A 01:03 Phone Call, a Long Ace Shift, and Fixing How My Websites Send Mail
A late night, a long shift, and a handful of real fixes to things that quietly weren't working right.
A 01:03 phone call
At 01:03 I got a call from a friend who'd just landed back from a trip. His ride home turned abusive on the way back, bad enough that he demanded to be let out and wait on his own. I went out and picked him up. He's back home safe.
A longer Ace shift than planned
Worked at Ace Hardware from 08:00 to 17:00. I was only scheduled until 13:00, but stayed longer to help get Tuesday's delivery put away.
Fixing how my websites send mail
Spent part of today fixing a real weak spot in how the contact form and agreement alerts on my business website work. A key that let anyone send mail was sitting in the page's own code, where any visitor could find it. Replaced that with a proper setup: the form now proves it's a real visitor with a "prove you're human" check and a few other spam filters, and mail only gets sent through my own server. The contact form still falls back to a plain email link if the send ever fails, so nobody gets stuck.
While I was in there, also fixed a printing problem: my site is dark-themed, and printing one of the service agreements kept that dark background and light text, so the headings and signature lines came out nearly invisible on paper. Fixed the print version to use dark text on a plain white page instead.
Also repriced the retainer plans I offer for ongoing IT support — the old prices actually worked out more expensive per hour than just paying for the hours directly, which defeats the point of a retainer. Fixed that so a retainer is a real discount again. And rewrote the page about my CARF experience (CARF is an accreditation for health and human services organizations) to describe what I've actually done — writing policy at a CARF-accredited employer — instead of implying things I can't back up.
A better way to catch real network problems
Also fixed something in my VPN software that had been filing too many false alarms. Before, one missed check could look like half the connection had failed and open a support ticket. Now it looks at the last minute of checks and asks how many actually failed, so it reports what's really happening instead of overreacting to a single blip.