ippValidateAttribute: Properly detect strings with an invalid last character

[vliaskov]

An ISO/IEC 8859-1 / Latin-1 string with an invalid last character whose UTF8 equivalent is a multi-byte sequence is currently not detected correctly in ippValidateAttribute, leading to a _dbus_check_is_valid_utf8 crash in relaxed conformance mode. The multi-byte sequence check in ippValidateAttribute expects that the following bytes exist, and this is not the case here.

To catch this case, I think ippValidateAttribute needs to check whether the string finishes midway in a multi-byte sequence check for IPP_TAG_TEXT* and IPP_TAG_NAME* strings.

[vliaskov]

If the approach in
vliaskov/cups@4f0f44f
is ok, I can make a PR, let me know.

[michaelrsweet]

Hmm, I'm not keen on those changes. But I can see how a truncated string or one that sortof looks like UTF-8 could cause this.

[michaelrsweet]

OK, I refactored the code so that truncated UTF-8 sequences will fail validation.

[master d9f301d] Fix UTF-8 validation (Issue #5509)

[branch-2.2 949c217] Fix UTF-8 validation (Issue #5509)