Until now, if json item files couldn't be read, the script was
crashing. This commit delete the corrupted files as they do not
provide any useful information.

monitor_config_write: CORS apache conf should be rendered from a slapos template instead of being hardcoded

Revert "resiliencytest: adds tests on services status after takover for the slaprunner" and its fixup!

This reverts commit 26c5ef0b.
This reverts commit edc91312.

resiliencytestsuite: sleeping should happen before doing takeover, otherwise clone instances may not be ready

resiliencytest: test suite for slaprunner now uses a special SR for test, lighter than the previous helloworld SR

This reverts commit 30b6b285

/reviewed-on nexedi/slapos.toolbox!9

Revert "runner: if no "slapos-software" is defined in parameters, the auto build&run feature shouldn't work"

The current behavior should stay until a correct way of restarting processes in the webrunner (ie: watchdog within slapproxy) is implemented

This reverts commit 50948d5b.

Adds tests for the Webrunner, and updates its code to make it easier to test. These tests are still incomplete, but they have the merit of existing. I don't really like mocking function calls, but globally it's a better way than spawning a supervisord server and recreating the whole slaprunner's system tree.  

Also adds tools (= scripts) which will be usefull for the monitoring in slapos : 
  * generation of RSS feed from status items, and management of them.
  * script for creating promises based on RSS feeds 

/reviewed-on nexedi/slapos.toolbox!8

generatefeed: new helper script to easily create RSS2 Feeds from a bunch of well-formated JSON files