Tweak readme to avoid bad formatting

Add a bit more info to the readme

Delete some now unused code in compress.py

Remove separate MAX_X_OFFSET/MAX_Y_OFFSET from compress

Code doesn't support different MAX_X_OFFSET and MAX_Y_OFFSET currently
so just have one param MAX_OFFSET. No functional change.

Add description of compression in readme

Update readme

Add description of decompressing raw captured stream
Add info about gbcap verilog testbench

Make gbcap verilator testbench less verbose

Just print a line per input frame, not per input line

Add frame count to gbcap verilator testbench output

Make verilator testbench VCD dumping easily switchable

./run gbcap verilator --trace

Will build the testbench with VCD dumping enabled.

Simplify verilator testbenches

gbcap testbench is pretty much the same as the compress testbench just
with FIFO+UART output, so drop the compress tb - removes a lot of
copy-paste of code.

Testbench now reads and writes binary rather than hex encoded strings,
simplified reading of lines.

dump_pgm: update argparse help to be accurate

Had copy-pasted this without correction...

recv: only support outputting raw data from serial port

Remove decompression, use decomp_stream

decomp_stream improvements

- Add support for raw and hex encoded input
- Add support for outputting animated gifs using imageio
- Make non-verbose by default
- Support reading from a file or stdin

Add explaination for local version of migen

Update migen

Fix a couple of off-by-one errors in the OUTPUT_DONE state

Would "leak" some lines on each frame and so eventually overflow the
buffer.

Handle input line being written whilst output is stalled

Could be in OUTPUT_DONE state but waiting for the last output byte to be
consumed. In that case, still want to do the increments of full_lines
and todo_lines.

Detect overflow in gbcap verilator testbench

Testbench for gbcap lines->compress->uart

Quieten verilator a little

All the verilog is generated by migen...