Make Archive.php use PHP CLI instead of Web, more reliable / performant

[mattab]

The implementation of our Cron Archive script should use Php CLI instead of doing Http requests to the webserver. This would remove one system in the middle and should make archive script slightly more reliable.

Because the piwik.php webcron also uses the CronArchive logic then either we change this logic, or keep the CURL wrapper in the CronArchive (when Php CLI not available or exec() disabled).

[tsteur]

In 238e71a: refs #4610 trigger archiving using cli instead of http

[tsteur]

In 19385b0: refs #4610 worked on executing cli commands in parallel but not tested yet

[tsteur]

In 65d4083: refs #4610 simplified the fallback if async is not supported, tests and fixes for output class

[tsteur]

In 095af95: refs #4610 fix headers

[tsteur]

In 9cba66f: refs #4610 use existing methods, lock does not work on Windows (yet), save locks in a separate folder

[tsteur]

In 279a4d6: refs #4610 I moved the Lock class to CliMulti as it does not 100% work on windows and is only used in CliMulti so far. As the Lock class can be quite useful and maybe needed in other use cases it would be nice to have it somewhere under core or so as it is not directly related to CLI. We might have to move it later again. Need to find a better solution for windows

[tsteur]

In 113e403: refs #4610 tests for cliMulti and loads of bugfixes

[tsteur]

In f06a7a1: refs #4610 Pid => Process

[tsteur]

In a2a586a: refs #4610 make sure the fallback mode works in case async is not supported

[tsteur]

In dffb297: refs #4610 use a command to execute a request, this way we do not have to bootstrap everything again and we already make sure it is only executed from the command line

[tsteur]

In fdc866a: refs #4610 make sure the command awk exist

[tsteur]

In 6247a10: refs #4610 store everything in climulti folder for easier cleanup later, some other optimizations

[tsteur]

In cb994c9: refs #4610 prevent some notices like "failed to read from stream" which happens if a subprocess tries to write to a file and the archiver just tries to file_get_content at the same time. So file_exists returned true just a few ms before but while trying to read it fails

[tsteur]

In 7947a56: refs #4610 if a process does not start after 8 seconds, finish it as the process probably will not start anyway

[tsteur]

In f86af8b: refs #4610 do not start the actual process in case it was marked as finished meanwhile