CVE-2022-31116

Published: 5 July 2022

UltraJSON is a fast JSON encoder and decoder written in pure C with bindings for Python 3.7+. Affected versions were found to improperly decode certain characters. JSON strings that contain escaped surrogate characters not part of a proper surrogate pair were decoded incorrectly. Besides corrupting strings, this allowed for potential key confusion and value overwriting in dictionaries. All users parsing JSON from untrusted sources are vulnerable. From version 5.4.0, UltraJSON decodes lone surrogates in the same way as the standard library's `json` module does, preserving them in the parsed output. Users are advised to upgrade. There are no known workarounds for this issue.

Notes

the embedded ujson code in pandas, eventhough containing similar
content as the upstream ujson code, seems to have diverged from the
ujson upstream project (they have fully forked ujson), since pandas
upstream is maintaining their own ujson bug fixes and changes without
re-syncing with the ujson upstream project. There is no indication
the ujson fork, as used in pandas, is vulnerable to the same issues
as the upstream ujson code.

according to https://github.com/ultrajson/ultrajson/releases/tag/5.4.0,
the commits that fix this issue are e0e5db9a46 and bc7bdff051. However,
release 5.4.0 does include a few other commits which seem to be related
to the above.

Status

Severity score breakdown

References

• https://github.com/ultrajson/ultrajson/security/advisories/GHSA-wpqr-jcpx-745r
• https://github.com/ultrajson/ultrajson/commit/67ec07183342589d602e0fcf7bb1ff3e19272687 (5.4.0)
• https://github.com/ultrajson/ultrajson/commit/67ec07183342589d602e0fcf7bb1ff3e19272687
• https://ubuntu.com/security/notices/USN-6629-1
• https://ubuntu.com/security/notices/USN-6629-3
• https://www.cve.org/CVERecord?id=CVE-2022-31116
• NVD
• Launchpad
• Debian