Normalize oversized array index conversion errors

[fhgffy]

A long numeric array token passes the index syntax check but can exceed Python's integer-string conversion limit. The resulting ValueError escapes resolve()'s JsonPointerException handler, so a supplied default is ignored.

With the default 4300-digit limit, resolve_pointer([0], '/' + '1' * 4301, default=None) currently raises ValueError. This patch returns None; without a default it raises JsonPointerException.

The change catches ValueError only around int(part), preserving Python's safety limit. Five regression tests cover resolution, nested defaults, to_last, both inplace modes, and long numeric mapping keys, which remain valid.

Validation:

• On the unmodified base, four new test methods fail (five errors across the inplace cases); the mapping-key control passes
• Full unittest/doctest suite: 36 passed on Python 3.12.14 and 3.13.5
• flake8 passes; configured coverage run passes with 98% aggregate coverage
• Clean patch application and all 36 tests verified against 8c0bfdd
• Python 3.10, 3.11 and 3.14 runtimes were not run locally. The new tests skip if the integer-limit API is missing or its limit is disabled

[Copilot]

Copilot review overview

🟢 Approval recommended

The focused exception handling preserves valid mapping keys and is comprehensively tested across affected APIs.

Review effort: Balanced
Findings: None

What changed in this PR

Normalizes oversized array-index conversion failures into JsonPointerException, allowing defaults to work consistently.

Changes:

• Wraps sequence index conversion errors.
• Adds regression coverage for resolution, assignment, nested defaults, and mapping keys.

File	Description
jsonpointer.py	Converts oversized index ValueError into JsonPointerException.
tests.py	Adds oversized numeric-token regression tests.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[stefankoegl]

Thanks!