Skip to content

Handle non-UTF-8 data from gpgv signature checking

Stuart Prescott requested to merge stuart/python-debian:tmp/gpg-encoding into master

The output from gpgv can contain bytes that are not UTF-8. While it is rare/not seen that keys and signatures from gnupg do so, we are seeing that signatures/keys from sequoia have a NOTATION_DATA field that is binary. When that data is read and decoded to UTF-8, exceptions are thrown. It is possible to decode the output from gpgv one field at a time, or it is possible to allow the decoder to replace the non-UTF-8 sequences with escape sequences instead. The latter is the approach taken here.

With thanks to Gianfranco Costamagna for the suggested fix.

Closes: #1095835, #1087991

Merge request reports