This is so that people do not get warnings about problems in existing
code when we turn on Hound style checks.
---
If git-blame brought you here you may want to read this, the
problem is with git-blame, not this change.
Try running these two lines just once:
```
git config --global alias.praise 'log -p -M --follow --stat --'
git config --global alias.praise-line 'log -p -M --pretty=format:"%h (%an %ai)" -L'
```
Now in future you can use
`git praise <path/to/your/file>`
or if you want to see the evolution of a specific line or range of lines
`git praise-line <start-line>:<end-line>:<path/to/your/file>`
Some examples you should try:
```
git praise lib/awesome_print/version.rb
git praise-line 8:8:lib/awesome_print/version.rb
```
Inspiration for these aliases:
http://blog.andrewray.me/a-better-git-blame/
This is also so that Hound can be introduced without it bugging
people about issues in existing code.
---
If git-blame brought you here you may want to read this, the
problem is with git-blame, not this change.
Try running these two lines just once:
``
git config --global alias.praise 'log -p -M --follow --stat --'
git config --global alias.praise-line 'log -p -M --pretty=format:"%h (%an %ai)" -L'
``
Now in future you can use
`git praise <path/to/your/file>`
or if you want to see the evolution of a specific line or range of lines
`git praise-line <start-line>:<end-line>:<path/to/your/file>`
Some examples you should try:
``
git praise lib/awesome_print/version.rb
git praise-line 8:8:lib/awesome_print/version.rb
``
Inspiration for these aliases:
http://blog.andrewray.me/a-better-git-blame/`
Use the ruby 1.9 hash syntax everywhere appropriate. This is to fix style
inconsistencies in the code base. This is also so that Hound can be used
without it bugging people every time they touch an older piece of code
but forget to update the syntax.
If git-blame brought you here you may want to read this, the
problem is with git-blame, not this change.
Try running these two lines just once:
```
git config --global alias.praise 'log -p -M --follow --stat --'
git config --global alias.praise-line 'log -p -M --pretty=format:"%h (%an %ai)" -L'
```
Now in future you can use
`git praise <path/to/your/file>`
or if you want to see the evolution of a specific line or range of lines
`git praise-line <start-line>:<end-line>:<path/to/your/file>`
Some examples you should try:
```
git praise lib/awesome_print/version.rb
git praise-line 8:8:lib/awesome_print/version.rb
```
Inspiration for these aliases:
http://blog.andrewray.me/a-better-git-blame/
Only using double quoted strings when needing interpolation.
If git-blame brought you here you may want to read this, the
problem is with git-blame, not this change.
Try running these two lines just once:
```
git config --global alias.praise 'log -p -M --follow --stat --'
git config --global alias.praise-line 'log -p -M --pretty=format:"%h (%an %ai)" -L'
```
Now in future you can use
`git praise <path/to/your/file>`
or if you want to see the evolution of a specific line or range of lines
`git praise-line <start-line>:<end-line>:<path/to/your/file>`
Some examples you should try:
```
git praise lib/awesome_print/version.rb
git praise-line 8:8:lib/awesome_print/version.rb
```
Inspiration for these aliases:
http://blog.andrewray.me/a-better-git-blame/
The previous stubbing of dotfiles was still allowing calls to fall
through and populate the AwesomePrint.defaults hash which would then
cause some tests to fail if run in a certain order.
e.g. https://github.com/awesome-print/awesome_print/issues/265
One place in particular where this would happen is if you ran the
action_view_spec before other specs as it did not stub the dotfile.
This change makes dotfile stubbing more reliable, specific and
ensures it is run for every spec. I've also removed all instances
where it was being called manually, as this is now not needed due
to it being automatic.
Simplify matching logic by adding a custom matcher which
handles the task of normalisation of object IDs instead of
calling gsub many times throughout the test code.
This allows our tests to focus more on what needs to be tested
instead obfuscating the tests with logic unrelated to what
the test is trying to cover.
It also allows making the normalisation code more robust so
there is less chance of an accidental match or typo slipping
in.
The change also replaces the arbitary default object IDs with a
string that makes it very obvious that it is just a placeholder.
Does not run rails specs when running mongoid specs
WIP Start of remove BRE to check tests
Remove old helper of active record
Use RSpec skip to skip AR specs
Run ActionView only when needed
Run ActiveSupport only when needed
Run Mongoid only when needed
Create ExtVerifier.require_dependencies
Run MongoMapper only when needed
Always load nokogiri specs
Always load ostruct specs
Run Ripple only when needed
Remove :: when check for defined constants
Require spec_helper directly
Remove 1.8.6 old monkey patch
Add some RSpec configs
* Get specs passing with latest versions of
* ActiveRecord 4.0.4
* ActiveSupport 4.0.4
* MongoMapper 0.13.0
* Mongoid 4.0.0
* ActionView 4.0.4
* Ripple seems to be abandoned so no changes made.
* Clean up ActiveRecord setup:
* Creating tableless AR objects seems even trickier in Rails 4 so
simplify testing setup by using an in-memory SQLite DB back-end.
Conflicts:
spec/ext/active_record_spec.rb
