[bug#27690] [PATCH] build: Allow mounting of entire disks.

[ludo@gnu.org (Ludovic Courtès)]

Hello,

Danny Milosavljevic <dannym@scratchpost.org> skribis:

> * gnu/build/file-systems.scm (disk-partitions): Also return entire drives.

In the subject line, please write “file-systems:” instead of “build:”
(the latter is for Makefile and configury.)

> --- a/gnu/build/file-systems.scm
> +++ b/gnu/build/file-systems.scm
> @@ -377,11 +377,14 @@ not valid header was found."
>      (string-ref str (- (string-length str) 1)))
>  
>    (define (partition? name major minor)
> -    ;; Select device names that end in a digit, like libblkid's 'probe_all'
> -    ;; function does.  Checking for "/sys/dev/block/MAJOR:MINOR/partition"
> -    ;; doesn't work for partitions coming from mapped devices.
> -    (and (char-set-contains? char-set:digit (last-character name))
> -         (> major 2)))                      ;ignore RAM disks and floppy disks
> +    ;; grub-mkrescue does some funny things for EFI support which
> +    ;; makes it a lot more difficult than one would expect to support
> +    ;; booting an ISO-9660 image from an USB flash drive.
> +    ;; For example there's a buggy (too small) hidden partition in it
> +    ;; which Linux rightfully refuses to mount.
> +    ;; In any case, partition tables are supposed to be optional so
> +    ;; here we allow checking entire disks for file systems, too.
> +    (> major 2))                      ;ignore RAM disks and floppy disks

I think you can remove ‘last-character’ and keep the comment about
/sys/dev/block.

Have you tested that this doesn’t cause any problems on actual hardware,
like matching things that should really be skipped, and then causing
failures down the road in label/UUID code?

Also, it seems that we’d be departing from what libblkid is doing, per
the comment above.  To me that raises a flag, though if we really need
it that’s fine of course.

Thoughts?  :-)

Thanks,
Ludo’.