Damien George

> However, the output of `co_lines` as currently implemented in [19bc295](https://github.com/micropython/micropython/pull/16989/commits/19bc295e010e4b3151ab852caa28bff6f8fe8a11), produces (a) extra 'spurious' entries corresponding to the `bc+0` info bytes sometimes used to skip the line number ahead...

> but wtf is going on with rp2 that makes this change +8 there? I think this is due to the refactoring of `mp_bytecode_decode_lineinfo()`. Maybe because rp2 is ARMv6 rather...

> Sounds good! I've tried a few other ways of refactoring it [...] If we wanted to be very strict with code size increase, could make it a `#define` instead...

This looks really good now. I think the last commit "py/objcode: Skip spurious and split entries in co_lines." should be squashed down into the second commit "py/objcode: Implement co_lines method."....

> My thought is to do that as a separate PR. The change itself doesn't seem complicated, but the testing feels like it would take at least a bit of...

Thanks for the contribution, this does look very good! One initial comment/question: it looks like `ULP.load_binary()` is gone, and now `ULP.run()` both loads and starts the ULP program. This is...

> What is the best approach for this now? The work you've done here looks really good, thank you. I think it's probably 90% ready to merge. It's just the...

Please rebase this PR now that #16630 is merged.

> I have compiled this bare-bones app to isolate the bug from MicroPython. Is there an upstream issue report in IDF that was for this bug, or at least could...

MicroPython now targets IDF 5.4.2, since 10ef3e4ac2767b68f746f0ebb995672076eee67d