Multiple unwrap() panics on valid XLSX files from LibreOffice and Apache POI

[developer0hye]

Summary

We're using umya-spreadsheet (v2.3.3) in office2pdf to parse XLSX files. When processing a corpus of ~2,800 real-world XLSX files (sourced from LibreOffice and Apache POI test suites), 15 files cause panics due to unwrap() calls in the library. We've worked around this with catch_unwind, but the panics should ideally be handled gracefully.

Related to #271 (panic reduction proposal) and #301 (FileNotFound on .xlsm).

Panicking files (15 files)

FileNotFound panics (7 files)

unwrap() on missing zip entries for hyperlinks/charts:

• chart_hyperlink.xlsx, hyperlink.xlsx, tdf130959.xlsx, tdf115192.xlsx (LibreOffice)
• 47504.xlsx, bug63189.xlsx, ConditionalFormattingSamples.xlsx (Apache POI)

Stack trace points to zip entry lookup by relationship ID where the referenced file doesn't exist in the archive (e.g., external hyperlinks).

ParseFloatError / ParseIntError panics (3 files)

unwrap() on str::parse::<f64>() or str::parse::<u32>():

• check-boolean.xlsx — Boolean cell value parsed as float
• functions-excel-2010.xlsx — Integer overflow parsing cell reference
• FormulaEvalTestData_Copy.xlsx — Integer overflow parsing cell reference

unwrap() on None panics (3 files)

Expected XML elements/attributes missing:

• tdf100709.xlsx, 64450.xlsx, sample-beta.xlsx

dataBar end element panics (2 files)

Conditional formatting parser panics via panic!():

• tdf162948.xlsx, NewStyleConditionalFormattings.xlsx

Error: Could not find dataBar end element

Expected behavior

All of these should return Err instead of panicking. Library consumers should be able to handle failures without catch_unwind.

Environment

• umya-spreadsheet v2.3.3
• office2pdf v0.3.0
• Rust 1.85+
• Files sourced from LibreOffice and Apache POI test suites (publicly available)