I am learning how to develop a Frama-C plugin. After reading the Frama-C Developer manual and doing the CFG plugin example, I tried to do a basic script that prints all annotations in a C file. I came up with this:
open Cil_types
class print_annot out = object
inherit Visitor.frama_c_inplace
method vstmt_aux s =
let annots = Annotations.code_annot s in
let anleng = List.length annots in
if anleng <= 0 then Format.fprintf out "Empty List\n"
else
Format.fprintf out "Number of Annotations: %d\n" anleng;
List.iter (fun annot -> Format.fprintf out " -> s%d\n" annot.annot_id) annots;
Cil.DoChildren
end
let run () =
let chan = open_out "annots.out" in
let fmt = Format.formatter_of_out_channel chan in
Visitor.visitFramacFileSameGlobals (new print_annot fmt) (Ast.get());
close_out chan
let () = Db.Main.extend run
It always says the list is empty even when the input file has ACSL annotations and never prints the annotations id. What am I doing wrong?
Edit:
An example with the following code:
/*# requires y >= 0;
# ensures \result >= 0;
*/
int g(int y){
int x=0;
if(y>0){
x=100;
x=x+50;
x=x-100;
}else{
x = x - 150;
x=x-100;
x=x+100;
}
return x;
}
void main(){
int a = g(0);
}
And invoking frama-c with:
$ frama-c -load-script annot_script.ml condi.c
Gives the following output:
Empty List
Empty List
Empty List
Empty List
Empty List
Empty List
Empty List
Empty List
Empty List
Empty List
Empty List
In your example, none of the annotations are attached to a statement. The requires and the ensures are part of a function contract and are attached to the function g, not to any statement of g.
An annotation that would be attached to a statement would for instance be /*# assert x == 150; */ after the line x=x+50;.
If I modify condi.c to insert this assertion, then with the same commandline, I get:
Empty List
Empty List
Empty List
Empty List
Number of Annotations: 1
-> s1
Empty List
Empty List
Empty List
Empty List
Empty List
Empty List
Empty List
It seems that your script is working as expected, for a value of “expected” that corresponds to printing annotations attached to statements.
Related
I have a list declared outside a for loop and then I assign some values to this list inside that for loop and its value is updated when printed inside the loop but when I print it after the loop, it gives me an empty list.
List<List<String>> chunkSizeCollection(List<String> followedList) {
int counter = 0;
int ongoingCounter = 0;
bool isLessThanTen = false;
List<List<String>> returnAbleChunkedList = [];
List<String> midList = [];
log("in followedlist");
followedList.forEach((element) {
if (counter == 0) {
int difference = followedList.length - ongoingCounter;
if (difference < 10) {
// log("in difference if: $difference");
isLessThanTen = true;
}
}
midList.add(element);
counter++;
ongoingCounter++;
if (counter == 10 || (isLessThanTen && ongoingCounter == followedList.length)) {
returnAbleChunkedList.add(midList);
log("returnAbleChunkedList in counter 10 after adding new val is: $returnAbleChunkedList");
//above log works properly and prints the updated list
midList.clear();
counter = 0;
}
});
//this log on the other hand, returns an empty list
log("returnAbleChunkedList: $returnAbleChunkedList before return");
return returnAbleChunkedList;
}
The output:
[log] returnAbleChunkedList in counter 10 after adding new val is: [[KQTuEPllbmRrlBNvYgZ7oUXwtA63, OZUZOE10IzT8quUFoZbNxZOynU32, fCYIlYemCvbLTc7SpNHw6fCHrcm1, CbcLrtDNOdYZyC23FzEehOrJbKx2, FFvvVHCpPGNKUiXPQD34QdoPqH32, Gk09y59vSVXa1HNhzYvc6Atqnt53, JDO356z8urYQvuktJmc6eNUYqSm2, YesvvNI43gUVYPMfqhG4uRO5t6K2]]
[log] returnAbleChunkedList: [[]] before return
In this line:
returnAbleChunkedList.add(midList);
you add a reference to midList to your output list. If your input is longer than 10, you'll end up adding more than one reference to midList to the output list (i.e. you might now have 2). Subsequently, you clear midList so you now have an output list that contains 2 references to a list that you've now cleared, so you end up with a list containing 2 empty lists.
