Re: [PATCH 16/26] rv/rvgen: fix unbound initial_state variable

On Tue, 2026-01-20 at 08:42 -0300, Wander Lairson Costa wrote:

On Tue, Jan 20, 2026 at 09:21:01AM +0100, Gabriele Monaco wrote:

quoted

I believe these newlines were added automatically by some tools, not sure if
we
really want them but they don't belong in this patch (since 1/26 added this
class).

There are python formatting tools that put 2 lines after a type
definition. Anyway, I can remove it in v2.

If that's common practice I'm fine keeping it, I just would create this class
directly with the newlines separators in 1/26 instead of adding these lines
later in the same series. Just to reduce the noise.

quoted

quoted

 class Automata:
     """Automata class: Reads a dot file and parses it as an automata.
 

@@ -31,7 +33,9 @@ class Automata:

         self.__dot_path = file_path
         self.name = model_name or self.__get_model_name()
         self.__dot_lines = self.__open_dot()
-        self.states, self.initial_state, self.final_states =
self.__get_state_variables()
+        self.states, self.initial_state, self.final_states = (
+            self.__get_state_variables()
+        )

There is no strict 80 character limit for python code and I personally find
this
less readable. Is this again what the tool suggested?

In general, 100 lines is assumed the to a good limit. Anyway, this is
minor and I can remove the change in v2 if desired.

Yeah 100 is probably also what checkpatch.pl uses (although I think it skips
python files), since this is below 90 columns, I wouldn't touch it.

Thanks,
Gabriele

quoted

quoted

         self.events = self.__get_event_variables()
         self.function = self.__create_matrix()
         self.events_start, self.events_start_run =
self.__store_init_events()

@@ -86,6 +90,7 @@ class Automata:

         # wait for node declaration
         states = []
         final_states = []
+        initial_state = None
 
         has_final_states = False
         cursor = self.__get_cursor_begin_states()

@@ -96,9 +101,9 @@ class Automata:

             raw_state = line[-1]
 
             #  "enabled_fired"}; -> enabled_fired
-            state = raw_state.replace('"', '').replace('};',
'').replace(',',
'_')
+            state = raw_state.replace('"', "").replace("};",
"").replace(",",
"_")

Ok this change is good.

quoted

             if state.startswith(self.init_marker):
-                initial_state = state[len(self.init_marker):]
+                initial_state = state[len(self.init_marker) :]

You fixed spacing in 12/26. We could either keep move this change there or
just
merge that patch with others touching the same files.

Np, I can move this change to there.

quoted

quoted

             else:
                 states.append(state)
                 if "doublecircle" in self.__dot_lines[cursor]:

@@ -111,6 +116,9 @@ class Automata:

 
             cursor += 1
 
+        if initial_state is None:
+            raise AutomataError("The automaton doesn't have a initial
state")
+

Yeah that's needed.

quoted

         states = sorted(set(states))
 
         # Insert the initial state at the beginning of the states

@@ -132,7 +140,7 @@ class Automata:

             #  ------------ event is here ------------^^^^^
             if self.__dot_lines[cursor].split()[1] == "->":
                 line = self.__dot_lines[cursor].split()
-                event = line[-2].replace('"', '')
+                event = line[-2].replace('"', "")
 
                 # when a transition has more than one labels, they are
like
this
                 # "local_irq_enable\nhw_local_irq_enable_n"

@@ -162,7 +170,9 @@ class Automata:

             nr_state += 1
 
         # declare the matrix....
-        matrix = [[self.invalid_state_str for x in range(nr_event)] for y
in
range(nr_state)]
+        matrix = [
+            [self.invalid_state_str for x in range(nr_event)] for y in
range(nr_state)
+        ]
 
         # and we are back! Let's fill the matrix
         cursor = self.__get_cursor_begin_events()

@@ -170,9 +180,9 @@ class Automata:

         while self.__dot_lines[cursor].lstrip()[0] == '"':
             if self.__dot_lines[cursor].split()[1] == "->":
                 line = self.__dot_lines[cursor].split()
-                origin_state = line[0].replace('"', '').replace(',', '_')
-                dest_state = line[2].replace('"', '').replace(',', '_')
-                possible_events = line[-2].replace('"',
'').replace("\\n", "
")
+                origin_state = line[0].replace('"', "").replace(",", "_")
+                dest_state = line[2].replace('"', "").replace(",", "_")
+                possible_events = line[-2].replace('"',
"").replace("\\n", "
")

All in all looks good, thanks.

Reviewed-by: Gabriele Monaco <gmonaco@redhat.com>

quoted

                 for event in possible_events.split():
                     matrix[states_dict[origin_state]][events_dict[event]]
=
dest_state
             cursor += 1