After discussion with @the-eg and @p3p it makes the most sense to port this to python.. since anyone who is going to use the output of this to compile marlin.. will already have python installed.

Right now the code just assumes PlatformIO is installed and in the default user profile location. If it's not there it will just fail to build.

Change behavior so if it can't find it there, throw an error.

Make an option to define the PlatformIO environment location in JSON settings section.

This option will create the marlin root directory if it does not exist. It will pull down the branch as defined in either the JSON or --branch flag. If neither are specified it will throw an error if using this flag.

I think it is important to inject metadata into the to of the modified .h files with the metadata from the run so there is evidence of not only a tool was used on it, it also doubles as a "how-to" should they want to get that config again.

From discord:

platformio.ini
 build_flags = 
  -DPWM_FREQUENCY=(168*10^6)/(2*255*1)

need to adjust logic of build commands to accommodate special characters and math objects.

when values are set in the json settings but there is nothing passed on the commandline it triggers a conflict error. using --preferargs after that allows it to continue on with broken values and the build fails.

When there is a different marlin root path defined in the json config and the command line arg, the command-line arg wins regardless of --preferargs. It should throw an error forcing use of the --preferargs flag.

parse the configuration file prior to modification to create a map of which directives are in which sections. store them into a new json file with the config version as the name for future reference..

the new python module argparse has its own mechanism for reading in conf settings for the flags. this actually helps a lot since one of the flags is the JSON config file.. created a bit of a catch-22 for me. This way the configuration for flags set by command-line arguments are seperated from the configuration for parsing.

In batch mode everything that would default to throwing an error, that is a logic choice, will default to the behavior that would allow it to progress. An example might be to add a missing directive automatically or enable --skipmissing. Or, there is a config conflict and the default action is to force an option, it might default to--preferargs rather than exit.

In interactive mode everything that causes an error or has a conflict will ask for a choice instead of forcing the command to be rerun with an argument.

similar to the --preferargs flag, add a --preferconf flag that would allow the override in the other direction, using what is in the json config file instead of the command-line.

If moving through the logic we are asked to enable an already enabled, or disable an already disabled directive, display a message and skip that item.

Used the config to json script to convert my existing config into the new format. Added the old format to the end of it. Working on the code now to parse it ... then will start working on the offsets.

currently unable to determine if a global value is a default or a conf (json) value if the json value has the same value as the default. need to define an array of what the defaults are as a constant that never changes so i can compare the values.

The idea is to have a main JSON File, and then other smaller files with a specific purpose.

They will be processed in the order they are created.

An example might be the stock creality cr10 s5 from the repo... with your changes... a month goes by and you want to add a bltouch. So you find a bltouch.json in contrib and include it...

need to write a fully functional library that will allow the two config files to be converted into a json, that can later be used to create new configuration files from.