If, instead, you changed this to:
returnAbleChunkedList.add(midList.toList());
you'd add a copy of midList to your output list and your end up with:
[[a, b, c, d, e, f, g, h, i, j], [k]]
as you expected.
The problem is this:
returnAbleChunkedList.add(midList);
midList.clear();
You add the object midlist to the list and then you clear it. So the object you added will be emptied. You have to create a copy or add the elements of the list.
returnAbleChunkedList.add(midList.toList());
// or
returnAbleChunkedList.add(List.of(midList));
// or
returnAbleChunkedList.add([...midList]);
I have a strange problem when I use a for loop to add element in to the java.util.ArrayList , but the list's address of the reference always changing
Here is the code:
var curntRow: Row = null
var startTime: lang.Long = null
//this is the list
var standTime: util.ArrayList[Row] = new util.ArrayList[Row]()
for (row <- usersCoorOrderByTime) {
if (curntRow == null) {
startTime = row.getAs[lang.Long](2)
} else if (!row.getAs[String](1).equals(curntRow.getAs[String](1))) {
//And I use the method list.add() right here
standTime.add(Row(row.getAs[String](0), row.getAs[String](1), row.getAs[DoubleType](4), row.getAs[DoubleType](5), curntRow.getAs[lang.Long](2) - startTime))
startTime = row.getAs[lang.Long](2)
}
curntRow = row
}
And please see the pic that I debug below:
addr is "7703"
Before get in the loop The list's addr is "7703"
When is get in the loop ,the address changes
change to "11268"
change to "11287"
The most strange things is when it end the loop, the address has changed back to where it was originally declared
change back to "7703"
finally I get an empty ArrayList
I found the error
the parameter of for loop is Dataframe, I should turn to Array or List then make for loop
Try using Mutable ArrayBuffer. Below is a simple example. I hope it helps.
val x = scala.collection.mutable.ArrayBuffer[String]()
x += "2"
x += "4"
println(x)
using UnityEngine;
using System.Collections;
public class NewMonoBehaviour1 : MonoBehaviour
{
void ConcatExample(int[] intArray)
{
string line = intArray[0].ToString(); // the line is the var of the first in array
for(i =1;i <intArray.Length; i++) // the length is unknown ?
{
line += ", " + intArray[i].ToString(); //
}
return line;
//each time allocate new in original place
}
}
How can this function work ? the length of array is unknown , so how the for loop works ?Besides, this is void function but shouldn't return anythings right ,or is there any exceptional case ,finally,according to the unity manual, it is said that the function will keep producing a string but with new contents in the same place , resulting in consuming large memory space .Why ?thx
What makes you think that the Length should be unknown? It is a property that any array simply has
Gets the total number of elements in all the dimensions of the Array.
Of course it is not unknown the moment you call your method with an according parameter!
The return line; will not even compile since as you say the method is of type void so it can not return anything. It should probably be private string ConcatExample
Then what the unity manual (don't know where exactly you read this) means lies in
line += ", " + intArray[i].ToString();
under the hood every string in c# is an immutable char[]. So everytime you do a string concatenation via stringC = stringA + stringB what happens under the hood is basically something similar to
char[] stringC = new char[stringA.Length + stringB.Length];
for(var iA = 0; iA < stringA.Length; iA++)
{
stringC[i] = stringA[i];
}
for(var iB = 0; iB < stringB.Length; iB++)
{
stringC[iB + stringA.Length] = stringB[iB];
}
so whenever dealing with loops especially with large data it is strongly recommended to rather use a StringBuilder like
private string ConcatExample(int[] intArray)
{
var builder = new StringBuilder(intArray[0]);
for(i =1; i < intArray.Length; i++)
{
builder.Append(", ").Append(intArray[i].ToString());
}
return builder.ToString();
}
The length of the array will be the length of the array of ints you pass into the function as an argument.
