Each tool needs to be independent, which
meant that every time a tool was called,
it would build the entire system graph.
This was an obvious waste of time and resources.

Now, the tools build the graph based
on the requested services, which
drastically reduces the time needed to
construct the necessary dependency graph.

This program is executed via the run script in the service's live directory. It sits between s6-supervise and the user's executable script, allowing for the preparation of the user's script execution. This introduces a dual capability:

    Define certain preparations through frontend keys, such as setting ownership, moving to a working directory, limiting resources used by the service, etc.
    Since 66-execute runs on each signal received and transmitted by s6-supervise, it is now possible to send messages to an event management daemon.

By default, this program performs previously defined actions such as logging redirections and managing environment variables (previously handled by excel-envfile), creation of the log destination directory, and so on.So, the user's file now contains only the exact copy of the @execute field.
Also, the name of the script for an start or
stop process are the same. This was a old behabvior
coming from s6-rc  manager.

Permit to set uid and gid for a run script.
Unknown key may occur at end of the file.
Avoid to changed pointer.
Typo fix.

These directories is used to properly define environment variable
at start of a scandir

This reverts commit 87bd69c6.

use tmpfs directory as service directory for the runtime. Deal with logger destination directory at init time

simplify interface of service_is_g and its functionality, provide service_get_treename as replacement