Hacker News new | ask | show | jobs
by idorosen 1446 days ago
I use getmail (similar to fetchmail) to routinely archive (i.e., sync without deleting + reindex) all of my emails from various free accounts, just in case. It can save to mbox, Maildir, mh, and other formats that are easy to import to any MUA/LDA. This is worth doing for all e-mail, and I have a patch to make it support OAuth. I don’t think it supports JMAP, but it’s great for IMAP, Gmail, and Yahoo mail and deduplicates messages by ID and content, etc. while preserving tags/mailboxes (if saved as Maildir). I highly recommend running something like that in a cron job somewhere once a week to sync locally with some sanity checks (e.g., did it save any new messages? did the folder grow? Etc.)

I also use it to save Spam/Junk folders, which then comes in very handy to train my local spam classifier for my self-hosted mail servers with lots of data. (Over 3TB of spam saved so far and about 20GB of ham.)

Gmail’s spam filter has had a higher false positive rate than usual for me lately, so I have a little report emailed to me once a week of likely ham in my gmail spam box, which has found at least 3 messages per week that I missed.