Refactoring

[puzrin]

In continue of https://github.com/stg/SYLT-FFT/issues/6#issuecomment-949832806 and below

See proposal draft for v2.0.0: https://github.com/speedcontrols/dc_sc_grinder/tree/dev/lib/SYLT-FFT

Goal: simplify use of SYLT-FFT as library.

Renamed files/folders

1. More descriptive.
2. Auto-compatibility with platformio, arduino etc (those build systems recognize such folders well). No need to add env-specific library declarations.
3. More extendable (can add test/ folder later)

May be include/sylt-fft/ instead of include? See etl for example.

Added lib namespace prefixes (incomplete)

Example: https://github.com/speedcontrols/dc_sc_grinder/blob/dev/lib/SYLT-FFT/include/config.h#L24

1. I did minimal changes, for example only. More prefixing required. Did not touched math, according to your notes.
2. Not sure about SYLT_FFT_ (too long). May be SYLT_? What "sylt" means?

Redeclared sine table

https://github.com/speedcontrols/dc_sc_grinder/blob/dev/lib/SYLT-FFT/include/config.h#L46-L82

1. More flexible. Now can add more tables of different length
2. Allow external const table

Added #ifdef guards (incomplete)

https://github.com/speedcontrols/dc_sc_grinder/blob/dev/lib/SYLT-FFT/include/config.h#L48

That allows to configure everything via -D, no need to edit config manually (+ need prefixes to avoid collisions).

Suppressed arch warnings

https://github.com/speedcontrols/dc_sc_grinder/blob/dev/lib/SYLT-FFT/include/intrinsics.h#L12

Hided warnings behind debug guard. If user use this lib with < M4 - no need to annoy on every build.

---

Notes:

• Please, add tag 1.0.0 to last commit. That will help users to access "previous" version, and creating CHANGELOG.md.
• First refactoring iteration require 2 things:
  • reorganize folders/files
  • decide final name for "library namespace prefixes". SYLT_FFT_? SYLT_? Other?

What do you think? Can you propagate acceptable changes to upstream?

PS. I'm ~ software architect, but NOT C programmer. So, i'd prefer to not intrude to your code via PRs, without experience of C specifics. But i can prepare samples/descriptions of proposals and other things (except coding).

[puzrin]

One more note about namespaceing. I remember, you wish to have simple math functions accessible.

My proposal is (two parts):

1. Namespace everything
2. Create fft_legacy.h file with something like this:

   #define old_name1 namespaced_name1
   #define old_name2 namespaced_name2
   ...
   

Then:

• library will be collision-free by default
• existing users can continue use "old-style" api, enable it with single line #include "fft_legacy.h"

=> all should be happy

[stg]

Sure thing!

[puzrin]

I think we have agreement about principal questions. Can i help somehow more to land changes into master?

[stg]

Dear puzrin,

I am not able to keep track of activity on my git account. Please fork the project and take it over, such that you have full control of it. I will amend the README to reflect this.

[puzrin]

@stg consider create organization, transfer project there and add me as second admin of org. Then i'll be able to curate development until your situation changes back.

If i do fork, that will look ~~like parasitism on your work~~ strange :). I'd like to avoid such glory. I'm really impressed by your efforts in sylt-fft and would like project to be continued in best possible way. IMO, org with 2 admins would be optimal.

Also, if you like to keep existing code as single-file-based, we could create second repo sylt-fft2. But again - in context of organization.

Let me know what do you think, or if i misunderstand something.