Added move constructor for MIPSImage

[Nydauron]

This PR adds a move constructor to MIPSImage.

There is some debate about whether to add a copy constructor (which would come in handy if we don't want to prepare the original assembly file, can technically (although seems rather inefficiently) allow the ability to rollback the state of the MIPS program to an earlier point in time).

Adding a copy constructor would require the use of malloc() to make sure that it is a deep copy rather than a shallow copy. It is sort of unclear as to what needs to be copied over as there are a lot of source pointers (pointers that were created by malloc()) and referential pointers (pointers that were shallow copied).

[philvukovic]

changes look good. but from my understanding, a move assignment operator should also be implemented when we add a move constructor, so I would suggest we add that

[philvukovic]

There is some debate about whether to add a copy constructor (which would come in handy if we don't want to prepare the original assembly file, can technically (although seems rather inefficiently) allow the ability to rollback the state of the MIPS program to an earlier point in time).

This could be added in the future if it's desired, but it doesn't seem necessary right now, and there are some potential complications in performing a full deep copy of the image. Move semantics fit the average use case better

[Nydauron]

This could be added in the future if it's desired, but it doesn't seem necessary right now, and there are some potential complications in performing a full deep copy of the image. Move semantics fit the average use case better

Sg. Created issue #14

[Nydauron]

Added assignment operators. Used the following code to test it:

diff --git a/spim/worker.cpp b/spim/worker.cpp
index 89fe55a..ea59df7 100644
--- a/spim/worker.cpp
+++ b/spim/worker.cpp
@@ -82,8 +82,10 @@ void reset(unsigned int max_contexts, std::set<unsigned int> &active_ctxs) {
         char file_name[64];
         sprintf(file_name, "./input_%d.s", i);
         if (read_assembly_file(new_image, file_name)) { // check if the file exists
-            new_image.reg_image().PC = starting_address(new_image);
-            ctxs.emplace(i, std::move(new_image));
+            MIPSImage img(i);
+            img = std::move(new_image);
+            img.reg_image().PC = starting_address(img);
+            ctxs.emplace(i, std::move(img));
         }
         yylex_destroy();
     }

Albeit it is rather inefficient, hence why its only being used for testing and not for production

For compiling for native, I also added the following to main() (mainly for the purposes of Valgrind):

diff --git a/main.cpp b/main.cpp
index d163359..417e703 100644
--- a/main.cpp
+++ b/main.cpp
@@ -1,2 +1,18 @@
 // Need to provide main() function for JsSpim to work
-int main() {}
+#include "spim/worker.h"
+#include <set>
+int main() {
+    std::set<unsigned int> active_ctxs({0, 1});
+
+    reset(2, active_ctxs);
+    reset(2, active_ctxs);
+    active_ctxs.clear();
+    active_ctxs.insert(1);
+    reset(2, active_ctxs);
+    active_ctxs.clear();
+    active_ctxs.insert(0);
+    reset(2, active_ctxs);
+    shutdown();
+
+    exit(0);
+}

[philvukovic]

wanted some clarification on the change of sink from & to *. maybe this is fine for our use case, but it should be well-documented how the ostream is managed as it is a bit counter-intuitive