When logging, always have something static

cert-shifter, when there is no change to the cert, mention that.

This makes log filtering easier when you can latch onto something.

Do not return 1 when doing cert-puller -s

This caused me some issues in an ansible script.

Allow user-specified fetch options

• Assign empty values to SERVICES, SERVICES_RELOAD, & SERVICES_RESTART.
• Provide methods to supply options for curl, fetch, and wget
• allow other services restarted as well #6

Ignore DAYS

Ignore the DAYS configuration item for cert-shifter. It is now no longer used.

Why? Because we cannot depend upon the mtime of a directory to indicate that
it contains a new file. Instead, we must look at the certs contained within
those directories.

For postresql, do a reload.

0.0.16

Resolved conflict

Preserve the old behavior

Set a default of 1 for the number of days we search back.

This preserves the old behaviour.

Document the new configuration item.

Add support to reload postgresql

Detect restart/reload errors and report them

Support for Apache24, better logging

Log a message if the mv or cp fails.

This will give you some clues if the expected certs are not showing up.

Add support for mosquitto

Add support for restarting mosquitto

While here, sort the services

Fix quoting

0.0.11

Add missing quote on user-agent for fetch