Add language support for RP2040 PIO ASM

[willeccles]

Adds support for the PIO assembler language for the RP2040, including embedded C code. I have left the "detected" language from the embedded code snippets generic and hard-coded C for now, as it's possible they could add support for other languages (i.e., C++, Python) in the future. To accomplish this, I have included new patterns and language contexts and so on where necessary.

However, I did not know how to handle the fact that it's an error for a % c-sdk { to be unterminated by a matching %}. At the moment, it simply treats it the same as Markdown--unbalanced is just treated as a valid block. Feedback would be appreciated. Relevant code snippets: here and here.

[willeccles]

Reproduction for the issue: consider a simple PIO program with the following code:

.program something
nop

% c-sdk {
void func(void);
// <---- code block NOT terminated!

The assembler will reject this program, but tokei will treat it the same way it does Markdown code fences: if it's not terminated, the rest of the file is C, and there is only one comment added for the % c-sdk {. This is technically correct in this situation, despite the program being ill-formed.

Now let's slightly tweak this file. If you simply start another PIO program underneath this, the counting is now broken:

.program something
nop

% c-sdk {
void func(void);
// <---- code block NOT terminated!

.program another_program
nop

This is also not accepted by the assembler, and again, tokei considers the rest of the file to be C code. However, in this example, this is not correct--this counts PIO instructions as C code.

I am not sure how this should be handled. Any thoughts? The question is really this: should tokei make an effort to properly handle ill-formed code? Personally, my vote is that this is probably "correct" behavior as-is. Garbage in, garbage out.

[willeccles]

Just realized that Python is supported in passthrough sections as per datasheet section 3.3.10.2. I'll update to include that.