fix: preserve whitespace in XML text content

[sueun-dev]

from_xml() strips leading and trailing whitespace from properties mapped to the node text with xml_name('.'), before applying the configured XML string handling. For example, a normalized-string value of " hello " is read as "hello", and a value containing only spaces becomes empty. This also changes Property.value when roundtripping a BOM through CycloneDX Python Library's XML output.

Remove the unconditional strip() and let the configured string handler decide normalization. TOKEN still collapses whitespace; plain strings and NORMALIZED_STRING retain surrounding spaces. The tests cover both affected types, whitespace-only text, and token/tab/plain-text controls.

Checked on Python 3.12/Linux arm64:

• Three regression failures before the change; all six regression/control tests pass after it.
• Serializer suite: 97 tests pass; flake8, isort, pyupgrade, and mypy (default and Python 3.8 target) pass.
• CycloneDX Python Library at 52cb3c94f023df887ac65a6125bce4d63ab7857e: 6,984 tests pass with the patched dependency, plus 120 generated JSON/XML BOM roundtrips checking component fields and dependency edges.

The other Python/OS CI combinations were not run locally. AI assistance: OpenAI Codex produced the investigation, patch, tests, and this description; no human review is claimed.