The design of this parser is similar to my vCard and Internet Mail Message parsers.
I’ve now tested this with an .ldif exported from Netscape Communicator 4.6 (Linux glibc). It looks like folding is interpreted correctly, as is rebuilding of folded lines.
Note that Communicator 4.6 isn’t using the latest draft spec of ldif, which is called: "The LDAP Data Interchange Format (LDIF) - Technical Specification" Filename: draft-good-ldap-ldif-04.txt. This spec was only released on 22nd June 1999, too late for Communicator 4.6. Due to this, the format that is exported is slightly different, though not enough to confuse the parser. The new spec adds a compulsory ‘version’ line. The parser will generate its own version line using version ‘0’ in this case. Using the current draft, the version should be set to 1.
Also note that URL lines aren’t recognised correctly as they are missing the ‘<’ character that identifies them. I guess this is the fault of the old spec and not Communicator.
You can test if a ValueSpec is Base64 encoded by calling: LDIF::ValueSpec::ValueType LDIF::ValueSpec::valueType(); The return value is LDIF::ValueSpec::Base64 if, er, it’s Base64 encoded. Use KCodecs::base64Decode() to get the data.