Interpreter #267
Interpreter #267
Conversation
Added dead command to trace
including: bitwise operations variable size arrays dereferencing dead objects dynamical objects input enumeration (trace backtracking)
Avoid crash from infinite array size
| default: { | ||
| out << "unknown type: " << type << std::endl; | ||
| assert(false); | ||
| } |
There was a problem hiding this comment.
I'm generally in favour of omitting the "default" clause when all cases are covered, so as to let the compiler warn about this. But that view isn't necessarily shared by others.
That said: the placement of { and } does not match coding guidelines.
| @@ -154,6 +154,12 @@ class goto_tracet | |||
| steps.push_back(step); | |||
| } | |||
|
|
|||
| //Retrieves the final step in the trace | |||
| inline goto_trace_stept &get_last_step() | |||
There was a problem hiding this comment.
No need to use "inline" within class definition.
| #include <algorithm> | ||
| #include <string.h> |
There was a problem hiding this comment.
Use cstring, not string.h
| @@ -31,8 +41,41 @@ Function: interpretert::operator() | |||
|
|
|||
| void interpretert::operator()() | |||
| { | |||
| show=true; | |||
| message->status() << "0- Initialize:" << messaget::eom; | |||
| initialise(true); | |||
There was a problem hiding this comment.
At some point it was agreed to use American English across the code base, so please use "initialize".
| { | ||
| message->error() << e << messaget::endl << messaget::eom; | ||
| } | ||
| while(!done) command(); |
There was a problem hiding this comment.
So the first error will be caught, but further errors won't?
| } | ||
| else | ||
| { | ||
| return false; |
There was a problem hiding this comment.
You may just want to use has_prefix(...) instead of this function
| @@ -136,27 +336,35 @@ void interpretert::step() | |||
| { | |||
| PC=call_stack.top().return_PC; | |||
| function=call_stack.top().return_function; | |||
| stack_pointer=call_stack.top().old_stack_pointer; | |||
| //stack_pointer=call_stack.top().old_stack_pointer; | |||
| //TODO: this increases memory size quite quickly. Should check alternatives | |||
There was a problem hiding this comment.
Should this be fixed right away?
| const struct_typet &struct_type=to_struct_type(real_type); | ||
| const struct_typet::componentst &components=struct_type.components(); | ||
| for(struct_typet::componentst::const_iterator it=components.begin(); | ||
| it!=components.end();++it) { |
| /* Used to determine whether or not an identifier can be built | ||
| * before trying and getting an exception */ | ||
| static bool can_build_identifier(const exprt &src); | ||
|
|
There was a problem hiding this comment.
I don't understand those changes.
| return static_cast<const member_exprt &>(op).symbol(); | ||
| } | ||
| return to_symbol_expr(op); | ||
| } |
There was a problem hiding this comment.
The root object need not be a symbol - it could, for example, be an array. Use object_descriptor_exprt to compute the root_object().
|
This PR has been superseded by the interpreter improvements in test-gen-support. Changes should be back-ported from there if desired. I keep this here as a reminder: #659 |
…-evs-pretty-printer SEC-105 Update pretty printer for access path new location
…gression-tests Fix dependence graph regression tests
No description provided.