Update to MicroPython v1.25 #313
Conversation
dlech
force-pushed
the
micropython-v1.25
branch
4 times, most recently
from
da2c161 to
9fb2e5d
Compare
|
Awesome!
Let me know if you'd like to continue here or would like me to join in. |
|
If you have some ideas already about how you would like to handle imports differently, please go ahead. Otherwise, I will probably get back to it on Friday. |
|
I haven't looked at how it works yet. Maybe more of it is included in MicroPython now. Or maybe we've freed up enough space to do it the normal way 😄 |
|
Well, we didn't free up any space yet. But the virtualhub has been updated to do it the new "normal" way. |
Replace `STATIC`` with `static`. The define is being removed in MicroPython and will cause a compile error in the future.
Upstream MicroPython switched from black to ruff, so we are following suit. This now includes all non-3rd-party .py files so quite a few files are reformatted that weren't being checked before.
Update to current MicroPython stable release v1.25. Includes changes required to compile without errors: - Handling of PYEXEC_FORCED_EXIT/SystemExit changed - mp_make_function_from_raw_code() replaced by mp_make_function_from_proto_fun() - Module import code changed - mp_lexer_new_from_file() function signature changed - mp_hal_stdout_tx_strn() function signature changed - MP_REGISTER_MODULE_DELEGATION() replaces MP_MODULE_ATTR_DELEGATION_ENTRY() - mp_obj_malloc_var() signature changed - mp_generic_unary_op was removed - m_new_obj_var_with_finaliser() replaced by mp_obj_malloc_var_with_finaliser() - Handling of 'u' prefix on builtin modules has changed
|
The stop button and imports are passing the smoke test, so I'll go ahead an merge this and I'm sure people will let us know if anything unexpected broke. |
3 participants
We've missed a few MicroPython releases, so it is about time we updated.
This passed locally with
make allso we'll see what else CI catches. Also, I haven't actually tried running this on a hub yet, so we'll need to do some manual testing to see if anything else broke.Some key things that changed to be tested are raising SystemExit/pressing stop button and module imports (both builtin and user),
Once we've fixed any regressions, I would like to set
MICROPY_QSTR_BYTES_IN_HASH=0on hubs where we are starting to run out of flash memory to regain some space.