Unicode Exceptions
Most of the EFT interface supports Unicode characters. (Refer to Unicode File Transfers for details. The table below lists exceptions to Unicode support in EFT.
Area |
Storage (Internal Representation on Disk) |
GUI (Allowed and Displayed) |
Usage |
Remarks |
General |
||||
All domain (host) fields |
Unicode |
Unicode |
7-bit ASCII |
IDN (Internationalization of Domain Names) support by converting Unicode to Punycode upon use. From a user (presentation) perspective and EFT’s internal representation, it will be Unicode. |
All e-mail address fields |
Unicode |
7-bit ASCII |
7-bit ASCII |
Conversion to Unicode for storage, but downgrade on usage. No risk/potential for loss of fidelity, as all chars limited to <128 ASCII. You can paste in Punycode ASCII characters directly for the domain portion if you must have Unicode domains for e-mail addresses. |
SMTP settings (username + password) |
Unicode |
7-bit ASCII |
7-bit ASCII |
Conversion to Unicode for storage, but downgrade on usage. No risk/potential for loss of fidelity, as all chars limited in GUI/usage to <128 ASCII. |
Installer |
8-bit ASCII |
8-bit ASCII |
8-bit ASCII |
See Installer – ASCII ONLY below for potential problems |
Keys and OpenPGP |
||||
SSL CN field |
8-bit ASCII |
8-bit ASCII |
8-bit ASCII |
RFC allows for Unicode but OpenSSL handles the value as ASCII |
SSL priv. key pass |
8-bit ASCII |
8-bit ASCII |
8-bit ASCII |
Passwords stored as octets sequence using same encoding as UI (ASCII) |
SSH priv. key pass |
8-bit ASCII |
8-bit ASCII |
8-bit ASCII |
Passwords stored as octets sequence using same encoding as UI (ASCII) |
OpenPGP priv. key pass |
8-bit ASCII |
8-bit ASCII |
8-bit ASCII |
Veridis limitation. Passwords stored as octets sequence, 8-bit ASCII. |
OpenPGP filename and pathnames |
n/a |
n/a |
8-bit ASCII |
Unicode filenames will be supported, but the will be temporarily converted to ASCII. |
OpenPGP key name |
8-bit ASCII |
8-bit ASCII |
8-bit ASCII |
OpenPGP module does not support Unicode |
OpenPGP public and private key ring paths |
8-bit ASCII |
8-bit ASCII |
8-bit ASCII |
OpenPGP module does not support Unicode |
RSA |
||||
RSA Conf.reg path |
8-bit ASCII |
8-bit ASCII |
8-bit ASCII |
RSA dll only takes ASCII path values |
RSA usernames and passwords |
Unicode |
Unicode |
8-bit ASCII |
RSA does not support Unicode. We will downgrade to ASCII on usage. Potential loss of fidelity resulting in failed authentication attempts. |
ARM |
||||
ARM report content |
Unicode (audited) |
8-bit ASCII (reported) |
8-bit ASCII (reported) |
Loss of fidelity for UTF-8 chars that don’t match local code page for ext. ASCII. The VSReport Designer does not support Unicode. |
AS2 |
||||
AS2 outbound |
n/a |
n/a |
8-bit ASCII |
AS2 does not support Unicode encoded filenames. We can’t downgrade to ASCII as we would be violating Drummond, thus we will simply disallow and log error. |
AS2 inbound |
n/a |
n/a |
8-bit ASCII |
AS2 does not support Unicode encoded filenames. Unlike offloads, EFT inbound can’t detect whether the incoming file is Unicode encoded or not, thus we will always hand the file off to the AS2 component, with potential for mixed results. The outcome will be: a) an ASCII encoded filename, b) a failed transaction, or c) an ASCII encoded unique filename. Reference the AS2 Inbound Operations Use Case Cheat Sheet for additional guidance. |
RADIUS |
||||
RADIUS NAS ID |
Unicode |
8-bit ASCII |
8-bit ASCII |
EFT 6.4 used 8-bit ASCII. EFT 6.5 will represent as UNICODE strings (internally) and then downgrade to 8-bit ASCII on use. Also limited to ASCII in UI. |
RADIUS special cases |
||||
RADIUS shared secret |
Unicode |
Unicode |
UTF-8 |
RFC says nothing about RADIUS password. EFT 6.4 used 8-bit ASCII. EFT 6.5 uses UTF-8. This difference (between earlier versions of EFT and version 6.5 and the resulting potential loss of fidelity is why this item is included on this list (even though it is UTF-8) |
RADIUS usernames |
Unicode |
Unicode |
UTF-8 |
RADIUS usernames and the shared secret can be UTF-8 strings. |
HTTP |
||||
EFT client action HTTP/S credentials |
Unicode |
n/a |
Base64 encoding of UTF-8 string |
|
EFT client action HTTP/S Proxy |
Unicode |
n/a |
Base64 encoding of UTF-8 string |
AS2 client compliance with RFC 2184 |
File encoding of source file on disk |
Resulting change to file encoding in transit |
EFT treatment (resulting encoding and/or loss of fidelity) |
Standard AS2 client. Does not comply with RFC 2184. It relies on filename=<text encoded in ASCII> NOTE: Majority of known AS2 clients |
ASCII |
ASCII (no change) |
Filename integrity maintained. No loss of encoding fidelity. Normal use case. This is essentially 7 and 8-bit ASCII transfers when working with standard ASCII only AS2 clients works perfect. |
Unicode |
ASCII – down converts to "??????" or nonsense characters. |
/n component will fail to process because "????" is an invalid filename. The transfer will result in a failure to write to disk. EFT must log this as an error in eft.log. |
|
AS2 client is compliant with RFC2184. That is, uses filename*= utf-8''<text encoded in UTF-8> |
ASCII |
ASCII (no change) |
Filename integrity is NOT maintained. /n component will process but will convert the filename to a unique 8-bit ASCII filename. However, there won’t be a loss of encoding fidelity. EFT will log as warning that filename was changed. |
Unicode |
Unicode (no change) |
Filename integrity is NOT maintained. /n component will process but will convert the filename to a unique 8-bit ASCII filename; There is also a loss of encoding fidelity as it is down converted to ASCII. EFT will log as warning that filename was changed. |
|
AS2 Other/Miscellaneous Limitations |
|||
AS2 ID (identifier) |
ASCII only |
ASCII only |
EFT displays a message indicating that Unicode is not allowed. |
Q:Why is there no AS2 outbound cheat sheet?
A: EFT controls outbound and disallows UTF-8 encoded filenames from being transferred (see main exceptions above). EFT is more lenient on the inbound side, depending on the use cases described above.
PGP SDA Exceptions
If source filename is: |
If source path is: |
Resulting SDA name: |
If extract path is: |
Result |
Unicode |
n/a |
n/a |
n/a |
Will fail to generate SDA |
ASCII |
Unicode |
ASCII |
ASCII |
Works perfect. Everything preserved |
ASCII |
||||
ASCII |
Unicode |
ASCII |
Unicode |
Will fail to extract |
ASCII |
EFT’s installer is not Unicode compliant. You cannot define Unicode values in the installer for app data path, admin name, etc.The values can be changed in EFT, but with serious ramifications.