TeX \input Stack and Archived Source Segments: Verifying Restored Files Preserve Pagination

To verify that restored TeX segments preserve original pagination, you must move beyond basic compilation checks. Because TeX treats `\input` as a transparent byte stream, silent failures—like truncation or re-encoding—can shift layout without triggering a fatal error if the structural integrity (closed environments and groups) remains accidentally intact.

Detecting Truncation vs. Short Files

The most reliable method for detecting a prematurely truncated file versus a legitimately short one is using the tracingall command. When tracingall is active, TeX logs every file read and the line number processed to the log file.

• Truncation Detection: If a file is truncated, TeX will likely throw a Run out of control sequence or Missing }group error if the cut occurs mid-macro or environment.
• Valid Short Files: If the file is simply short but valid, the log will show a clean end-of-file (EOF). By comparing the line count in the log against the original build log, you can identify exactly where the stream diverged.

Fidelity Checks: Auxiliary and PDF Checksums

Comparing auxiliary files and PDF checksums is a high-fidelity check, but it has specific nuances:

• PDF Checksums: If the MD5/SHA-256 of the PDF matches the original, the pagination and content are identical. This is the gold standard for verification.
• Auxiliary Files: Comparing .aux, .toc, and .log files is useful for detecting shifts in cross-references. However, two different sources could theoretically produce identical checksums if the changes occur in metadata or non-visual comments that do not affect the layout engine.

Impact of LaTeX Format Versions

Restoring and building under a different LaTeX format version (e.g., moving from TeX Live 2022 to 2024) can change pagination despite files being byte-identical. This happens because:

• Catcodes: Newer formats may change default catcodes for certain characters, affecting how macros parse input.
• Encoding: Changes in the underlying font metrics or Unicode-mapping between versions can subtly alter kerning and word spacing, which eventually pushes a line of text to the next page.

Recommended Verification Workflow

1. Run the original build with \tracingall and archive the resulting .log.
2. Restore the segments and recompile using the exact same engine and format version.
3. Use a `diff` tool to compare the new log against the original to ensure line-number parity.
4. Verify the PDF checksum to confirm the final visual output is identical.