Critical Persistence of Conversion Error Across Windows 11 Devices and Registry Clean

[ATHENA007]

Dear Kovid Goyal,

I am writing to report a conversion issue that seems to defy standard troubleshooting. Despite performing what should be a complete reset, a corrupted Regular Expression setting continues to haunt my Calibre environment.

System Environment: Windows 11

The Persistent Error:
Due to a faulty Regex applied previously, unique IDs in my HTML tags (such as h2 id) are being deleted during conversion. This causes chapters to merge without page breaks and breaks the Table of Contents.

Actions Taken (All Failed to Resolve the Issue):

* Path Changes: I have changed the installation and library storage paths between internal and external hard drives.
* System Cleaning: I have manually cleared all cache files and deleted Calibre-related keys from the Windows Registry.
* Reinstallation: I have performed multiple clean uninstalls and reinstalls.
* Hardware Migration: Most surprisingly, I moved to a completely new computer with a fresh Windows 11 installation, yet the same broken conversion behavior reappears as soon as I use my existing library.

My Conclusion and Request:
It appears this ghost setting is not stored in the program files or the OS registry, but is embedded directly within the library's metadata database (metadata.db) or the individual book folders.

1. How can I forcefully purge these per-book conversion histories and Search and Replace memories from the library itself?
2. Is there a way to reset the entire library's conversion properties to factory defaults without losing the books themselves?

I look forward to your expert advice on how to finally erase these persistent, corrupted settings.

[kovidgoyal]

https://manual.calibre-ebook.com/con...for-conversion

[BetterRed]

Edit metadata -> Edit metadata in bulk



Added: I use the attached template function to show an icon if the book has stored conversion options.



BR