Dropbox repository downloads files larger than max_cache_size on *every* run of cron

The dropbox repository cron function has this logic:

for all files in dropbox repo {

  if the file isnt in moodles file pool {

    download the file to check its file size

    if file is less than max cache size (say 1mb) store it in the file pool

  }

}

What this means in practice - is that all files larger than the cache size get their filesize checked on every invocation of the cron. So if you have a 1GB file for example - Moodle will download that every time the legacy cron tasks run, forever.

Relatedly, the import uses file_put_contents - which requires that the content being saved be able to fit into the memory limit of the cron process.