I’m not too eager to put in an initramfs as it will slow down the boot and could break some things. If it’s to be done, it would be done near the end of the year.
For the systems I manage I just modified /etc/fstab to run it every time. On a clean file system it is near instantaneous. This saves me a lot of headache and gas (driving to fix family members PI’s.)
Just to add, I have some memory cards that were stuck in read-only mode with corruption. I edited the fstab in a VM, booted and the filesystems were repaired and then booted!
We are investigating the safety of enabling it by default.
It’s not as simple a change as it appears though, first of all we would have to parse your (possibly customised) copy of fstab and attempt to change only those two lines (this can be done with sed, but we have to be very careful) secondly enabling automatic fsck on boot potentially means that if the corruption is severe enough that it will stop the bootup process and ask the user some questions on how to proceed with the repair - which is a big problem if you (a) don’t have a keyboard to answer the questions and (b) can’t read the questions due to the overscan on your TV…
Automated fsck with forced repair does come with it’s own risks.
However if you have reached the point of corruption, your main choice to repair the filesystem is to unmount the microsd, mount it elsewhere, repair it, then return it.
For average joe with no huge skills or access to other linux machines, this means the usual solution is to completely reimage.
For user-friendlyness, this option should be changed.