That way the auto scroll flag can be removed and rspec-mode can defer to the user's compilation scroll preferences

Is it possible to make output scrolling configurable?

https://github.com/pezra/rspec-mode/blob/master/rspec-mode.el#L663

Is it possible to turn this to a variable? My ouput always auto scrolls to the bottom, which makes it hard to see the actual error on a single spec run.

Is this project still maintained?

I see some pull request that are 8 months old, and no updates for the last 4 months.

Is there something that can be done to speed things up? Maybe adding some collaborators?

I was thinking of adding some functionalities and tests, but I don't want to spend time on an unmaintained project. I also don't want to work on my own fork, because the efforts I make would be wasted on the community.

rspec-verify et. al. not split window if target compile buffer already in another frame

Not sure how to mark this as a 'feature' or 'suggestion'

I usually work with one frame with emacs but I'm toying with multiple frames (and I'm also toying with testing a great deal recently). I put the "rspec-compilation" buffer in its own separate frame and then run (or re-run) rspec-verify (or any of the other similar test centric commands) and the separate frame is updated like I want but the code still splits the current frame (the one with the spec in it where I issued the command) and displays another copy of the same buffer in the new window.

For me, I'd like for that to not happen. Just the one instance of the buffer it its own frame is all I want to see and leave the originating frame alone. In this way, I can set up frames and windows with what I want to see and work on and doing the test could just update the appropriate windows but does not change the window configuration.

I predict that there are others that want to see it both places. I don't know if this is a feature of Emacs' compilation or if rspec is doing this.

rspec-from-project-root should not be a macro

AS metioned in the emacs lisp reference
" When a macro call appears in a Lisp program being compiled, the Lisp compiler calls the macro definition just as the interpreter would, and receives an expansion. But instead of evaluating this expansion, it compiles the expansion as if it had appeared directly in the program. As a result, the compiled code produces the value and side effects intended for the macro, but executes at full compiled speed. This would not work if the macro body computed the value and side effects itself--they would be computed at compile time, which is not useful. "

The macro with a runtime variable would cause some side effect.

and since the macro contains "(or (rspec-project-root) default-directory)",when we compile the rspec-mode,the runtime variables would be assigned and become static. and it would cause the spec run in a wrong root directory.

we need to change the macro to function to fix the issue.

cheers!

dyld: Symbol not found: _system_free

some of my request specs, when run from rspec-mode, generate the following error:

bundle exec rspec /Users/fharbec/projects/cellars/spec/requests/landing_pages_spec.rb --options /Users/fharbec/projects/cellars/.rspec
dyld: lazy symbol binding failed: Symbol not found: _system_free
Referenced from: /Users/fharbec/.rvm/gems/ree-1.8.7-2011.03@cellars/gems/nokogiri-1.5.2/lib/nokogiri/nokogiri.bundle
Expected in: flat namespace

When I run the same command through the command line it runs ok. All my rspec model specs seem to fine using rspec-mode as well. Anyone else encounter this ?

• thanks

provide hook for conditioning environment prior to running tests?

I'm not an RVM user -- instead I have a customized sandbox directory with all the executables and gems required for a project. I use a SETUP file that exports a BUNCH of environment variables, notably PATH and various passwords and access keys.

Is there a straightforward way to exec my SETUP file prior to running the rspec command? If not, what's the preferred emacs idiom for creating a before hook?

FWIW, my truly awful hack at present is to patch rspec-runner as follows:

(defun rspec-runner ()
  "Returns command line to run rspec"
  (let ((bundle-command (if (rspec-bundle-p) "PATH=/Users/r/Developer/Chalcedony/usr/bin:${PATH} bundle exec " ""))
    ... etc...

Even that's not sufficient since it doesn't set up all my other environment variables. But it's enough to run tests.

zeus command not found, ./bin/zeus works

Because

• bundle exec adds a few seconds
• the path does not seam to be getting set from an environment variable in .bashrc
• the default directory is set,
• bundler is using binstubs

using ./bin/zeus works where 'zeus' doesn't.

