I added linux dbgsym and got better names in my flame graphs. However, I did not get more depth in my stack traces. Most of the graph has depth 4 or 5.

add tips for setting up perf properly so that flame graphs are improved

Adding an input buffer did not help. BufRead already has an input buffer.

Fixed performance issue due to unbuffered output stream.

Performance analysis shows write is the issue. Try wrapping output in a buffer.

Merge pull request #1 from josephmate/add_unit_test_to_figure_out_append_bug

Added unit tests that allowed me to fix the bug.

part way through adding unit test and escaping characters

Performance tests reveal bug with --append and --append is much slower than match. We are no where near close to sed.

Completed untested implementation of --prepend --append, and --replace.

Refactored to allow unit testing on --append, --prepend, --and replace when they are implemented

Refactoring to enable append and replace complete with tests passing.

Refactored to enable split and replace support. However, tests are failing. Needs debugging.

Refactor duplicated code so that adding split and replace will be easier in future patches.

My plan to use the tuple did not work. I need change the iterator interface to return String or char instead of a tuple.