Damien George
Damien George
Well, at least this means we can go ahead with a straight forward implementation of MRO, and get these details correct at the very end, if at all.
It requires a little bit of work to support derived types on rhs of assignment. Work around: use `b[0:4]` or `memoryview(b)` on RHS.
Using a memoryview (pre created perhaps) should not impact performance at all (or very little).
It's possible to improve the situation slightly: the compiler currently uses the head element of the dict/set to report the line number for the error, but it can in fact...
Confirmed. This looks similar to issues with the RTC not starting ofter DFU programming, and requiring an rtc.init() call to get it working. I guess programming leaves some registers in...
> In this sense, the behavior is correct: someone calling timeout_sock.read(100) wanted one of: a) exactly 100 bytes; b) less bytes in case of EOF; c) error (exception), in particular...
> There's nowhere to push data back. You can cache it in the socket object. But of course that is not a very nice way to solve the problem. >...
Thanks @hoihu! I can confirm that the above script leads to a hard crash, and it only takes 10-50 tries.
I'm looking at it now. It is related to double ctrl-c, and is hence related to PENDSV ISR. I've disable the PENDSV stuff and it works fine. Using the hard-fault...
I can't make progress on this, but the problem goes away with the following hack: ``` diff --- a/stmhal/pendsv.c +++ b/stmhal/pendsv.c @@ -56,9 +56,11 @@ void pendsv_nlr_jump(void *o) { if...