Inline markup: How does org identify nested code/verbatim?

[<c.buhtz@posteo.jp>]

Hi folks,

this is a question about org(mode) development itself.
It is magic to me how you do this. ;) And I would like to learn it
because I do write kind of an org parser in Python.

Here is a nested code-in-verbatim text.

    This =is ~code~ in verbatim= text.

Exporting this to html (via org-html-export-as-html)

    This <code>is ~code~ in verbatim</code> text.

Awsome! :D

The point is myself I'm able to identify code or verbatim with regex
including three catch groups for the content before, between and
after the inline markers.

    for verbatim: "(^|[ .,;:\-?!({\"'])=(.*?)=([ .,;:\-?!)}\"']|$)"
    for code:     "(^|[ .,;:\-?!({\"'])~(.*?)~([ .,;:\-?!)}\"']|$)"
 
But they don't work together. In the example above I need to use the
verbatim regex first to make it right. 

If I would use the code regex first it wouldn't work because it would
find the ~code~ but without knowing that it is surrounded by ~verbatim~.

I don't know what my users inputs to my software: verbatim in code or
code in verbatim. So I have to figure out which regex to use first.

How does org solve this problem? I don't need a full working solution
but just an idea.

One approach in my mind is to run both regex separate and then compare
the results "somehow":

    Verbatim: ['This', ' ', 'is ~code~ in verbatim', ' ', 'text.']
    Code :    ['This =is', ' ', 'code', ' ', 'in verbatim= text.']

"Somehow"!

Another approach in my mind is to do something I would call nested
regex. Constructing a regex pattern looking for verbatim with code in
it. And the other way around of course.