WEBVTT 00:00:00.740 --> 00:00:03.650 I will start this lesson with a simple experiment. 00:00:04.040 --> 00:00:07.520 I declared an array with three integers, and I'm wondering 00:00:07.530 --> 00:00:10.790 what will be the output of this sizeof operator when I 00:00:10.790 --> 00:00:12.650 pass it the name of the array. 00:00:13.040 --> 00:00:16.550 Since we said that the name of the array works like a pointer, 00:00:16.560 --> 00:00:20.170 this should probably return 8, because on this system 00:00:20.170 --> 00:00:22.660 we need 8 bytes to store a pointer. 00:00:23.040 --> 00:00:27.150 However, as you can see, the result is not 8, it is 12. 00:00:27.540 --> 00:00:29.080 Like we mentioned before, 00:00:29.090 --> 00:00:33.010 array names can be used like pointers in certain situations, 00:00:33.010 --> 00:00:34.950 but this is not one of them. 00:00:35.340 --> 00:00:39.820 When you declare an array, C++ will add a little bit of magic, 00:00:39.830 --> 00:00:44.530 a little bit of syntactic sugar to it. Compiler will actually know that 00:00:44.530 --> 00:00:49.960 this name is assigned to an array and it will return 12 bytes, because 00:00:49.970 --> 00:00:52.760 that is the actual size of the whole array. 00:00:53.140 --> 00:00:58.150 Each integer needs 4 bytes, so 3 integers make an array of 12 bytes. 00:00:58.740 --> 00:01:02.210 This is the size of the array in memory, but you may have 00:01:02.210 --> 00:01:04.599 also noticed that in the previous lesson, 00:01:04.610 --> 00:01:08.720 I used the word size to refer to the number of elements. 00:01:08.720 --> 00:01:11.490 Terminology for this is not consistent, 00:01:11.490 --> 00:01:15.050 but I guess I could have said that the number of elements is the 00:01:15.050 --> 00:01:19.110 length of the array. But since I'm only going to use the sizeof 00:01:19.110 --> 00:01:21.610 operator with arrays in this lesson, 00:01:21.620 --> 00:01:24.760 I will keep referring to the array length as size. 00:01:25.140 --> 00:01:29.760 Now, let's say that I want to pass this array to a function as an argument. 00:01:30.140 --> 00:01:33.280 The name of the function will be printArray and the 00:01:33.280 --> 00:01:35.560 purpose of the function is obvious. 00:01:35.940 --> 00:01:40.140 The function will print the elements from the given array, but 00:01:40.140 --> 00:01:43.850 notice the syntax for using the array as a parameter. 00:01:44.240 --> 00:01:47.360 This is the correct way to do this and it will work. 00:01:47.740 --> 00:01:52.060 The parameter will accept the array of integers of any size, 00:01:52.440 --> 00:01:56.070 but we still need to know the size of the array, because we will need 00:01:56.070 --> 00:01:58.860 it to create a for loop inside of the function. 00:01:59.240 --> 00:02:02.910 Since this expression gives us the size of the whole array, 00:02:02.920 --> 00:02:08.259 maybe we can divide it by the size of one integer to get the number of elements. 00:02:08.840 --> 00:02:14.200 Let's try to compile this, and it works. Now, let's do the 00:02:14.200 --> 00:02:16.660 same thing from inside of the function. 00:02:17.840 --> 00:02:22.830 I will pass this parameter to the sizeof operator, compile 00:02:22.830 --> 00:02:27.490 this to make sure that it's still working, and it seems that 00:02:27.490 --> 00:02:32.160 it's not, because we get 2. But why? 00:02:33.140 --> 00:02:37.840 Also, it's hard not to notice this big warning message which says that the 00:02:37.840 --> 00:02:42.050 sizeof operator will return the sizeof a pointer to an integer. 00:02:42.440 --> 00:02:45.750 This is because of a somewhat dangerous and really common 00:02:45.750 --> 00:02:49.610 occurrence of something called a pointer decay. Whenever 00:02:49.610 --> 00:02:51.440 you pass an array to a function, 00:02:51.450 --> 00:02:54.610 the parameter that should accept the array will decay 00:02:54.610 --> 00:02:57.260 from an array to a simple pointer. 