Partial data flows of MemberExpression nodes from/to their base

[oxisto]

This PR changes the way we define the DFG for member expressions. Previously, we connected all of them directly to their FieldDeclaration. This is problematic and inconsistent with the way we handle variables/locals in the control sensitive DFG.

This new approach connects the member expression with their base and introduces DFG granularities properties on the DFG edge to denote that this flow is only "partial" and only matches a certain field.

It also includes a very experimental DFG mermaid printer.

Loading

%%{init: {"flowchart": {"htmlLabels": false}} }%%
flowchart TD
-834991509["`s
(VariableDeclaration)
field.c(8:19-8:20)`"]-->|"DFG"|-1496149577["`s
(Reference)
field.c(9:3-9:4)`"]
-1496149577["`s
(Reference)
field.c(9:3-9:4)`"]-->|"DFG"|-1150393309["`s
(Reference)
field.c(11:15-11:16)`"]
-1150393309["`s
(Reference)
field.c(11:15-11:16)`"]-->|"DFG (partial, myStruct::field1)"|-98834102["`myStruct::field1
(MemberExpression)
field.c(11:15-11:23)`"]
1721867210["`myStruct::field1
(MemberExpression)
field.c(9:3-9:11)`"]-->|"DFG (partial, myStruct::field1)"|-1496149577["`s
(Reference)
field.c(9:3-9:4)`"]
1390425348["`
(Literal)
field.c(9:14-9:15)`"]-->|"DFG"|1721867210["`myStruct::field1
(MemberExpression)
field.c(9:3-9:11)`"]
-98834102["`myStruct::field1
(MemberExpression)
field.c(11:15-11:23)`"]-->|"DFG"|460955417["`i
(ParameterDeclaration)
field.c(5:18-5:23)`"]

[oxisto]

It would be interesting, whether we could introduce a "full" DFG flow from the member expression in 9:3 to the one in 11:15. not sure if this is possible?

[oxisto]

I added an even more hacky approach to this, to have something which is somewhat object-sensitive as well:

Loading

%%{init: {"flowchart": {"htmlLabels": false}} }%%
flowchart TD
-741817423["`s
(VariableDeclaration)
field.c(8:19-8:20)`"]-->|"DFG (full)"|2076022928["`s
(Reference)
field.c(9:3-9:4)`"]
548976981["`myStruct::field1
(MemberExpression)
field.c(9:3-9:11)`"]-->|"DFG (partial, myStruct::field1)"|2076022928["`s
(Reference)
field.c(9:3-9:4)`"]
2076022928["`s
(Reference)
field.c(9:3-9:4)`"]-->|"DFG (full)"|-1873188100["`s
(Reference)
field.c(11:15-11:16)`"]
-1873188100["`s
(Reference)
field.c(11:15-11:16)`"]-->|"DFG (partial, myStruct::field1)"|-1271724331["`myStruct::field1
(MemberExpression)
field.c(11:15-11:23)`"]
2120835341["`
(Literal)
field.c(9:14-9:15)`"]-->|"DFG (full)"|548976981["`myStruct::field1
(MemberExpression)
field.c(9:3-9:11)`"]
548976981["`myStruct::field1
(MemberExpression)
field.c(9:3-9:11)`"]-->|"DFG (full)"|-1271724331["`myStruct::field1
(MemberExpression)
field.c(11:15-11:23)`"]
-1271724331["`myStruct::field1
(MemberExpression)
field.c(11:15-11:23)`"]-->|"DFG (full)"|-1602965580["`i
(ParameterDeclaration)
field.c(5:18-5:23)`"]

[oxisto]

Here is another example using two structs:

struct myStruct {
  int field1;
};

void doSomething(int i) {}

int main() {
  struct myStruct s1;
  struct myStruct s2;

  s1.field1 = 1;
  s2.field1 = 1;

  doSomething(s1.field1);
  doSomething(s2.field1);
}

This results in the following DFG:

