fix: parse ova disk path from import output instead of fragile unused0 lookup

[endavis]

Description

Fix the OVA template's disk import phase to parse the disk path from qm disk import stdout (successfully imported '...' line) instead of reading unused0: from qm config. The unused0 approach is fragile when importing multiple VMDK disks in a loop — after the first disk is attached, subsequent imports may not appear as unused0.

Also adds set -e for fail-fast behavior and explicit error handling if disk import fails.

Addresses #370

Type of Change

• Bug fix (non-breaking change which fixes an issue)

Changes Made

• Parse disk path from qm disk import output instead of qm config unused0 lookup
• Add set -e to the import script for fail-fast on errors
• Add explicit error check if disk path extraction fails

Testing

• All existing tests pass
• Manually tested the changes

Checklist

• My code follows the code style of this project (ran doit format)
• I have run linting checks (doit lint)
• I have run type checking (doit type_check)
• All new and existing tests pass (doit test)
• My changes generate no new warnings

[github-actions]

Benchmark

Details

Benchmark suite	Current: 0a80c3a	Previous: cd476bd	Ratio
tests/benchmarks/test_placeholder.py::test_import_time	6964.507224353655 iter/sec (stddev: 0.000012275805525138683)	6872.315979448821 iter/sec (stddev: 0.000018039316691679812)	0.99

This comment was automatically generated by workflow using github-action-benchmark.