Is there another way to get the ./bin directory on the path?

see commit:teaforthecat@20bbc8f

FactoryGirl: rspec--include-fg-syntax-methods-p

Just noticed that this checks the spec helper for FactoryGirl::Syntax::Methods. More recent versions of rspec-rails will create a spec_helper.rb and rails_helper.rb file, and the appropriate place for putting this include is in the latter. Should the rails helper also be checked (assuming of course that it's a rails project, and that the rails helper is there)?

Cannot locate project root

Ever since commit f972c4a, rspec-mode has been unable to locate the project root of my rails projects. Wherever I run the rspec-verify-* functions, they always report that rspec-mode "Could not determine the project root."

I am using rspec-mode 20120320 from MELPA. The project I am testing against is Rails 3.2 with both a Gemfile and a Rakefile in the root.

If I can provide additional debugging information, let me know how. Thanks for the great work, this is by far the best rspec mode I have found.

key binding issue

i've been trying out rspec-mode in emacs and I quite love it. I'm having issues with certain key bindings. Most of the bindings work as expected however the one for rspec-verify-single does not seem to work. I get the message "C-c , s is undefined". However, if I do meta-x rspec-verify-single, that works just fine. If I toggle rspec-mode off and on then it seems to work.

I'm guessing that this is probably some strange issue with my setup but I thought I would throw it out there in case anyone else has encountered and fixed this?

Hide the *compilation* buffer

Hello,

Is it possible to automatically hide the *compilation* buffer that appears in tandem with the *rspec-compilation* buffer? This is one thing that's been nagging me, but otherwise, this mode is perfect and has proven invaluable in development. Thank you so much for it!

rspec-verify-single broken with Rspec 3

rspec-verify-single errors with a message

invalid option: --line

Rspec 3 changes syntax for running a single test from using e.g. '--line 23' as a param to appending the linenumber to the filename e.g. my_cool_spec.rb:23

how to bind re-run

Hi,

I want to bind the re-run command to a different key, what function do I need to call ? It's not obvious to me from the source code.

Thanks

global-set-key: Key sequence C-c , r starts with non-prefix key C-c ,

I get that when I try to run any of the *-verify functions, while the point is in a rspec buffer.

Spring no longer works

Per: https://github.com/rails/spring/blob/c57942857210d1ed6c242b140f81cd5147cccb32/CHANGELOG.md#092

The tmpfile directory is no longer nested under the APP_ROOT.

Is there an issue with just simply checking for the spring gem by running bundle list or something similar?

compilation-error-regexp-alist and compilation-error-regexp-alist-alist not set right

My symptom is that if I load rspec-mode and then do an rgrep, the resulting buffer is colored half red and half green. Looking at compilation-error-regexp-alist-alist, it seems to be goofed up but also, looking at other code, they tend to do something like this:

  (set (make-local-variable 'compilation-error-regexp-alist)
       ruby-compilation-error-regexp-alist)

So, the rspec specific patterns would be used only in buffers created by rspec-compile

bad requirement in elpa

Hi, I've just try to install the rspec-mode by elpa. unfortunately, I found the rspec-mode is indeed dependent to the snippet.el and mode-compile, and all this are not either in elpa or list in the package. Could you please review and fix them? thanks

format output

Hi, is there a way to format the output so that it can fit in the error buffer (when a test fails) or an easy way to visit each error without actually needing to switch to the error buffer and then pgup/pgdown to find the errors ?

Does not work with RSpec 2 (Rails 3)

When used with RSpec 2 (Rails 3 project). It does not work fine.

running a spec causes emacs to switch the current buffer to config/application.rb

Not sure what's going on here, but this always happens seemingly between the time that the rails environment has loaded, and when the test actually executes.

Rakefile dependency

rspec-mode will not operate if there is no file called "Rakefile" at the "root" of you ruby scripts. I discovered this since my project did not need a Rakefile.

rspec-compile: Invalid function: rspec-from-project-root

Recently upgraded rspec-mode and I'm getting this error when running any specs. The process begins to run, but the command is not executed from within the context of the project so certain files that deal with paths appear to be missing.

