OCR It 0.3.0 — Firefox support

Firefox support: one source tree now builds for both browsers.

Both stores are reviewing this version. Until they finish, the packages below
install by hand.

Firefox — download ocr-it-firefox-0.3.0.zip, open
about:debugging#/runtime/this-firefox → Load Temporary Add-on… and pick the
zip. Firefox unloads a temporary add-on when you quit, so this is a stopgap until
the signed AMO listing lands.

Chrome — download ocr-it-chrome-0.3.0.zip and unzip it — Chrome cannot
install a zip. Then chrome://extensions → turn on Developer mode →
Load unpacked → select the unzipped folder. It stays installed, though Chrome
will nag about developer-mode extensions on each launch.

In this release

• Firefox build. The OCR engine moved to src/ocr/, and the one real fork is
  where it runs: an offscreen document on Chrome, the background event page on
  Firefox.
• Full-size viewer for the crop behind a page's thumbnail.
• Benchmark against raw Tesseract and RapidOCR: https://thiagotigaz.github.io/ocr-it/bench/
• Page records no longer store the URL or title of the tab they came from, and
  updating strips both from anything an older version saved.

Requires Firefox 140+ or Chrome 116+. OCR runs entirely on your machine.