Some PDFs have CropBox or MediaBox entries stored as a PDFDict
instead of the expected PDFArray, causing pdf-lib to throw during
lookup.
Wrap both box lookups in try-catch and fall back to A4 dimensions
when neither can be parsed
Upgrade @libpdf/core from 0.2.12 to 0.3.3, which includes:
- WebCrypto SHA-256 replacing pure-JS @noble/hashes (10x signing
speedup)
- Iterative collectReachableRefs (fixes stack overflow on large PDFs)
- Iterative Math.max helpers in xref writer (fixes remaining stack
overflow)
Extract duplicated FontLibrary.use() calls from render-certificate,
render-audit-logs, and insert-field-in-pdf-v2 into a shared
ensureFontLibrary() helper with has() guards so fonts are only
registered once per process.
This PR introduces automatic detection and placement of fields and
recipients based on PDF placeholders.
The placeholders have the following structure:
- `{{fieldType,recipientPosition,fieldMeta}}`
- `{{text,r1,required=true,textAlign=right,fontSize=50}}`
When the user uploads a PDF document containing such placeholders, they
get converted automatically to Documenso fields and assigned to
recipients.
Migrate from @documenso/pdf-sign and @cantoo/pdf-lib to @libpdf/core
for all PDF manipulation and signing operations. This includes:
- New signing transports for Google Cloud KMS and local certificates
- Consolidated PDF operations using libpdf API
- Added TSA (timestamp authority) helper for digital signatures
- Removed deprecated flatten and insert utilities
- Updated tests to use new PDF library
Generate certificates and audit logs using Konva instead of browserless.
This should:
- Reduce the changes of generations failing
- Improve sealing speed
When flattening PDF forms, some appearance streams lack the required
/Subtype /Form dictionary entry needed when used as XObjects. This
causes
corruption in Adobe Reader which fails to render these flattened fields.
Per PDF spec, Form XObject streams require:
- /Subtype /Form (required)
- /FormType 1 (optional)
The normalizeAppearanceStream function ensures these entries exist
before
adding appearance streams as XObjects to the page content stream.
Fixes rendering issues where flattened fields don't display in PDF
viewers.
We were previously omitting cmaps meaning that
when signing documents with certain UTF-8 characters or CJK characters
they would appear as outlined squares in the pdf viewer despite the
actual pdf looking as expected with the characters displaying correctly.
This PR is handles the changes required to support envelopes. The new
envelope editor/signing page will be hidden during release.
The core changes here is to migrate the documents and templates model to
a centralized envelopes model.
Even though Documents and Templates are removed, from the user
perspective they will still exist as we remap envelopes to documents and
templates.
During our field rework that makes fields appear
more accurately between signing and the completed pdf we swapped to
using text fields. Unfortunately as part of that we dropped using the
Noto font for the text field causing ANSI encoding issues when
encountering certain characters.
This change restores the font and handles a nasty issue we had with our
form flattening reverting our selected font.
Improves the existing document rejection process by actually marking a
document as completed cancelling further actions.
## Related Issue
N/A
## Changes Made
- Added a new rejection status for documents
- Updated a million areas that check for document completion
- Updated email sending, so rejection is confirmed for the rejecting
recipient while other recipients are notified that the document is now
cancelled.
## Testing Performed
- Ran the testing suite to ensure there are no regressions.
- Performed manual testing of current core flows.
Adds the ability to align text to the left, center or right for relevant
fields.
Previously text was always centered which can be less desirable.
See attached debug document which has left, center and right text
alignments set for fields.
<img width="614" alt="image"
src="https://github.com/user-attachments/assets/361a030e-813d-458b-9c7a-ff4c9fa5e33c"
/>
N/A
- Added text align option
- Update the insert in pdf method to support different alignments
- Added a debug mode to field insertion
- Ran manual tests using the debug mode
## Description
<!--- Describe the changes introduced by this pull request. -->
<!--- Explain what problem it solves or what feature/fix it adds. -->
## Related Issue
<!--- If this pull request is related to a specific issue, reference it
here using #issue_number. -->
<!--- For example, "Fixes #123" or "Addresses #456". -->
## Changes Made
<!--- Provide a summary of the changes made in this pull request. -->
<!--- Include any relevant technical details or architecture changes.
-->
- Change 1
- Change 2
- ...
## Testing Performed
<!--- Describe the testing that you have performed to validate these
changes. -->
<!--- Include information about test cases, testing environments, and
results. -->
- Tested feature X in scenario Y.
- Ran unit tests for component Z.
- Tested on browsers A, B, and C.
- ...
## Checklist
<!--- Please check the boxes that apply to this pull request. -->
<!--- You can add or remove items as needed. -->
- [ ] I have tested these changes locally and they work as expected.
- [ ] I have added/updated tests that prove the effectiveness of these
changes.
- [ ] I have updated the documentation to reflect these changes, if
applicable.
- [ ] I have followed the project's coding style guidelines.
- [ ] I have addressed the code review feedback from the previous
submission, if applicable.
## Additional Notes
<!--- Provide any additional context or notes for the reviewers. -->
<!--- This might include details about design decisions, potential
concerns, or anything else relevant. -->
Since we allow checkboxes and radio fields without a label (which we use for the value) we
had an issue where multiple checkboxes with no value would exist and items would not end
up checked on the resulting document.
This change fixes that by adding a placeholder value for these empty checkboxes and labels.
Use Noto Sans to gracefully handle inserting custom text
on PDF's. Previously we were using Helvetica which is a
standard PDF font but that would fail for any character
that couldn't be encoded in WinANSI.
Noto Sans was chosen as it has support for a large number
of languages and glyphs with challenges now being adding
support for CJK glyphs.
Previously we used the form flattening method from PDF-Lib
but unfortunately when it encountered orphaned form items
or other PDF oddities it would throw an error.
Because of this certain documents would fail to seal and
be stuck in a pending state with no recourse available.
This change rewrites the form flattening handler to be
more lenient when coming across the unknown opting to skip
items it can't handle rather than abort.
David Nguyen
Ephraim Duncan
Lucas Smith
Catalin Pit
Ted Liang
Catalin Pit
Catalin Pit
Ephraim Atta-Duncan