fix(linbo_mkgrubimg): handle hosts of non-default schools

[TomlDev]

Part of #178. As agreed there, this fixes existing bugs in linbo-mkgrubimg for non-default schools first, before the follow-up PR switches it to the school-qualified group.

Problem

linbo-mkgrubimg -n <host> -w <school>.devices.csv [-s] did not work for a school other than default-school:

• getHostname() returns the hostname from devices.csv, lowercased and without a school prefix. linuxmuster-import-devices writes non-default schools' hosts with a <school>- prefix in DHCP and hostcfg. The image was created as hostcfg/pc01.img instead of hostcfg/abc-pc01.img, and --setfilename never found the host abc-pc01 DHCP entry, so it silently changed nothing.
• -n was looked up while the options were still being parsed. With -n pc01 -w abc.devices.csv, the host was searched in default-school's devices.csv.
• --setfilename dropped option nis-domain when it rewrote the DHCP host entry, although import-devices writes it next to extensions-path.

Change

• New getSchoolFromDevicesFile() derives the school from the workstations file name, using the same path convention as base7's readDevicesCsv() (<school>/<school>.devices.csv, otherwise default-school). No new option is needed; -s is already taken by --setfilename.
• -n is looked up after all options are parsed. For a non-default school the hostname gets the <school>- prefix.
• The rewritten devices.csv row keeps the hostname exactly as it was read. Before, it was lowercased, and with the prefix it would have become abc-pc01, which the next import turns into abc-abc-pc01.
• The rewritten DHCP entry keeps option nis-domain.

default-school: the only difference is that the devices.csv row keeps the hostname's original case.

Testing

• pytest tests/python: 102 passed, including new tests for getSchoolFromDevicesFile().
• Ran main(['-s', '-n', 'pc01', '-w', '<tmp>/abc/abc.devices.csv']) against temp files with the real base7 functions. The environment paths pointed at the temp dir, and grub-mkstandalone, the dhcp restart and base7's third-party imports (IPy, ldap3, ...) were stubbed. The image is hostcfg/abc-pc01.img, the host abc-pc01 entry gets its filename and keeps nis-domain, and the devices.csv row keeps pc01.

Not changed (existing, separate)

• With -s, a host whose IP is DHCP is not updated in devices.csv, because ip is replaced by 0.0.0.0 before the old row is built for the search.
• The check 'host ' + hostname in content also matches longer hostnames (pc01 matches pc010).

[HappyBasher]

Thanks, I reviewed this and ran the tests locally: pytest tests/python gives 102 passed. The school lookup follows the same path convention as base7's readDevicesCsv(), and keeping the devices.csv row as read is the right call.

We'll merge once the lab test is done. For this PR that means a real run of linbo-mkgrubimg -s -n <host> -w <school>.devices.csv on a server with a non-default school, checking that hostcfg/<school>-<host>.img is created and the DHCP entry for <school>-<host> gets its filename and keeps nis-domain.

Before merging, please change two things in the commit:

1. Add a Signed-off-by: trailer. It has to come from you as the author, so we can't add it for you. git rebase --signoff origin/7.4 and git push --force-with-lease will do it. We merge contributor PRs with a merge commit, so the commit message stays in the history.
2. Replace Co-Authored-By: Claude <noreply@anthropic.com> with Assisted-by: Claude, as in your file headers. We don't use Co-Authored-By trailers.

One note for the follow-up PR: groupcfg, start.conf.<group> and extensions-path still use the unqualified group here. The follow-up that switches to the qualified id has to land together with the base7 helper, otherwise --setfilename writes a group that no longer matches the import.

[TomlDev]

Lab test on a linuxmuster 7.4 server (linuxmuster-linbo7 7.4.17, linuxmuster-base7 7.4.8), school agy, host PC02 (devices.csv row R101;PC02;win11_2026;..., existing hostcfg/agy-PC02.cfg and DHCP host agy-PC02). I ran the script from a scratch copy against a copy of agy/agy.devices.csv: linbo-mkgrubimg -s -n PC02 -w <copy>. The installed script and the real devices.csv were not touched, and the hostcfg and DHCP files were restored afterwards.

The first version of this PR failed the criterion for host names with uppercase letters. getHostname() returns host.lower(), while readDevicesCsv() and import-devices write <school>-<name> with the case of the devices.csv row, so the image became agy-pc02.img and the case-sensitive 'host agy-pc02' in content match found no DHCP entry. The updated commit takes the prefixed name from the devices.csv row (getQualifiedHostname(), with tests); default-school keeps the lowercased name.

Result with the updated commit:

• hostcfg/agy-PC02.img is created ("successfully created").
• The DHCP entry host agy-PC02 gets filename "boot/grub/hostcfg/agy-PC02.img" and keeps option host-name, option extensions-path "win11_2026" and option nis-domain "win11_2026".
• The -w copy gets the filename in field 8 of the PC02 row.
• default-school's host pc02 entry is unchanged (identical before and after).

For comparison, the installed 7.4.17 script with the same arguments creates hostcfg/pc02.img and rewrites default-school's host pc02 DHCP entry with agy's MAC and IP, dropping its nis-domain.

Unrelated to this PR: getHostname() fails with list index out of range (printed, then the usage text) when the devices.csv contains a blank line.