Repository navigation
SignuDoc 0.5.2
Twenty-four bug fixes. Seventeen fix reported issues (sixteen from @coygeek, one from @seanGSISG), and ten came as pull requests from five outside contributors. This release adds one database column: see the end of these notes.
Signing
- Opening an old invitation after the envelope was signed, completed, declined or voided says what happened to it, instead of asking the signer to request a new link the sender cannot send (#110, PR #178).
- Reloading the signing page before Finish asks first, instead of throwing away the signature and answers without warning (#99, PR #169).
- A long Text answer no longer widens the signing page past the window (#103, PR #177).
- Closing the signature dialog, including with Escape, returns focus to the field instead of the top of the page (#105, PR #181).
- "Resend code" says when the code was not sent, instead of confirming an email during the wait between codes (#113, PR #172).
- A signer name that is a lone emoji, or uses scripts with combining marks, gives whole-character avatar initials instead of a broken one that could stop the page from loading (#115, PR #173).
Sending
- Going Back from Add signers returns to the same draft, instead of a blank upload where a second upload made a second draft (#94, PR #166).
- A field placed just before pressing Back on Place fields is saved (#102, PR #165).
- On a phone, the Field tools sheet stays on screen with Done visible, and Escape closes it (#111, PR #180).
- A long unbroken title in the review and send dialogs wraps, so "Send now" stays on screen (#104, PR #168).
- On wide screens the envelope page's action buttons wrap instead of covering a long title (#131, PR #132 by @seanGSISG, carried in PR #170).
- A new sender on an install without email is told so at sign-up, instead of being told to check their inbox (#95, PR #167).
- Uploading a cut-off or damaged logo is refused, instead of replacing the working logo with a broken one (#91, PR #124 by @HimanshuMahto, carried in PR #171).
Audit trail and certificates
- Password resets, role changes and account closure requests record the IP address and user agent in the site audit trail (#98, PR #164).
- The daily job's scheduled account deletion is logged as the system's, not as a user action (#109, PR #176).
- The certificate says a locally captured email was captured, not accepted by a mail provider (#106, PR #175).
Administrators and settings
- The last administrator cannot close their own account (PR #153 by @MattJackson).
- Saving settings can no longer be undone by an older cached copy (PR #152 by @MattJackson).
- A cancelled deletion's email always arrives before the notice of a deletion scheduled again (PR #160 by @RayFernando1337).
- A scheduled deletion whose files could not be removed is reported as a failure (PR #162 by @RayFernando1337).
- Switching between Mine and Everyone's on the dashboard clears the old search (PR #161 by @RayFernando1337).
Hosting
- The functions that draw PDFs ship only the files they need instead of a copy of the whole project, and CI checks this after each build (#108, PR #119 by @UsamaKhan, carried in PR #174).
- An empty
STORAGE_LOCAL_DIRuses the default.data/storagefolder instead of the working directory (PR #151 by @MattJackson). - A backup or restore whose copy fails lets go of its download (PR #150 by @MattJackson).
The seven outside PRs #150-#153 and #160-#162 were merged through PR #179.
Thank you to everyone who helped with this release. @coygeek filed sixteen of these reports, each with the cause and a clear way to reproduce it, which is a huge gift to the project. @MattJackson sent four careful fixes for the edges of backups, storage, settings and account closure. @RayFernando1337 sent three fixes for scheduled deletion and the dashboard. @UsamaKhan made the PDF functions much smaller, @HimanshuMahto made logo uploads safe against damaged files, and @seanGSISG both reported the covered title and sent the fix.
This release changes the database (20261119090000_spent_signing_links.sql, one new column on signers that lets an old link say what happened). On Vercel the production build applies it. On other hosts, run npm run db:migrate as on every deploy.