[BUG] Wrong emoji deserialization

[Guzhuu]

Describe the bug
Emojis in an HL7 message are not deserialized correctly when using non-strict parser, which causes an error and prevents normal message processing.

To Reproduce
Steps to reproduce the behavior:

1. Create channel, Source outbound message HL7 non-strict parser
2. Insert the emoji somewhere in the Source outbound message (msg['MSH']['MSH.4']['MSH.4.1'] = "😀")
3. Create 2 destinations, the first one with HL7 inbound, strict parser; the second one, HL7 inbound, non-strict parser. Create a transformer in each destination otherwise the message will not be parsed and no error will occur.
4. Process a message

Expected behavior
Correct emoji deserialization, for example:

• Original -> XML escaped (Source Transformed) -> Original (Source Encoded & Destination Raw) -> XML escaped (Destination Transformed)
• 😀 -> 😀 -> 😀 -> 😀
• 💄 -> 💄 -> 💄 -> 💄

Actual behavior
Incorrect emoji deserialization, for example:

• Original -> XML escaped (Source Transformed) -> Original (Source Encoded & Destination Raw) -> XML escaped (Destination Transformed)
• 😀 -> 😀 -> 😀 -> ��
• 💄 -> 💄 -> 💄 -> ��

Screenshots
Source transformer

Source Transformed

Source encoded

Destination strict parser Raw

Destination strict parser Transformed

Destination non-strict parser Raw

Destination non-strict parser Transformed

Destination non-strict parser Error

Environment (please complete the following information):

• OS: [Linux (RHEL 8.9), Windows 10]
• Java Distribution/Version [OpenJDK 21, Java 8 (201)]
• Connect Version [at least 3.12.0, 4.4.1 and 4.5.2]

Workaround(s)
Using strict parser, and changing the channel accordingly; which is not always an option.

[pacmano1]

Nice write up!

Maybe post a sample message as a file to make this a bit easier to reproduce. And you left out what encoding you are using, e.g. UTF-8?

[Guzhuu]

Hello,

The emoji was being inserted in the HL7 in the source transformer, so any message (e.g. MSH|^~&|MSH3|😀|B|C|D|E|MDM^T02^MDM_T02|||2.5||||||UNICODE UTF-8)

Encoding UTF-8 is being used.