ANSI color output for --color option

The output displays ANSI color codes such as ^[[32m. This is even when calling ansi-color.el's ansi-color-for-comint-mode-on prior to compile.

I've been trying to get this working and will give a pull request if I figure it out.

GNU Emacs 23.1 OSX.

The header comment formatting is non-standard

With f5281ed, it goes firmly against the library headers conventions. Specifically, I mean replacing three semicolons with two and a line with dashes.

As a result, the code below now returns nil when evaluated in the rspec-mode.el buffer, and Melpa, for example, won't show the commentary in the package description buffer before the user can install it.

(require 'lisp-mnt)
(lm-commentary)

The section naming was mixed up before that commit, though. As you can see from the text, there's no need to separate the header fields from the copyright notice. The commentary section goes after that (should be named Commentary, not Documentation). The change log section is fine, aside from the semicolons and dashes.

Incorrect spec file path for files in nonstandard directories

In addition to the standard Rails directory structure, I have other root level directories like services or exhibits:

├── app
        ├── assets
        ├── controllers
        ├── helpers
        ├── mailers
        ├── models
        └── views
├── config
├── db
├── exhibits
├── Gemfile
├── Gemfile.lock
├── lib
├── log
├── public
├── services
├── spec
└── tmp

When I execute rspec-toggle-spec-and-target in a file in one of those nonstandard directories like services outside of app, it opens a buffer for a new file directly inside of spec, instead of the existing file in /spec/services.

^C , v in dired mode

How about hooking up ^C , v in a dired buffer open to, for example, spec/models to run all the specs in spec/models which would simply be calling "rspec spec/models"?

Anyone else running into exit with code 2 frequently?

I'm frequently running into this problem where zeus exits early with code 2, as described in burke/zeus#469.

I'm running into this more often with rspec-mode than when running zeus rspec from iTerm.

Anyone else running into this? Any suggestions on how it could be fixed?

Spaces in file name are not quoted

I just did C-c , v on a buffer in a directory with a space in the name and got a "load: cannot load such file" error.

Checking the generated command I saw

bundle exec rake spec SPEC_OPTS='--options /path/Name With Space/project/.rspec' SPEC='/path/Name With Space/project/spec/requests/orders_api_spec.rb'

which also fails on the command line. The problem is that the spaces in the path should be quoted. This would work

bundle exec rake spec SPEC_OPTS='--options /path/Name\ With\ Space/project/.rspec' SPEC='/path/Name\ With\ Space/project/spec/requests/orders_api_spec.rb'

I checked the source code but I'm too rusty on elisp to try to fix it. I'll work around it with a symlink.

Possible to make recompile behave like rspec-rerun?

The recompile command from compile.el runs the latest compilation again. When the latest compilation was triggered by rspec-mode, it runs it however in regular compilation-mode instead of the nicely formatted rspec-compilation-mode. Is there an easy way to make recompile to behave like rspec-rerun?

I'm currently using the following workaround in my custom save-and-recompile function to make it work with rspec and non-rspec reruns:

(defun save-and-recompile () "Save and recompile" (interactive)
  (save-buffer)
  (if (string-match "^bundle exec rspec " compile-command)
    (rspec-rerun)
    (recompile))
)

This workaround does the trick for me now, so feel free to close this issue if there is no easy fix, I was just wondering if there is a cleaner solution.

Pry support

Hi,

Is it possible to support pry in rspec-mode result buffer?

Currently I have to open a term buffer and insert the rspec command manually so that pry gets responsive. Although this is not crucial, it's kind of annoying.