00:02:57.640 --> 00:03:01.050 This is known as decay because a little bit information 00:03:01.050 --> 00:03:03.940 has been lost. Inside of this function, 00:03:03.950 --> 00:03:07.920 the compiler no longer adds that syntactic sugar to our array 00:03:07.920 --> 00:03:10.660 that lets us get the size of all elements. 00:03:12.040 --> 00:03:15.110 So this is the same as defining this parameter as a 00:03:15.110 --> 00:03:17.050 simple pointer to an integer. 00:03:17.440 --> 00:03:21.480 The square bracket syntax is allowed because of programmers, so that we 00:03:21.480 --> 00:03:25.150 can be more explicit about our intentions in the code. 00:03:25.540 --> 00:03:27.050 If you use square brackets, 00:03:27.050 --> 00:03:30.680 compiler will figure out that your intention is to pass an 00:03:30.680 --> 00:03:34.690 array, so, that is why we got that big warning message that 00:03:34.690 --> 00:03:36.550 warns us about pointer decay. 00:03:36.940 --> 00:03:40.810 You can also use a simple pointer to pass an array, but 00:03:40.820 --> 00:03:42.660 only if you know what you're doing. 00:03:43.040 --> 00:03:46.410 So, using this sizeof operator to get the number of 00:03:46.410 --> 00:03:49.660 elements inside of a function is not possible. 00:03:50.040 --> 00:03:51.790 Whenever you want to pass an array, 00:03:51.800 --> 00:03:55.460 you also need to pass another argument which will inform the function 00:03:55.470 --> 00:03:58.560 about the number of elements inside of that array. 00:03:58.940 --> 00:04:02.340 I will continue referring to this as size, so that's 00:04:02.340 --> 00:04:04.060 how I will name the parameter. 00:04:04.440 --> 00:04:06.180 So now that we know the size, 00:04:06.180 --> 00:04:09.660 we can print all of the elements by looping through the array. 00:04:10.540 --> 00:04:13.920 I use the pointer arithmetic to do this, just like we 00:04:13.920 --> 00:04:16.360 learned in one of the previous lessons. 00:04:16.740 --> 00:04:17.390 However, 00:04:17.399 --> 00:04:21.990 if you know that this decayed pointer will always represent an array, you can 00:04:21.990 --> 00:04:25.560 use the square brackets operator just like you normally would. 00:04:25.940 --> 00:04:30.330 So, other than passing this size explicitly, everything else works in 00:04:30.330 --> 00:04:34.310 the same way. You could have also put the expected size here in the 00:04:34.310 --> 00:04:37.260 square brackets, but this wouldn't do anything. 00:04:37.270 --> 00:04:39.360 This would still be just a pointer. 00:04:39.740 --> 00:04:43.130 I also need to mention that you should not use this approach for 00:04:43.130 --> 00:04:47.890 calculating size, even in the main function, It is not reliable, 00:04:47.890 --> 00:04:50.180 and once we learn about dynamic arrays, 00:04:50.180 --> 00:04:53.860 you will see that this will only work with static arrays like this one. 00:04:54.240 --> 00:04:57.550 You should always personally keep track of the size and pass it 00:04:57.550 --> 00:05:00.750 to functions that require array arguments. 00:05:01.140 --> 00:05:04.760 Now, let's implement this new concept in our calculator. 00:05:06.140 --> 00:05:12.220 I will create a special function for each operation, sum function will add all 00:05:12.220 --> 00:05:16.960 of the elements, and multiply will, of course, multiply them. 00:05:17.340 --> 00:05:22.160 The first argument is the array of numbers, and the second one is the 00:05:22.160 --> 00:05:27.210 size of the array, just like we learned. The function logic is the same 00:05:27.210 --> 00:05:32.170 as before, we just need to define the result in each function and then 00:05:32.170 --> 00:05:33.760 print it out to the terminal. 00:05:34.740 --> 00:05:38.350 Now I can remove the result variable from the main function 00:05:38.360 --> 00:05:41.360 and delete the logic from every case. 00:05:41.740 --> 00:05:43.360 We don't need this anymore. 00:05:44.140 --> 00:05:48.520 Instead, I will just call our two new functions and pass 00:05:48.520 --> 00:05:51.360 them the numbers array and its size. 00:05:52.240 --> 00:05:55.840 This will make our main function more readable, but we will 00:05:55.840 --> 00:05:58.560 continue to optimize it in the next lesson.