Re: [PATCH v1] [gdb/python]: Add StepEndedEvent (simplifies DAP)

[Tom Tromey <tromey@adacore.com>]

>>>>> "Simon" == Simon Farre <simon.farre.cx@gmail.com> writes:

Simon> Adds the StepEndedEvent which signals that one of the thread
Simon> finite state machines finished. Matches the behavior of what is
Simon> generated by MI; the "end stepping range".

Simon> This should simplify some of the DAP code, where "expected stop
Simon> reason" is being tracked.  This logic should be handled by the
Simon> Python interpreter to begin with, instead.

If possible it would be better to go all the way and add the stop reason
to the stop event all the time.  This has been wanted for a long time,
see

https://sourceware.org/bugzilla/show_bug.cgi?id=13587

Simon> +#include "mi/mi-common.h"

I think this isn't needed in the current patch, but I guess would be
needed if this called async_reason_lookup.

Simon> +static gdbpy_ref<> maybe_create_step_ended () 
Simon> +{
Simon> +  const auto tp = inferior_thread();
Simon> +  const auto fsm = tp != nullptr ? tp->thread_fsm() : nullptr;
Simon> +  if (fsm != nullptr && fsm->finished_p () && 
Simon> +	fsm->async_reply_reason () == EXEC_ASYNC_END_STEPPING_RANGE)

The stop reason idea is basically -- why single out one kind of stop
reason when we could do something more generic.

We wouldn't need a subclass, I guess, we could just stick the extra info
into the current event.

Not sure though.  This area is kind of messy due to some unfortunate
choices early on.

Tom