Loading

flowchart TD
-1453381603["s1
(VariableDeclaration)
field.c(8:19-8:21)"]-->|"DFG (full)"|2023738053["s1
(Reference)
field.c(11:15-11:17)"]
-1453381603["s1
(VariableDeclaration)
field.c(8:19-8:21)"]-->|"DFG (full)"|1681679713["s1
(Reference)
field.c(13:3-13:5)"]
2023738053["s1
(Reference)
field.c(11:15-11:17)"]-->|"DFG (partial, myStruct::field1)"|-998413250["myStruct::field1
(MemberExpression)
field.c(11:15-11:24)"]
1681679713["s1
(Reference)
field.c(13:3-13:5)"]-->|"DFG (full)"|2028360463["s1
(Reference)
field.c(16:15-16:17)"]
940621758["myStruct::field1
(MemberExpression)
field.c(13:3-13:12)"]-->|"DFG (partial, myStruct::field1)"|1681679713["s1
(Reference)
field.c(13:3-13:5)"]
-998413250["myStruct::field1
(MemberExpression)
field.c(11:15-11:24)"]-->|"DFG (full)"|800622645["i
(ParameterDeclaration)
field.c(5:18-5:23)"]
202134477["myStruct::field1
(FieldDeclaration)
field.c(2:7-2:13)"]-->|"DFG (full)"|-998413250["myStruct::field1
(MemberExpression)
field.c(11:15-11:24)"]
2028360463["s1
(Reference)
field.c(16:15-16:17)"]-->|"DFG (partial, myStruct::field1)"|-850496130["myStruct::field1
(MemberExpression)
field.c(16:15-16:24)"]
940621758["myStruct::field1
(MemberExpression)
field.c(13:3-13:12)"]-->|"DFG (full)"|-850496130["myStruct::field1
(MemberExpression)
field.c(16:15-16:24)"]
1496727722["
(Literal)
field.c(13:15-13:16)"]-->|"DFG (full)"|940621758["myStruct::field1
(MemberExpression)
field.c(13:3-13:12)"]
-850496130["myStruct::field1
(MemberExpression)
field.c(16:15-16:24)"]-->|"DFG (full)"|800622645["i
(ParameterDeclaration)
field.c(5:18-5:23)"]
-792283555["myStruct::field1
(MemberExpression)
field.c(17:15-17:24)"]-->|"DFG (full)"|800622645["i
(ParameterDeclaration)
field.c(5:18-5:23)"]
998834333["myStruct::field1
(MemberExpression)
field.c(14:3-14:12)"]-->|"DFG (full)"|-792283555["myStruct::field1
(MemberExpression)
field.c(17:15-17:24)"]
2057914096["s2
(Reference)
field.c(17:15-17:17)"]-->|"DFG (partial, myStruct::field1)"|-792283555["myStruct::field1
(MemberExpression)
field.c(17:15-17:24)"]
998834333["myStruct::field1
(MemberExpression)
field.c(14:3-14:12)"]-->|"DFG (partial, myStruct::field1)"|1711233346["s2
(Reference)
field.c(14:3-14:5)"]
1525386664["
(Literal)
field.c(14:15-14:16)"]-->|"DFG (full)"|998834333["myStruct::field1
(MemberExpression)
field.c(14:3-14:12)"]
1711233346["s2
(Reference)
field.c(14:3-14:5)"]-->|"DFG (full)"|2057914096["s2
(Reference)
field.c(17:15-17:17)"]
-1452428260["s2
(VariableDeclaration)
field.c(9:19-9:21)"]-->|"DFG (full)"|1711233346["s2
(Reference)
field.c(14:3-14:5)"]

[KuechA]

I tried to add another "full" DFG edge between the two occurrences of o.i in the example above. However, I'm not sure if the flow from o to o.i makes sense or if it should be turned around to a flow from o.i to o (which seems to be more consistent to me).

Loading

flowchart TD
-665835678["o
(VariableDeclaration)
nested.c(12:16-12:17)"]-->|"DFG (full)"|1132347668["o
(Reference)
nested.c(13:3-13:4)"]
1132347668["o
(Reference)
nested.c(13:3-13:4)"]-->|"DFG (full)"|1478103936["o
(Reference)
nested.c(15:15-15:16)"]
1132347668["o
(Reference)
nested.c(13:3-13:4)"]-->|"DFG (partial, outer::i)"|381457095["outer::i
(MemberExpression)
nested.c(13:3-13:6)"]
1478103936["o
(Reference)
nested.c(15:15-15:16)"]-->|"DFG (partial, outer::i)"|-1439244217["outer::i
(MemberExpression)
nested.c(15:15-15:18)"]
381457095["outer::i
(MemberExpression)
nested.c(13:3-13:6)"]-->|"DFG (full)"|-1439244217["outer::i
(MemberExpression)
nested.c(15:15-15:18)"]
-1091946730["inner::field
(MemberExpression)
nested.c(13:3-13:12)"]-->|"DFG (partial, inner::field)"|381457095["outer::i
(MemberExpression)
nested.c(13:3-13:6)"]
410745339["outer::i
(FieldDeclaration)
nested.c(6:16-6:17)"]-->|"DFG (full)"|381457095["outer::i
(MemberExpression)
nested.c(13:3-13:6)"]
-1439244217["outer::i
(MemberExpression)
nested.c(15:15-15:18)"]-->|"DFG (partial, inner::field)"|-784138326["inner::field
(MemberExpression)
nested.c(15:15-15:24)"]
-1091946730["inner::field
(MemberExpression)
nested.c(13:3-13:12)"]-->|"DFG (full)"|-784138326["inner::field
(MemberExpression)
nested.c(15:15-15:24)"]
-1476331749["
(Literal)
nested.c(13:15-13:16)"]-->|"DFG (full)"|-1091946730["inner::field
(MemberExpression)
nested.c(13:3-13:12)"]
-117842857["
(ConstructExpression)
null"]-->|"DFG (full)"|410745339["outer::i
(FieldDeclaration)
nested.c(6:16-6:17)"]
-784138326["inner::field
(MemberExpression)
nested.c(15:15-15:24)"]-->|"DFG (full)"|1001880489["i
(ParameterDeclaration)
nested.c(9:18-9:23)"]
-674809415["inner::inner
(ConstructorDeclaration)
null"]-->|"DFG (full)"|-117842857["
(ConstructExpression)
null"]

[sonarqubecloud]

Quality Gate passed

Issues
4 New issues
0 Accepted issues

Measures
0 Security Hotspots
88.1% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarCloud