[JFWooten4]

Summary

Preserve open PDF tabs across unpacked-extension reloads, including local file:// PDFs and web PDFs.

• Register each PDF's original source URL and current physical page without storing document bytes.
• Track tab window, index, pinned state, and active state; reopen Chrome-closed tabs or refresh surviving tabs in place.
• Support both the direct MIME-handler path and the older extension-page redirect fallback, without depending on temporary stream/blob URLs.
• Skip local-file reopening when Chrome's Allow access to file URLs permission is disabled, and never overwrite unrelated tabs.
• Remove intentionally closed tabs from the recovery registry, avoid duplicate tab recreation during the second recovery pass, and skip incognito persistence.
• Add six regression tests and document the file-permission requirement.

Verification

• Verified all six committed file changes and that the branch's Git tree differs from main.
• Confirmed the recovery worker script parses successfully and the package manifest remains valid JSON.
• Added tests/pdf-tab-recovery.test.mjs to npm test. Automated Node tests have not been executed in this environment.

Limitations

• The first reload after installing this feature cannot recover tabs that an earlier viewer version never registered.
• Local PDFs need file-scheme access permission and must still exist at the saved path.
• Reloading a surviving tab may re-fetch the PDF; recovery does not cache PDF bytes.