Handling blank lines in LLP Listener

[jeffkeller]

I have an LLP listener which seems to work fine unless an incoming HL7 message has a blank line in it, in which case Mirth appears to split the message, parses the first part as best it can, then errors out on the second part since it isn't a valid HL7 message by itself.

Any thoughts on how to get around this? I've tried "Wait for End of Message Char" and that doesn't seem to help.

If you need a sample HL7 message, let me know, but it looks to me as if any blank line in a given HL7 message will cause this.

[jeffkeller]

Re:Handling blank lines in LLP Listener

To clarify a bit further, by "blank line" I mean back-to-back "0a" characters.

My delimiters are set as follows in the channel:
Start of Message Char: 0x0B
End of Message Char: 0x1C
Record Separator Char: 0x0D
End of Segment Char: 0x0D

[jeffkeller]

Re:Handling blank lines in LLP Listener

This appears to be caused by this line

if ((line.length() == 0) || line.equals((char) endOfMessage) || line.startsWith("MSH")) {

in BatchMessageProcessor.processHL7Messages(). Is this the intended behavior on the part of Mirth? Is there a way to modify this behavior via the interface?

[chrisl]

Re:Handling blank lines in LLP Listener

There is no way to modify this directly, however if you add it to the Jira, we can make the batch processor and option on the LLP side.

Thanks!

[jeffkeller]

Re:Handling blank lines in LLP Listener

Done.

http://www.mirthproject.org/index.ph...pper&Itemid=74

Thanks for checking this out, Chris.

[brendanh]

Re:Handling blank lines in LLP Listener

Wait for end of message character should take care of this if you use it in conjunction with the non-strict LLP validation UNLESS your sending program is putting the end of message character prematurely.

[jeffkeller]

Re:Handling blank lines in LLP Listener

That's what I had hoped.

I've attached a simple channel and 2 messages (blankline.txt and noblankline.txt) that will produce this behavior. Maybe you'll see something in there that I missed...
blanklinetest.gz (1414 bytes)