Because now addressable would consider them invalid anyway.

Reduces the number of queries needed to redact notes to which
the user does not have access. Also includes an N+1 query test
as a guard against future issues.

This is a follow-up from https://gitlab.com/gitlab-org/gitlab-ce/merge_requests/14327#note_40976854.

Because of a change in GitLab 9.5.4 to prevent users from assuming control of
a repository already on disk, the import task broke. Imports would fail with
the message, "There is already a repository with that name on disk".

This change skips the validation when the import is done from the
command-line.

Closes #37682

This reverts commit 1d3c33b5.

`Repository#has_visible_content?` used to rely on the cached count of local branches, but
since it is now an independently cached value it needs to be invalidated on its own.

Closes #38646

When calling pre-receive, post-receive, and update hooks, add the GitLab
username as the GL_USERNAME environment variable.

This patch only handles cases where pushes are over http, or via
the web interface.  Later, we will address the ssh case.

In line with other changes in GitLab, make email address validation
properly case insensitive. The email address in the commit may be in
any case, so it needs downcasing to match the address stored in GitLab
for the user. Without this change the comparison fails and commits are
not marked as verified.

See #37009.

Signed-off-by: NRémy Coutable <remy@rymai.me>

The first attempt didn't migrate all rows on GitLab.com, due to a couple of
issues:

1. Some rows in merge_request_diffs had truly huge numbers of commits and diffs
   serialised - one in particular had 26,000 commits!
2. The jobs were sometimes on Sidekiq hosts with frequent OOM errors, leading to
   the job being lost.

The previous commit adds more logging, and a more robust insertion method. This
commit reschedules the jobs, with a generous pause between each.