say you pass it
Int[] ints = {1,2,3}
ConcatExample(ints); //the length of the array is now 3
add a debug.log() function to the ConcatExample method
void ConcatExample(int[] intArray)
{
string line = intArray[0].ToString();
for (int i = 1; i < intArray.Length; i++)
{
line += ", " + intArray[i].ToString(); //
Debug.Log(line);
}
}
debug.log would produce the following in the console
1, 2
1, 2, 3
and finally the return line; at the end would just result in an error because yes you are correct void returns nothing
This function CANNOT work, unless it gets the data it expects. A NULL passed to this function, for example, would generate a runtime null-reference exception. Passing a valid integer array, of length zero would generate an invalid index error on the first line.
You are correct, the function returns nothing, and appears pointless. In fact, I would have expected return line; to generate a complier error.
The string type appears "dynamic" meaning, it will indeed allocate more and more memory as needed. Technically, it is actually the string "+" operator, (a function that takes two strings as parameters) that is allocating this space. This function returns a new string, of the appropriate size. The garbage collector will DEallocate "old" strings when they are no longer referenced by any variables.
How can I apply constrains to list of list, similarly to what I can do to simple list:
list_size: uint;
my_list: list of uint;
keep my_list.size() == list_size;
keep for each (item) using index (item_index) in my_list { item == item_index;};
My intention is to create something like:
list_size:uint;
grosslist_size:uint;
my_grosslist: list of list of uint;
keep my_grosslist.size() == grosslist_size;
keep for each (grossitem) using index (grossindex)in my_grosslist {
grossitem.size() == list_size;
// keep for each (item) using index (item_index) in grossitem {
// item == item_index + grossindex * 100;
// };
};
How can I write 3 lines commented above using Specman syntax?
Please note that constrains are for instance only, in reality I'll need to apply much more sophisticated ones rather than indexing list items...
Thanks in advance.
The code you wrote is indeed the correct usage of list-of-list. Note that there was a missing space and the additional 'keep' is not needed for the internal for each. other than that, it works.
<'
extend sys {
list_size:uint;
grosslist_size:uint;
my_grosslist: list of list of uint;
keep my_grosslist.size() == grosslist_size;
keep for each (grossitem) using index (grossindex) in my_grosslist {
grossitem.size() == list_size;
for each (item) using index (item_index) in grossitem {
item == item_index + grossindex * 100;
};
};
};
'>
I'm using getElementsByTagName to return all the select lists on a page - is it possible to then filter these based upon an option value, ie of the first or second item in the list?
The reason is that for reasons I won't go into here there are a block of select lists with number values (1,2,3,4,5 etc) and others which have text values (Blue and Black, Red and Black etc) and I only want the scripting I have to run on the ones with numerical values. I can't add a class to them which would more easily let me do this however I can be certain that the first option value in the list will be "1".
Therefore is there a way to filter the returned list of selects on the page by only those whose first option value is "1"?
I am pretty sure that there is a better solution, but for the moment you can try something like:
var allSelect = document.getElementsByTagName("select");
var result = filterBy(allSelect, 0/*0 == The first option*/, "1"/* 1 == the value of the first option*/);
function filterBy(allSelect, index, theValue) {
var result = [];
for (var i = 0; i < allSelect.length; i++) {
if(allSelect[i].options[index].value == theValue ) {
result.push(allSelect[i]);
}
}
return result;
}
I managed to get this working by wrapping a simple IF statement around the action to be performed (in this case, disabling options) as follows:
inputs = document.getElementsByTagName('select');
for (i = 0; i < inputs.length; i++) {
if (inputs[i].options[1].text == 1) {
// perform action required
}
}
No doubt there is a slicker or more economic way to do this but the main thing is it works for me.