Summary
The XLSX benchmark case classic164_indic_scripts still showed a visible mismatch: the MiniPdf output kept the Indic sample column blank while the reference rendered Devanagari, Tamil, Bengali, Telugu, and Gujarati text.
Root Cause
MiniPdf embedded Unicode fonts by mapping Unicode code points directly to CID glyphs, but Indic scripts require shaping. A code point to CID mapping is not enough for these scripts, so the visual output could be missing or incorrectly formed even when text extraction stayed correct.
Proposed Flow
- Preserve extractable Unicode text in the PDF.
- Render Indic XLSX cell text through the platform text shaper into transparent PNG overlays.
- Commit the rendering fix with the refreshed benchmark artifacts.
- Push the commit and trace the benchmark result.
- Close this issue after the pushed fix is verified.
Validation Target
Run scripts/Run-Benchmark.ps1 -Filter "classic164" and confirm classic164_indic_scripts renders Indic glyphs in classic164_indic_scripts_p1_minipdf.png.
Summary
The XLSX benchmark case
classic164_indic_scriptsstill showed a visible mismatch: the MiniPdf output kept the Indic sample column blank while the reference rendered Devanagari, Tamil, Bengali, Telugu, and Gujarati text.Root Cause
MiniPdf embedded Unicode fonts by mapping Unicode code points directly to CID glyphs, but Indic scripts require shaping. A code point to CID mapping is not enough for these scripts, so the visual output could be missing or incorrectly formed even when text extraction stayed correct.
Proposed Flow
Validation Target
Run
scripts/Run-Benchmark.ps1 -Filter "classic164"and confirmclassic164_indic_scriptsrenders Indic glyphs inclassic164_indic_scripts_p1_minipdf.png.