We have a legacy system that outputs flat files in fixed format on Unix. The target OS for the data is Windows and Unix end of line characters are problematic as you expect from that file format. The data format is not changed between source and target systems (i.e. they both currently accept fixed format).
My question is;
Does BizTalk come with a filter that converts LF to CRLF, this seems like a common problem when interfacing legacy heterogeneous systems?
We are using BizTalk 2009, though an upgrade to 2011 is in progress.
I wouldn't write two identical schemas with different line endings just to satisfy two different architectures.
Instead, I would write a single flat file schema with one of the two line endings. Then, write an encode send pipeline component and change the line endings to the 'other' system as you are transmitting the message. This will be the closest you will come to a LF to CRLF filter in BizTalk.
If your end systems require different message structures, I would then consider creating two schemas.