Don't load all of GitLab in mail_room

Fixes #12731

cc @jacobvosmaer 

See merge request !3005

Fix wrong anchor in PROCESS.md



See merge request !3060

Fix import from gitlab.com fails

_Originally opened at !2896 by @kazsw._

- - -

Fixes #12652

CGI.escape encodes '/' by default.
Second argument can be removed.

See merge request !2988

[ci skip]

Fixes #12652

Fix help keyboard shortcut for relative URL setups

Fixes gitlab-org/gitlab-ce#12751

See merge request !3016

Bump sprockets to `~> 3.3.5`

Closes #6069

See merge request !3055

Change doc styleguide to use blockquotes for notes



See merge request !3009

Updated SAML and Omniauth docs

Updated SAML and Omniauth docs with information on the new behavior recently introduced in !2882

/cc @ernstvn @DouweM 

See merge request !2950

Closes #6069

Logged out search

Fixes #13676  

See merge request !3045

Add a paragraph on "Implementing design & UI elements" in CONTRIBUTING.md

Closes gitlab-com/www-gitlab-com#544

See merge request !2983

Signed-off-by: Dmitriy Zaporozhets <dmitriy.zaporozhets@gmail.com>

Also, link to this new paragraph in PROCESS.md and improved a bit this
document.

[ci skip]

Add customer label explanation and change label references to markdown

Add customer label plus change valid labels to markdown format so they render.

The unchanged labels do not have actual/valid labels behind them. We should consider removing them since they're obviously not used.

cc/ @sytses @dzaporozhets @DouweM What do you think about removing the unused labels?

Label names are now rendered

![Screen_Shot_2016-02-25_at_10.59.03_AM](/uploads/4869416c120422b605df8982a46f6d61/Screen_Shot_2016-02-25_at_10.59.03_AM.png)

See merge request !2972

Return empty array when commit has no statuses in API

This makes API endpoint for Commit Statuses return empty array instead
of 404 when commit exists, but has no statuses.

Closes #3080

See merge request !3010

Fix relative URL

See https://github.com/gitlabhq/gitlabhq/issues/10053

 1. Same configuration way for relative URL like with Omnibus
 2. Loading the relative configuration from Rakefile as Rails do not load initializers for `asset:precompile`

First point has another positive side effect: no collisions (due to git controlled `application.rb`) any more during the upgrades of source based installations and relative url configuration

 - [x] tests on the source based installation
 - [x] tests on the centos&ubuntu omnibus packages

Fixes: gitlab-org/gitlab-ce#13730, gitlab-org/gitlab-ce#13727, gitlab-org/omnibus-gitlab#1143 and https://github.com/gitlabhq/gitlabhq/issues/10053

See merge request !2979

Add bottom margin for flash container on login page
Signed-off-by: Dmitriy Zaporozhets <dmitriy.zaporozhets@gmail.com>

cc @jschatz1 @skyruler 

Fixes #13935 

See merge request !3040

Fix issue with overlap of sidebar links.

Thanks @iamphill for the help with this one. 

![Screen_Shot_2016-03-01_at_10.19.52_AM](/uploads/f203fde79ae397ad18f23c4108f1c306/Screen_Shot_2016-03-01_at_10.19.52_AM.png)

cc @iamphill @alfredo1 @dzaporozhets @rymai 

See merge request !3043

Replace Gollum `[[_TOC_]]` tag with result of TableOfContentsFilter

Closes #2494

See merge request !2952

Only validate avatar cropping values when avatar is present AND changed

Closes #13905

See merge request !3018

Properly check for a pre-existing error in `User#namespace_uniq`

Rails adds an empty error to the Errors object even if you just call
`#[]` on it:

```
[1] pry(main)> u = User.last
[2] pry(main)> u.errors.keys
=> []
[3] pry(main)> u.errors[:username].present?
=> false
[4] pry(main)> u.errors.keys
=> [:username]
```

See merge request !3017

Don't show any "2FA required" message if it's not actually required

Prior, if the user had enabled and then disabled 2FA, they would be
shown a "You must enable Two-factor Authentication for your account."
message when going back to re-activate it, even if 2FA enforcement was
disabled.

See merge request !3014

Precompile assets before all spinach tests

Try to fix failing tests.

cc @jschatz1 

See merge request !3041

Signed-off-by: Dmitriy Zaporozhets <dmitriy.zaporozhets@gmail.com>

Signed-off-by: Dmitriy Zaporozhets <dmitriy.zaporozhets@gmail.com>

Signed-off-by: Dmitriy Zaporozhets <dmitriy.zaporozhets@gmail.com>

Changed # to ! on merge requests on activity view

Closes #9082

![Screen_Shot_2016-02-23_at_11.59.55](/uploads/87b859ae63282ae24025168404beb158/Screen_Shot_2016-02-23_at_11.59.55.png)

See merge request !2938

Batch size >1000 does not pay off

We did a small experiment to see how a full scan of the Redis keys on
gitlab.com speeds up as we increase the batch size. The values on the
right are time in seconds for a full scan (no delete operations).

```
count: 10);      284.500529021
count: 100);      86.21216934
count: 1_000);    60.931676195
count: 10_000);   60.96355610
count: 100_000);  62.378172541
```

It looks like 1,000 is a good number.

See merge request !2970

Explain why we mangle blob content types

Closes https://gitlab.com/gitlab-org/gitlab-ce/issues/9079

See merge request !2956

Closes #9082