[UPDATE] This does not work as expected. Doesn’t seem to run as a plain AppleScript and when re-worked into an Automator folder action workflow (word salad anyone?) it might work… It has and it also hasn’t. But, even when it has I have the sneaking suspicion iCloud does not propagate deletes from macOS. Wouldn’t be the only inexplicable behavior from that service.

[UPDATE]This has been moved to a Google Drive folder. I’ve had enough of ghost folders taking up all the space. The principle is however the same, except now it just works. Boom. Roasted.

I have a security camera that FTPs a new folder everyday into a local iCloud folder on my home server. FYI, the real location of your iCloud Drive is within ~/Library/Mobile Documents, which is a package folder.

This storage process has not proved ideal since iCloud doesn’t handle automation very well. It works with Finder windows and Save dialogs, but background processes not so much. The result for me has been constant reminders that my iCloud drive is “full” despite deleting the bulk of folders. Most of the space never comes back and many times folders just “deleted” reappear.

Maybe adding another automated process to this soup is not going to help and may make things worse (will bird even notice when items are removed this way?), but what the hell. This script will delete files oldest first until the number of items it counted when it was called reaches the “keep” count (set to “5” here).

This may be attached to any folder, and items may be sent to the Trash or deleted immediately with rm. but be aware that it does not send files to the Trash. They are removed with rm.

-- http://strawhousepig.net/

on read_array(the_dom, the_key, the_val, cycle)
try
set last_run to (do shell script "defaults read " & the_dom & " " & the_key)
on error -- Will error if there is nothing to read.
return false
end try
if the_val - last_run is less than cycle then
return true
else
return false
end if
end read_array

[UPDATE] Didn’t realize how dangerous this is if you actually need to get a backup file(s) off the drive rather than restore from catastrophic failure. Now prompts to confirm sync before proceeding. Should also prompt to eject when complete.

The previous AppleScript post is an earlier version of nearly the same thing. I posted it here to be able to write a new script for where I work. Its intention is to keep an on-site duplicate of an off-site backup. This can of course be used for simple backups and/or folder syncing operations if given the appropriate trigger(s).

To have this script run automagically attach it as a Folder Action to the Volumes folder (/Volumes).

Also, as the previous post, this script calls an rsync binary in a custom install location. Please be aware of that.

Because I can: rsync manual

-- The name of the volume to watch for.
property trigger : "Backup"

on do_it(skip)
tell application "System Events" to set trigger_path to POSIX path of every disk whose name is trigger -- When a drive is mounted and its previous mount point was not removed the new mount point will be given a unique number to follow the name (which isn't shown in the Finder), e.g. "/Volumes/Backup 1". This asks the System for the path rather than us blindly using text.
if trigger_path is {} then
display dialog "The drive '" & trigger & "' does not appear to be mounted. :("
return
end if
set rsync_options to rsync_options & " "
repeat with i in theFolders
tell application "Finder"
if not (exists (i as POSIX file)) then
set end of skip to i as text
end if
end tell
log skip
if i is not in skip then
try
do shell script "/usr/local/bin/rsync " & rsync_options & quoted form of i & " '" & trigger_path & "/'"
set end of skip to i as text
on error the_err
if i is last item in theFolders then
display dialog "rsync encountered a problem: " & the_err
else
if button returned of (display dialog "rsync encountered a problem: " & the_err & return buttons {"Cancel", "Continue"}) is "Continue" then
set end of skip to i as text
do_it(skip)
end if
end if
end try
end if
end repeat
set skip to {}
end do_it

This script includes error handling that isn’t well tested. Getting it to work with it at all was a bit of a trick (particularly if not (exists (i as POSIX file)) since if not (exists POSIX file i) does not work as expected). So that whole “skip” idea may not work as intended if you are asked to continue. The last do_it(skip) call may need to be my do_it(skip) instead.

This is to correct file times on photos that were imported from cameras with bad date settings or when they have been transferred to-and-fro and picked up the wrong data some where or after re-encoding high bit rate video to save a few GBs. [This does not alter the date atoms encoded in (most?) video files. ffmpeg can do that, however.]
Continue reading →