Version 1.0.1
Latest 1 September 2026A correctness and reliability release. No new tools — everything here comes from a full audit of the app: work that could be lost, output that was quietly wrong, and a Store hiccup that could stamp the free-tier footer onto a document you had already paid to save clean.
Fixed — lost work and crashes
- Undo could silently discard a merged document. Undoing back to the start marked a document as unchanged even when it had never been saved — and Merge and Images → PDF build entirely in memory — so closing the tab threw the work away with no prompt. Undoing past a Save As had the same effect. A document now tracks what is actually on disk.
- Random crashes after using Annotate. An internal method name collided with one Windows reserves for cleanup, so every annotate window ended up running that code later, on the wrong thread.
Ctrl+Ccould close the app. Copying selected text failed hard when another program was holding the Windows clipboard. It now retries and reports instead.- An unexpected error no longer ends the session. Your open documents stay open, and the technical detail is written to a log file you can attach to a bug report.
Fixed — wrong output
- Form filling. Check boxes and radio buttons were written in a form no viewer recognised, so they never appeared ticked anywhere, and clearing a form wrote an empty value where “off” belongs. Filled values are now also visible in ExecutiveDesk’s own page view, not just stored.
- Rotated pages. Watermarks, page numbers, Bates numbering, headers and footers were placed against the page’s unrotated shape, so on a rotated or landscape-scanned page they ran down the side of the sheet.
{file}in headers and footers printed an internal working filename instead of your document’s name.- Sanitize left the originating software’s name in the file it reported as clean, and missed JavaScript attached to links and form fields.
- A damaged file was reported as password-protected — the app asked for a password, then rejected five attempts, for a file that had none.
- Page image export and OCR silently dropped annotations and filled form values, because those rendering paths did not include them.
Fixed — purchases
- A Store hiccup could put the free-tier footer on a paid save. The check treated any failure to reach the Store as “this user has not bought Pro”, on every document you opened. It now tells “the Store said free” apart from “the Store could not answer”, never downgrades on a failure, and remembers a confirmed purchase across an offline launch.
- Batch processing ignored the free-tier footer entirely, and overwrote a previous run’s
merged.pdfwithout asking.
Changed — security
- Encryption is now AES-256 (it was AES-128), and document metadata — title, author — is encrypted with the rest of the file instead of being left readable without the password.
- Permissions can no longer be decorative. Restricting what a reader may do requires an owner password different from the open password; if they are the same, everyone who can open the file holds owner rights and the restrictions mean nothing. ExecutiveDesk now says so before you save, rather than producing a file whose settings do nothing.
- A protected document is no longer auto-saved for crash recovery. Opening one with its owner password produces a decrypted working copy, and that must not be written to disk. The trade-off is stated plainly: an encrypted document lost to a crash cannot be recovered. Recovery data left behind by a crash is also deleted after seven days.
Added
- Open PDFs directly. ExecutiveDesk now appears in Windows’ Open with, can be set as your default PDF reader, and opening a second file adds a tab to the running window instead of starting another copy of the app.
- Progress and Cancel. Long operations show a progress bar with a Cancel button in the status bar. OCR reports its progress page by page.
- Plain-language errors, with the technical detail written to the log rather than shown in the status bar.
- Ctrl+O opens a file. Print already answered to Ctrl+P; Open had no shortcut at all.
Performance
- Opening a long document is about twice as fast. Every page render used to re-read the whole document; the thumbnail pass, image export, flatten and OCR now make a single pass. Measured on a 200-page document: 2.0 s down to 1.0 s.
- Searchable PDFs are much smaller. Pages were embedded losslessly, so making a scan searchable used to make it bigger. A three-page test now produces 0.37 MB from a 0.64 MB scan.