I am getting the following error =>

before_session hook failed: ZeroDivisionError: divided by 0
/home/renanranelli/.rbenv/versions/2.1.2/lib/ruby/gems/2.1.0/gems/pry-0.10.0/lib/pry/pager.rb:210:in `/'
(see _pry_.hooks.errors to debug)

Any idea?

rspec-spring-p no longer works

In newer versions of spring, the spring pidfile is located at SYSTEM_TMP_DIR/spring/[md5(APP_ROOT)].pid. This is not necessarily /tmp—on OS X, Dir.tmpdir (Ruby) returns /var/folders/yc/b7x57kz950vcn4wwvtfl22wdsjczmm/T.

rails/spring@ea436ea

\C-c t is undefined

After typing "\C-c t", I'm getting the following message:

C-c t is undefined

Now, is the \C == Control-C ?

ELPA package does not install snippets

When rspec-mode is installed via M-x list-packages, snippets does not get installed.

rails3 toggling doesn't

In rails3 toggling between specs and target does not work for controllers or views, due to name changes.

font lock on various keywords

I was hoping that "describe" and other rspec words would be colored for me.

Does anyone else want that?

I might try and cobble that together -- hoping to steal from cucumber's "feature-mode".

error line not linking correctly

When a test has errors, the 'links' do not work correctly.

Example:

Failures:

1. Failure/Error: page.should have_css('form#new_perkkk') expected #has_css?("form#new_this_should_not_work") to return true, got false # ./spec/requests/perks_spec.rb:55:in `block (3 levels) in '

Finished in 5.83 seconds

1 example, 1 failure

In the above example the following is clickable => " # ./spec/requests/perks_spec.rb:55:i"
when the link is clicked it gives an error: "Find this error in (default # ./spec/requests/perks_spec.rb):

I'm using carbon emacs on Mac OS X 10.6.7 with Rinari and some other modes.

Please publish in melpa-stable

Hello, this mode is available in melpa, but not in melpa-stable yet. Request you to add it to the stable repo with an appropriate version number.

Basically you need to tag it appropriately. Steps are here.

Not changing default-directory to project root before running test in spec mode

I am having an issue where the root of compilation mode is not changing prior calling spec when using spec mode (as opposed to rake mode). Here is the output when I execute a verify-all.

-*- mode: compilation; default-directory: "/Projects/proj-web/spec/helpers/" -*-  
Compilation started at Fri Dec 17 00:33:31

spec /Projects/proj-web/spec/helpers/agencies_helper_spec.rb --options /Projects/proj-web/spec/spec.opts
...

This seems to be a problem with compilation mode more than rspec as the arguments passed to spec are accurate. I also discovered that if I M-x cd to the project root before kicking off the test, it works fine.

Any clues on why this is happening and what to do about it?

This is emacs 23.2, latest rspec mode, osx, rvm with ruby 1.8.7.

Do not display pending specs as errors

When i run pending specs the compilation buffer displays them as errors. It would be great to just mark them as pending (highlight yellow ore something like this) and let the compilation be successfull when all other specs pass.

Spring detection does not work

On OSX, rspec-mode does not find the Spring PID anymore.

This might be related to https://github.com/rails/spring/blob/master/CHANGELOG.md#092

compilation-error-regexp-alist-alist

(add-to-list 'compilation-error-regexp-alist-alist
             '(rspec "\\([0-9A-Za-z@_./\:-]+\\.rb\\):\\([0-9]+\\)" (expand-file-name 1) 2))

This regexp is affected all 'rb' files, not only spec codes.
So expand-file-name fails on normal grep result buffer.
Could you change like this?

(add-to-list 'compilation-error-regexp-alist-alist
             '(rspec "rspec.*\\([0-9A-Za-z@_./\:-]+\\.rb\\):\\([0-9]+\\)" (expand-file-name 1) 2))

or something to restrict spec codes only.

Thanks.

yet another snippet question / issue

I have this at init time:

(rspec-install-snippets) ; ditto for rspec
(yas-global-mode 1)

but it does not work. I do the same sort of thing for feature-mode and it works. In the message buffer, it says:

[yas] Prepared just-in-time loading for /Users/pedzan/.emacs.d/el-get/rspec-mode/snippets

If I do the M-x (rspec-install-snippets) from the rspec buffer, then the snippets get enabled but shouldn't what I have work? What am I missing?

