Add a `gemojione:sprite` Rake task

This task will generate a standard and Retina sprite of all of the
current Gemojione Emojis, with the accompanying SCSS map.

It will not appear in `rake -T` output, and the dependent gems are not
included in the Gemfile by default, because this task will only be
needed occasionally.

See merge request !3130

Fix error 500 in Todos

Closes #14095 

Closes #14075 

Closes #14109

Closes #14151

See merge request !3141

This task will generate a standard and Retina sprite of all of the
current Gemojione Emojis, with the accompanying SCSS map.

It will not appear in `rake -T` output, and the dependent gems are not
included in the Gemfile by default, because this task will only be
needed occasionally.

[ci skip]

Snippet visibility badge spacing

Closes #14074 

![Screen_Shot_2016-03-08_at_08.40.32](/uploads/f55ca724fbf0e247b299a2702c3b55f5/Screen_Shot_2016-03-08_at_08.40.32.png)

See merge request !3122

Commit limits

Prevent timeouts when creating/viewing Merge Requests with many
commits. Also, reduce the number of commits shown in the UI from 500
to 100.

Closes https://gitlab.com/gitlab-org/gitlab-ce/issues/14031

This MR avoids Unicorn timeouts in some places and takes 4.5 seconds
off load times in others, when manually testing with a 4500-commit
compare/MR in the linux repo. It does not limit the number of Commit
objects instantiated in memory, just the amount of HTML generated. It
seems that having 4500 commits objects is not a problem in itself. If
it ever becomes one we could do something for Commit objects like we
did for Diff (introduce a CommitCollection or something).

See merge request !3095

Bump poltergeist to `~> 1.9.0`



See merge request !3116

Retry failed tests

This is to make the CI builds more reliable due to large amount of timeouts and deadlocks recently.


See merge request !1793

Fix bug where Bitbucket 'closed' issues were imported as 'opened'

_Originally opened at !2930 by @iurisilvio._

- - -

Bitbucket has an undocumented status `closed`.

https://confluence.atlassian.com/bitbucket/issues-resource-296095191.html

Related to https://gitlab.com/gitlab-com/support-forum/issues/559

See merge request !2973

See merge request !3136

This wasn't included in gitlab-org/gitlab-ce!3000.

Fix master branch link to upgrader.md



See merge request !2967

Make better use of the `visibility_level` factory traits



See merge request !3131

Remove `Snippet#expires_at`

This was removed from the interface in
https://github.com/gitlabhq/gitlabhq/pull/6027 but its implementation
lingered around for two years.

See merge request !3103

Use dedicated methods for counting branches and tags

This started out as "Lets add two methods to count and cache some data" and ended up in a clean-up/fix of some existing code. The two problems were:

1. Different code was used for adding/removing branches/tags via Git and the UI
2. The code used for the UI didn't have any RSpec tests, and I couldn't find any Spinach tests either (though grepping for Spinach stuff is hard)

This MR addresses the following:

1. `Repository#branch_count` and `Repository#tag_count` are used to count and cache the number of branches/tags, these methods are then used on the branches/commits/tags pages.
2. `Repository#add_tag`, `Repository#add_branch`, `Repository#rm_tag` and `Repository#rm_branch` now all the appropriate before/after hook methods instead of calling a random single cache expiration method. This ensures caches are properly flushed when adding/removing tags/branches via the UI.
3. RSpec tests were added for the above methods.

This fixes gitlab-org/gitlab-ce#13459

See merge request !3128

Expose User#last_sign_in_at and User#confirmed_at for admins

Closes #840

See merge request !3090

Merge branch 'improve/3102-memoize-group-instance-variable-in-admin-groups-controller' into 'master'

Remove instance variable `@group` multiple assignment

See merge request !3127

Remove duplicate 'cache_store' configuration in test environment

[ci skip]

See merge request !3061

This ensures the right caches are flushed when adding a branch via the
UI, instead of only flushing this one random cache.

This ensures Repository#add_tag calls Repository#before_push_tag instead
of just 1 random cache expiration method.

This ensures that Repository#rm_branch calls
before_remove_branch/after_remove_branch instead of just 1 random cache
expiration method.

The methods used for this are Repository#tag_count and
Repository#branch_count which cache their output in Redis as well as
memoizing it in an instance variable. Both methods have a corresponding
methods/hooks to flush the caches at the right time.

Added basic SQL guidelines

This is mostly based on the various changes I've made over the past few months.

See merge request !3091

update slack instructions

the previous instructions no longer apply, there is no "Incoming WebHooks" row anymore.

luckily there exists direct link to web hooks, so the previous steps aren't even neccessary

See merge request !3052