EDIT: Whilst I’m very attached to my theory, it’s wrong
See Tidda’s comment in the next post.
This bit is validish though:
The noauto
@philip1 mentions is a related to file system mounting. As such, you wont see it show up in a partitioning tool, since it’s actually not a partition flag or label etc (it basically lives at least one level higher in the heirarchy, as it’s associated with the file system that sits on top of a partition). You can see these options listed against the mount points here:
grep '/dev' /proc/mounts
Incorrect stuff below
My understanding at present (and I could be wrong here), but I think Clear only looks at the partition labels (for instance CLR_SWAP) as part of the installation process of clr-installer
(or it’s GUI equivalent).
Those labels just serve as a way of identifying to Clear which partitions the user wants to assign for different purposes. As part of the install script for OS config, I guess it probably ceases to use the labels and likely correlates each known label, to it’s specific UUID or PARTUUID and uses that from there on out. UUID and PARTUUID is consistent across reboots, and much less likely to suffer name/identity collisions…at least; less likely than user applied and editable partition labels.
So, that’s my reasoning as to why they would probably do it, but I don’t know, it’s a bit beyond my knowledge to hunt through the source of the install scripts at this stage.
So to spell out my theory, I guess if you label a partition as CLR_SWAP, during the install script the UUID or PARTUUID specific to that label is probably used to build native systemd unit files (thus avoiding /etc/fstab) to mount that resource (such as swap). The corollary then would be (based on my limited understanding), that this is a one shot deal during the install process, it probably does not implement boot time scripts for successive boots in order to look for user partitions that get added at a later date, instead assuming you will take care of that in /etc/fstab, with your own systemd unit files, or (as per @miguelinux’s comment) by relying on systemd implementations that work for more common topologies:
Usually systemd mounts swap automatically when it is in the same
drive as the root partition.
Anyway, if one of the dev team are bored, they could probably clarify all this, but I suspect they are busy with other things.