Emacs crashes on *rspec-compilation* buffer

Hi! I'm using this mode for a while, good job!
I'd like to point out a minor issue I encountered recently. Running a long spec suite inside emacs crashes the app with this error:

Backtrace:
emacs[0x4fb23b]
emacs[0x4e0410]
emacs[0x4fa078]
/lib/x86_64-linux-gnu/libpthread.so.0(+0xfbd0)[0x7f7ea1630bd0]
emacs[0x451ee8]
emacs[0x43268d]
emacs[0x4329f2]
emacs[0x443911]
emacs[0x44a012]
emacs[0x45c426]
emacs[0x45e666]
emacs[0x55477b]
emacs[0x44af6f]
emacs[0x4eaf72]
emacs[0x4ed35d]
emacs[0x4ef35e]
emacs[0x554633]
emacs[0x4e2fbe]
emacs[0x554510]
emacs[0x4e4377]
emacs[0x4e4695]
emacs[0x41717f]
/lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xf5)[0x7f7ea1279ea5]
emacs[0x417c8d]

[1]  + 16515 segmentation fault (core dumped)  emacs --debug-init

After many tries I realized is the buffer line length that cause the error. I use this conf:

(set-default 'truncate-lines t)

and setting it to false solves my problem:

(add-hook 'rspec-compilation-mode-hook
          (lambda ()
            (setq truncate-lines nil)
            ))

Is it my config fault, or is it a problem with rspec-compilation-mode?

Doesn't use rvm.el gemset setting?

I'm not sure whether this is actually an rvm.el problem or rspec mode problem so apologies if it turns out to be a rvm.el problem.

The only maybe related issue I could find was senny/rvm.el#27.

I have the following in my init.el:

(setq rvm--gemset-default "my-desired-gemset")
(rvm-use-default)
(defadvice rspec-compile (around rspec-compile-around)
  "Use BASH shell for running the specs because of ZSH issues."
  (let ((shell-file-name "/bin/bash"))
    ad-do-it))

(ad-activate 'rspec-compile)

When I do an rspec-verify I get:

-- mode: rspec-compilation; default-directory: "~/git_repos/" --
RSpec Compilation started at Wed Feb 11 15:59:40

bundle exec rake spec SPEC_OPTS='--options /Users//git_repos//.rspec' SPEC='/Users//git_repos//spec/controllers/_spec.rb'
Could not find parser-2.2.0.2 in any of the sources
Run bundle install to install missing gems.

RSpec Compilation exited abnormally with code 7 at Wed Feb 11 15:59:40

My gemset has all gems required (including parser-2.2.0.2) but the global gemset doesn't. Doing a bundle install while using the global gemset "fixes" this issue.

I am on Mac OS X using the latest version of RVM, rvm.el, and rspec-mode with the brew installed stable emacs.

Any ideas on how to fix this? I specifically do not want to pollute the global gemset.

Emacs Error with simple example

if: Wrong type argument: stringp, nil
using C-c ,v

How to do “ Emacs path setup ”， “rspec-verify-all ” is no work

I am use rvm & zsh & mac emacs 24.4， but I don't know how to do "Emacs path setup"

➜  ~  echo $PATH
/Users/jim/.rvm/gems/ruby-2.2.0/bin:/Users/jim/.rvm/gems/ruby-2.2.0@global/bin:/Users/jim/.rvm/rubies/ruby-2.2.0/bin:/Library/Java/JavaVirtualMachines/jdk1.8.0_25.jdk/Contents/Home//bin:/Users/jim/bin:/usr/local/bin:/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin:/Users/jim/Downloads/apache-maven-3.2.5//bin:/Users/jim/.rvm/bin
➜  ~  cat .zshenv
export ZSH=$HOME/.oh-my-zsh
ZSH_THEME="robbyrussell"
export PATH=$HOME/bin:/usr/local/bin:$PATH
source $ZSH/oh-my-zsh.sh