Hi,
1. The RetroReplay clockport activation
IS part of the Contiki driver for the RR-Net. However the MMC64-specific stuff
IS NOT.
2. Contiki 2.x for 6502 machines uses a binary configuration file called contiki.cfg. I've created an online generator for this file:
http://www.a2retrosystems.com/contiki.htm.
3. The general desktop stuff in the Contiki 2.x codebase was broken because the Contiki project in general moves on to different areas. I repaired it to make sure that Contiki 2.x desktop stays a valid option. The test and demo environment is the Win32 console environment. The same is true for the dynamic loader using *.PRG and *.DSC files: Broken -> Repaired -> Win32 console as test environment.
So why isn't there a Contiki 2.x C64 desktop? Simply because I don't see a point in doing so. The Contiki desktop apps didn't receive any improvement since 1.x so it would be a significant amout of work yielding something already available.
In fact it would be hard to even archive the Contiki 1.x state as there are quite some C64 specific hacks written in assembler in the 1.x codebase: Graphics mode desktop, file and directory I/O, mouse support, screen savers, ... All this stuff isn't present in the 2.x codebase anymore and would need to be ported - just to get to the same result as with 1.x
However I thought that there could be something in between the 1.x desktop and the current "non-UI" server-type 2.x apps: Standalone UI client-type apps !
Why that? The C64 doesn't have enough memory to load more than one significant UI app from the desktop. So it might after all make more sense to have standalone UI apps loaded directly from the BASIC prompt. That allows to omit all the dynamic loading code, the window management code and the menu handling code. So I created a Contiki 2.x build option to have a single UI app running in "fullscreen mode". This is a single statically linked binary containing both the "lightweight fullscreen desktop" and the application itself. Again I tested this initially in the Win32 console environment.
Then I ported this approach to the C64 (and Apple2). In contrast to 1.x this time without special assembler hacks but rather making consequent use of cc65 assets: The UI runs in text mode completely relying on the cc65 conio library.
Obviously the by far most interesting UI app is the web browser. As a standalone text mode fullscreen app it uses much less memory than inside the 1.x desktop so I was able to activate the mouse support existing both in the 1.x and 2.x codebase. And here again no assembler hacks but usage of the stock cc65 loadable mouse drivers!
Sounds great so far - doesn't it? The Contiki web browser as a simple app with mouse support running both on the C64 and the Apple2 with only a handful of machine specific code

The issue is that there's a nasty bug that I'm just not able to find:
- Contiki 1.x C64 web browser: fine
- Contiki 2.x C64 web server: fine
- Contiki 2.x Win32 web browser: fine
- Contiki 2.x C64 web browser: lots of lost ethernat packets

If someone is willing to tackle with that it would be just great. I'd certainly be able to assist in using the Contiki 2.x build system to generate the standalone web browser app.
Best, Oliver