py/dynruntime: add mp get buffer

[dpgeorge]

No description provided.

[github-actions]

Code size report:

   bare-arm:    +0 +0.000% 
minimal x86:   -24 -0.013% 
   unix x64:   -24 -0.003% standard
      stm32:    -8 -0.002% PYBV10
     mimxrt:    +0 +0.000% TEENSY40
        rp2:    -8 -0.002% RPI_PICO
       samd:    -8 -0.003% ADAFRUIT_ITSYBITSY_M4_EXPRESS

[codecov]

Codecov Report

Merging #12679 (a1be5e1) into master (516385c) will decrease coverage by 0.01%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master   #12679      +/-   ##
==========================================
- Coverage   98.39%   98.39%   -0.01%     
==========================================
  Files         158      158              
  Lines       20951    20950       -1     
==========================================
- Hits        20614    20613       -1     
  Misses        337      337              

Files	Coverage Δ
py/nativeglue.c	94.50% <ø> (ø)
py/obj.c	98.13% <100.00%> (-0.03%)	⬇️
py/obj.h	100.00% <100.00%> (ø)

[andrewleech]

This looks like a nice optimisation on the get_buffer function, I like the idea of just being able to pass a flag to raise or return.

Am I right in assuming this change would break existing built dynamic C modules? Is that what the updated sub_version is flagging, to raise an error on import of older mpy files that have C code?

[dpgeorge]

Am I right in assuming this change would break existing built dynamic C modules? Is that what the updated sub_version is flagging, to raise an error on import of older mpy files that have C code?

Yes, that's correct.