Code size report:

Reference:  tools/mpy_ld.py: Allow overriding the internal MPY file name. [b49f098]
Comparison: rp2: Enable PPP for Pico W. [merge of 533a154]
  mpy-cross:    +0 +0.000% 
   bare-arm:    +0 +0.000% 
minimal x86:    +0 +0.000% 
   unix x64:    +0 +0.000% standard
      stm32:    +0 +0.000% PYBV10
      esp32:    +0 +0.000% ESP32_GENERIC
     mimxrt:    +0 +0.000% TEENSY40
        rp2: +24632 +2.669% RPI_PICO_W[incl +752(bss)]
       samd:    +0 +0.000% ADAFRUIT_ITSYBITSY_M4_EXPRESS
  qemu rv32:    +0 +0.000% VIRT_RV32

Update to enable in mpconfigport.h by default, with the option to set MICROPY_PY_NETWORK_PPP_LWIP and disable it.

Looks like there's a significant code size overhead, though!

Edit: Looks like I need to limit to ports with LWIP enabled 🤦

[  2%] Generating genhdr/pins.h, pins_ARDUINO_NANO_RP2040_CONNECT.c
[  2%] Generating genhdr/qstr.i.last
/home/runner/work/micropython/micropython/micropython repo/extmod/network_ppp_lwip.c:38:10: fatal error: lwip/dns.h: No such file or directory
   38 | #include "lwip/dns.h"
      |          ^~~~~~~~~~~~

Actually or maybe I have an order of operations thing going on in the header.

Edit: I think this is the classic disagreement in patterns between:

1. Include all source files and use .h defines to determine whether their contents are enabled
2. Conditionally include things based on CMake options and then set a define, eg:

   micropython/ports/rp2/CMakeLists.txt

   Lines 318 to 327 in 744270a

   	if (MICROPY_PY_LWIP)
   	target_link_libraries(${MICROPY_TARGET} micropy_lib_lwip)
   	target_include_directories(${MICROPY_TARGET} PRIVATE
   	lwip_inc
   	)
   	target_compile_definitions(${MICROPY_TARGET} PRIVATE
   	MICROPY_PY_LWIP=1
   	)
   	endif()

Wait a minute, what's the diference between:

MICROPY_PY_LWIP_PPP - https://github.com/search?q=repo%3Amicropython%2Fmicropython%20MICROPY_PY_LWIP_PPP&type=code

and

MICROPY_PY_NETWORK_PPP_LWIP - https://github.com/search?q=repo%3Amicropython%2Fmicropython+MICROPY_PY_NETWORK_PPP_LWIP&type=code

The latter seems to be set just for lwipopts_common.h -

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 98.51%. Comparing base (b49f098) to head (533a154).
⚠️ Report is 2 commits behind head on master.

@@           Coverage Diff           @@
##           master   #16445   +/-   ##
=======================================
  Coverage   98.51%   98.51%           
=======================================
  Files         176      176           
  Lines       22904    22904           
=======================================
  Hits        22563    22563           
  Misses        341      341           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

@Gadgetoid this PR looks good now, but you still have it in draft status. Any reason it's in draft?

If you move it out of draft and rebase, it should be good to merge.

@Gadgetoid see above.

Any reason it's in draft?

PR anxiety! And a dash of completely forgetting I raised this.

Rebased and converted.

PR anxiety! And a dash of completely forgetting I raised this.

😄

I think some wires got crossed during the rebase, see comments above.

Okay these commits being conceptually backwards really hurt my brain, I've swapped them and hopefully fixed them too.

thousand yard stare

😭

edit: Okay I had confused this in my head with enable LWIP PPP for all Pico boards, since we have PPP-based modules that could conceivably run on a Pico, Pico 2 and CYW43 support is tangental to that. As such I changed the mpconfigport.h from a conditional enable to always enable 🤦 Fixed!

Of course this built locally for me because I didn't try a non-W build.

Th... fo.. fifth time lucky?