Fix AC/battery detection logic on FreeBSD.

[fernape]

The logic for the detection of the battery/AC line was not completely correct for FreeBSD. Launching conky on console shows the following:

Cannot read sysctl "hw.acpi.battery.time" Cannot read sysctl "hw.acpi.battery.life" Cannot read sysctl "hw.acpi.battery.state" Cannot read sysctl "hw.acpi.acline" Unknown battery state 8!

In a PC, the hw.acpi.battery MIB does not exist. Also, the hw.acpi.acline is only present if supported by the hardware. In addition, some variables were used uninitialized and that causes strange behavior: in a PC it showed it worked on battery and the % of charge was an ridiculous big number.

This patch addresses the issue. It fixes the problem in the PC. It has also being tested in a laptop running FreeBSD current plugin and unplugging the AC line and also snatching the battery mercilessly to see if something breaks.

Checklist

• [x] I have described the changes
• [x] I have linked to any relevant GitHub issues, if applicable
• [x ] Documentation in doc/ has been updated
• [x] All new code is licensed under GPLv3

Description

• Describe the changes, why they were necessary, etc
• Describe how the changes will affect existing behaviour.
• Describe how you tested and validated your changes.
• Include any relevant screenshots/evidence demonstrating that the changes work and have been tested.

[netlify[bot]]

✅ Deploy Preview for conkyweb canceled.

Name	Link
🔨 Latest commit	f1f9c41cd527e798b7b6574f8a290114d7f9bd7a
🔍 Latest deploy log	https://app.netlify.com/sites/conkyweb/deploys/666c7ffd921e8d00088f9e7f

[fernape]

Overall it's good, but since you're making changes here we should replace those fprintf() calls with NORM_ERR instead, if you wouldn't mind.

Thank you for the suggestion. I didn't know about NORM_ERROR and the fprintf() were already there. I'll take that into account.