Add support for ignoring leading/trailing junk when parsing messages

[da4089]

It's very common for FIX log files to be written with one message per line, and a direction (send/recv) and timestamp prior to the message.

It'd be nice if FixParser were able to parse such files "out of the box" without having to trim the lines manually first.

Add an optional FixParser behaviour to enable it to ignore spurious characters before the 8=FIX and after the 10=xxx|, such that it can read a FIX log file into a sequence of messages with a single call to append_buffer().

[Pk13055]

Hi, what's the current status of this feature? This would be awesome to have ASAP, as it would help avoid most of the manual checks and truncation.

[da4089]

Hi Pratik -- can you post a sample of the text you'd like to extract the message from? It'd be helpful to have some test cases for this functionality.