Changed url validation in template
Moved owner & repo identification to the backend

- The parsing of the Github url (owner/repo/folder) was moved to the
  backend and is now prformed by processGithubUrl function
- The onwer & repo names are returned to the frontend after the
  pre-processing, along with the header names found in the repo

…would it be hardcoded to empty string?)

- Can now fetch a the list with the blobs and trees that belong to a provided
  folder.
- Removed some unncessary else statements
- Fixed brackets

- Use Github's tree API in order to get the complete list of the files
  under a provided repo or a repo directory
- Include only blobs to the returned filelist
- Does not work in current state (WIP)

…AJAX calls

Made the backend of the library manager capable of doing the following:
- Fetch the whole tree structure of a Github repo using Github's tree
  API
- Return the directory structure in a way that is understandable by the
  jsTree jQuery plugin (used by the frontend)
- Provide useful metadata such as in-repo full path and sha hash for
  every node of the returned tree (will make directory selection easier)

Also renamed some functions and removed some unused ones.

This reverts commit b6cda83.

…the jsTree structure

…one pointed by the url

…he user

…sitory's description

The removed sha key would be used when calling Github's contents API in
order to fetch the contents of each of the selected files.
Unfortunatelly this is not supported.
Instead, the root path of the library within the repository (might be
the same as the repo root directory) should be provided.
In order to achieve that, the path form field is updated each time the
jsTree selection event is fired.

Moved all the form validation and handling into a separate function and
added flashbag errors that will be returned upon failure (previously it
just dumped the error as a JSON stirng).

- validateFormData function makes sure only zip or git data are
  contained in the form and determines the type of library addition
  (git/zip)
- addLibrary performs the actual library addition, fetches the library
  files (either extracts them from the zip or fetches them from git) and
  saves the new library.

…e repo)

Made the function compatible with the getLastCommit function. However,
the current state of the checkGithubUpdates function does not support
checking the git branch of an already uploaded library before checking
for the last commit. Thus the master branch was used for all cases and a
TODO was added in order to check the function code again.
Note: this function is not actually used ever. We need to figure
out what we want it to return.

The getGithubRepoCode function handles the fetching of the code of a
folder (given a repo-branch combination). For every file in the folder,
getGithubFileCode function will be called in order to get the actual
content of the file. For every subfolder, the getGithubRepoCode will be
called recursively, until all the code is fetched.

…es for optional parameters

…lename

…l requests

…es created by the zip-process

… encoded.

…ke advantage of the 100 MB size limit of the 'blobs' API

Instead of converting all files' contents to UTF-8 (which breaks
binaries, executables, etc) write the file, then identify the mime-type
and if dealing with a text file that is not encoded in UTF-8, convert it
to UTF-8 and write the file again.
Might take longer to execute, but this is the safest way of converting
only text files to UTF-8.

…ary Github repos)

…to the library view

…properly

This function fetches the parameters needed for Github authentication
from the container and then performs the actual request to Github.
Enabled removing duplicated code from all Git-related functions

…ary description and humanName

As we already discussed, all data should be stored in our
databases/filesystems as is. That is without changing the character encoding
of the files, or any other of their properties.
Thus, the function was removed, and the file encoding should be changed
only while fetching the files of a library.

…stead of their utf-8 encoded versions

The method is now capable of using the branch and in-repo path (if
available) when checking the last commit of a library stored in library manager's
database.

…o be updated too.

`arduino_library_directory` parameter was replace by `builtin_libraries`
(for builtin libraries and examples, tracked on Github) and
`external_libraries` (for untracked external libraries, stored on AWS S3)

Made the library manager use the `builtin_libraries` and
`external_libraries` paths. Note that the external_libraries path is the
root to all the external libraries, unlike the previous implementation,
which was like
`arduino_library_files/external_libraries/libraries_directories`

- Added actual files to be used as library and example fixtures.
- Added a Symfony2 command, capable of purging pre-existing files from
  `external_libraries` directory and copying the fixture files there.