[refactoring] BSD-License-compliant flashloader rewrite

[chenguokai]

This will finally close #915

[chenguokai]

Hope this rebase will help

[chenguokai]

My local test shows that the current code is known to be working on stm32f103 and stm32f401. I am dealing with some minor formatting issues and a critical improvement about the compile process. After that the code will be ready for test.

Edit: All done.

[Nightwalker-87]

I think we should have more people looking at this as it is a key component.
What exactly do you mean with "cleanroom" - why did you choose this naming?
I don't think that explanations should go to /doc/developer.txt as this is only a quick and dirty copy-paste-doc created "once upon a time". I actually feel we should have a separate document /doc/flashloaders.md for it. Explanations which were present as comments in the old flashloader files could go into this document, maybe in form of a parameter | explanation table format - something like this. Maybe someone even has a better idea for that. In the end users should find a place to gain knowledge on how this works.

[chenguokai]

What exactly do you mean with "cleanroom" - why did you choose this naming?

Because it is what kind of method this PR have followed.

From https://opensource.stackexchange.com/questions/6216/can-i-cleanroom-code-by-myself-if-public-specifications-already-exist

A clean room design is a preventative defence strategy in anticipation of copyright infringement claims. The idea is that a new creative work cannot infringe on the copyright of an earlier work if the new work's authors did not view the earlier work. A clean room design process implements this isolation by having different people handle reverse-engineering of the earlier work and creating the new work, with the two groups only communicating via a spec that can be more easily reviewed for non-infringement.

I don't think that explanations should go to /doc/developer.txt as this is only a quick and dirty copy-paste-doc created "once upon a time". I actually feel we should have a separate document /doc/flashloaders.md for it. Explanations which were present as comments in the old flashloader files could go into this document, maybe in form of a parameter | explanation table format - something like this.

That's what I wanted, too. This clean room document provide only essential information for another one to implement a BSD code. What I wanted to add to any new document (either /doc/flashloaders.md or some file with another proper name) will and should include a more detailed explanation of why we should follow those restrictions.

[Nightwalker-87]

Ah I understand. To be honest I have never come across this topic before, thus I could not associate the name. Ok, so this reads as being a somehow intermediate step towards a complete solution of this issue. I'm fine with that. Thx for explaining. 👍
It would be desirable though to merge a complete solution if possible.

[chenguokai]

Hope the latest commit will help

[Nightwalker-87]

I have the impression that someone seems to have deeply focussed on this topic. 👀
If this whole stuff works (what I can't verify by myself here) we can go with it, apart from some minor changes.

[chenguokai]

All "my"s have been replaced.
Added a note at the head of flash_loader.c

[Nightwalker-87]

I'm tired of waiting any longer for a second review. 👎
If nobody cares, I don't do either...

[chenguokai]

Hope that untested devices work well🤔

[Nightwalker-87]

It's ok and if so, to me it's not in your responsibility.