Fixed
- Text recovery for pages 10 and up. The per-page
.pagetext fragments are numbered in decimal, not hex like the background images; the extension had built them in hex since 2.9.0, so its own fetch for any page past 9 silently returned 403 and those pages only rendered when the viewer happened to fetch them itself. Verified against the viewer's own requests on a 63-page premium document. - The reader is no longer left at page 50 after the automatic page-load pass. The site's viewer wrappers do not scroll (the window does), so the saved position was restored to the wrong element. The download capture restores the position the same way.
- Studocu's own Download button is removed again. Its class hash changed (
Button-module-scss-module__…) and the match now goes by the button's label in any of the site's languages ("Downloaden", "Scarica", "Télécharger" and so on), while links such as "Download the app" are left alone. - Text fragments reference their figure layer relatively (
src="bga.png"), which resolved against the site URL and 403'd until the next sweep. The reference is rewritten to the signed CDN URL on insertion. - Printed PDFs use the document's own page size. Each sheet is sized from the page it carries (named
@pagerules, so mixed orientations work) instead of the browser's default paper, which left a blank strip at the bottom of every A4 page printed on Letter. - An empty premium banner wrapper that survived the banner removal is hidden too.
Added
- Toolbar popup with settings: the download button, automatic page loading, ad hiding, AI toolbar hiding and the logo replacement can each be switched off. Settings are stored in
chrome.storage.sync. - Esc closes the download preview; the preview shows the page count and a hint to choose "Save as PDF".
- Test suite (
npm test) with jsdom fixtures for the four document shapes: native PDF with premium-locked pages, native PDF fully readable, native PDF fully locked, and scanned. - Build script (
npm run build) that packagessrc/intodist/, a CI workflow, and a release workflow that builds, signs and publishes on av*tag. CONTRIBUTING.md,docs/ARCHITECTURE.md, issue and pull request templates.
Changed
- The extension source moved into
src/; content scripts are split intocommon.js,cleanup.js,pages.js,download.jsandmain.jsand share oneStudocuHacknamespace instead of duplicating selectors and URL logic. - Ads, the AI toolbar and Studocu's native download button are hidden by CSS gated on
html[data-sh-*]attributes, so toggling a feature needs no stylesheet changes. - The "could not find the document pages" alert is now an in-page notice.
- Fetched page text is inserted as sanitised DOM nodes rather than through
innerHTML. - The periodic re-check pauses while the tab is hidden.
- Manifest declares
data_collection_permissions: nonefor Firefox, which raises the minimum Firefox version to 142.
Removed
- The packaged
.xpiand.zipfiles from the repository. They are attached to each GitHub release and produced bynpm run build. remove-previewBanner.js, whose selectors and observer were a subset of the main script.- The in-page version button, which targeted a header layout Studocu no longer uses. The version is shown in the popup.
firefox-add.txt; the manifest has carried the Firefox settings since 2.7.0.
Install
Chrome / Brave / Edge: download studocuhack-v2.11.0.zip, extract it, open chrome://extensions, enable Developer mode, click Load unpacked and select the extracted folder.
Firefox: studocuhack-v2.11.0.xpi is unsigned in this release; release Firefox will refuse it until a signed build is attached. Use Developer Edition or Nightly with xpinstall.signatures.required off, or load it temporarily through about:debugging.
See the README for the full instructions.