Disabled fstrim on BTRFS by default

[Torxed]

• Fixes No need to enable periodic TRIM for btrfs #1837
• Fixes i am encountering fstrim error #1841

[lavafroth]

Looks great. I overlooked the case where the filesystem is not BTRFS. Good catch!

[kojq]

looks good; just heads up that this could go beyond btrfs as new filesystems such as bcachefs, PuzzleFS, and/or SSDFS could be implemented into the kernel later in some way, and some future filesystem(s) may not need periodic trim as well. You could consider changing the variable name to _need_periodic_trim or something else, and consider reversing the boolean values (this also remove the not).

I will approve otherwise, if desired.

[Torxed]

Good idea, changed it to _disable_fstrim to be more generic. As we for most filesystems enable it, I thought of keeping the naming consistent with the if-logic and default value of being False for now.

[kojq]

Good idea, changed it to _disable_fstrim to be more generic. As we for most filesystems enable it, I thought of keeping the naming consistent with the if-logic and default value of being False for now.

Great. Also, is there a particular reason we want a variable, as I don't see anything technical like a scope issue with directly using self.enable_periodic_trim() when checking for btrfs?

[Torxed]

Are you referring to why we don't do the check outside before calling enable_fstrim?
The only reason is to keep minimal_installation small and easy to read :)

And if enable_fstrim is called as part of a custom script it's good if we do the check in the function to not cause issues for lib users.