fix: support multi-space object headers

[vitormattos]

Summary

• Accept indirect object headers that use one or more whitespace characters between object number and generation (for example 1 0 obj).
• Keep robust fallback for slightly inaccurate xref offsets by searching nearby expected object headers.
• Add regression fixture and integration test.

Problem

• Some malformed PDFs contain object headers with extra spaces.
• The parser matched only a single whitespace in object headers, causing objects to resolve as null.
• This propagated to missing catalog/page detection.

Fix

• Relax object header pattern to accept one or more PDF whitespace characters.
• Use the actual matched object-header token length when advancing parsing offset.
• Preserve nearby-header recovery when xref offset is near the real object header.

Regression coverage

• Added fixture: samples/bugs/PullRequestNearbyObjectHeaderOffset.pdf
• Added test: DocumentIssueFocusTest::testParseFileWhenObjectHeaderIsNearXrefOffset

Validation

• make run-phpunit ARGS="tests/PHPUnit/Integration/DocumentIssueFocusTest.php --filter testParseFileWhenObjectHeaderIsNearXrefOffset"
• make run-phpunit ARGS="tests/PHPUnit/Integration/DocumentIssueFocusTest.php"

PDF Source

• Corpus: veraPDF-corpus
• Fixture origin: Isartor test files/PDFA-1b/6.1 File structure/6.1.8 Indirect objects/isartor-6-1-8-t01-fail-a.pdf
• Source URL: https://github.com/veraPDF/veraPDF-corpus/blob/staging/Isartor%20test%20files/PDFA-1b/6.1%20File%20structure/6.1.8%20Indirect%20objects/isartor-6-1-8-t01-fail-a.pdf
• Local copy used in tests: samples/bugs/PullRequestNearbyObjectHeaderOffset.pdf

[vitormattos]

Superseded by the RawDataParser consolidation chain in the fork.

This fix is included in vitormattos#30 (stacked on PR796 base via PR797→798→799→800).

Fixture consolidated to samples/bugs/rawdata/, test migrated to RawDataParserTest data provider.