- when doing batch processing we are not using multi core.
    What can we do about it?

- error handling.
    This is a big one.
    I want every processor to know how to parse the output of the tools it is running (if they are exernal) and collect errors in a organized fashion: file, line number, desription of the error, stack etc.
    then we can do something like: rsconstruct errors edit
    and it will launch an editors of my choice on all the previous errors.
    Internal processros will surely be able to collect all the errors into that predefined structure with no problems.
    rsconstruct errors show  will show the latest errors

- doing checksum of files.
    could we store the checksum next to the file on disk as extra data?
    are we using mmap when doing the checksum? Maybe we could speed up the checksum calculation this way.

- I want to add a variant feature to every process.
    - what is the difference between two variants and the same processor twice in the same file?
    - the cargo processors seems to have variants, what does that mean?

- I assume we have a centalized place where we draw tables. If we do not then arrange a place like that (src/tables.rs). In that place always put a horizontal line
    after the headers of the table
    in "rsconstruct status" I do not see a horizontal line after the headers. 

- want a config system like the kernels config system.
    This is a big item.

- call the errors EXIT_[XXX]. It wail make it clearer.

- There is knowledge that we know when we run the tera processor (for instance, how many files are under the charge of each prcoessor) - we can make that knowledge available to the processor so we woulnd't have to do things like counting files.
