WEBVTT 00:00.490 --> 00:06.350 Cplusplus inherits from a C the idea of argument list. 00:06.370 --> 00:11.470 So to do this you use the ellipsis syntax like this here. 00:13.320 --> 00:14.280 Three dots. 00:14.290 --> 00:21.730 So as the last parameter to indicate that the color can provide zero or more parameters. 00:21.730 --> 00:28.030 So the compiler will check how the function is called and will allocate space on the stack for these 00:28.060 --> 00:29.080 extra parameters. 00:29.080 --> 00:43.160 So to access the extra parameters, your code must include the header file here include include zstd 00:43.810 --> 00:47.050 i g here as you can see here. 00:47.050 --> 00:55.210 So this your code must include the uh, this header file which has macros that you can use to extract 00:55.210 --> 00:57.910 the extra parameters of the stack. 00:57.910 --> 01:04.420 So this is inherently type unsafe because the compiler cannot check that that parameters that the function 01:04.420 --> 01:10.900 will get of the stack at runtime will be the same type as the parameters put on the stack by the calling 01:10.900 --> 01:11.140 code. 01:11.140 --> 01:21.860 For example, let's make another function here named some integer, some integer first and after with 01:21.860 --> 01:23.150 the comma after the comma. 01:23.180 --> 01:26.210 Operator We're going to paste the three dots here. 01:26.210 --> 01:30.440 And in this in this, some function, we're going to create an integer. 01:32.400 --> 01:34.020 Integer sum zero. 01:34.320 --> 01:37.110 Var list here arguments. 01:37.960 --> 01:39.490 Verse start here. 01:40.030 --> 01:41.770 We're going to this takes two arguments. 01:41.770 --> 01:45.970 The first is arguments and we're going to last here. 01:45.970 --> 01:50.230 We're going to pass the first and integer is first. 01:51.090 --> 01:59.550 Enter here first, while the E here, while E is not equal to minus one. 01:59.580 --> 02:09.480 Then the sum here increment a by one here one increment by one. 02:10.320 --> 02:17.490 And equals to f a arguments args here. 02:25.780 --> 02:26.140 Here. 02:26.140 --> 02:29.800 And after that, we're going to use the VA. 02:30.110 --> 02:32.560 VA and. 02:33.500 --> 02:37.820 Pass these arguments here and return the sum lastly here. 02:37.820 --> 02:43.550 So the definition of the function must have at least one parameter so that the macros work here. 02:44.180 --> 02:49.910 In this case, the parameter is called the first, and it's important that your code leaves this stack 02:49.910 --> 02:51.440 in a consistent state. 02:51.440 --> 03:00.260 And this is carried out using a variable named Val type here named arguments that Val has the list type 03:00.290 --> 03:00.470 here. 03:00.470 --> 03:07.750 So this variable is initialized at the beginning of the function by calling the vastart here macro via 03:07.760 --> 03:16.070 start macro and the stack is restored to its previous state at the end of the function by calling the 03:16.070 --> 03:18.890 var and macro here. 03:18.890 --> 03:26.600 So the code in this function simply iterates through the argument list and maintains a sum and the loop 03:26.600 --> 03:36.150 here finishes when the parameter has the value of minus one and there are no macros, there are no macros 03:36.840 --> 03:41.310 to give an indication of the type of the parameter on the stack. 03:41.310 --> 03:52.140 So your code has to assume the type of the variable to provide the desired type with the var args here. 03:53.310 --> 03:57.210 Var arg macro here. 03:57.210 --> 03:59.820 As you can see here we are macro here. 03:59.820 --> 04:04.290 So in this example the vendor var arg is called assuming. 04:04.290 --> 04:12.020 Then the A is assuming that the every parameter on the stack has has type of integer here. 04:12.030 --> 04:19.860 So once our parameters has been read of the stack, the in this code here, we're going to call the 04:20.790 --> 04:26.040 previous code that we've write it here C out some minus one. 04:26.940 --> 04:29.550 And of course end line. 04:29.820 --> 04:32.010 See out some. 04:33.450 --> 04:44.790 Uh, minus one, minus six, minus six, minus five, minus four, minus three, minus two and less 04:44.790 --> 04:45.900 than minus one. 04:46.950 --> 04:52.740 And end line here and see out some. 04:54.270 --> 04:56.100 And here. 04:59.190 --> 05:01.590 And lastly minus one here. 05:01.800 --> 05:05.940 And this gonna end line here so. 05:08.390 --> 05:09.260 Here. 05:09.950 --> 05:11.360 Firstly, we're going to. 05:12.170 --> 05:12.680 Okay. 05:13.610 --> 05:19.310 So here, once parameters have been has been read off the stack. 05:19.340 --> 05:25.630 The code calls var and rear before returning to. 05:26.420 --> 05:29.120 Before returning to the sum here, as you can see here. 05:29.850 --> 05:32.450 Uh, so the function can be called like this. 05:32.480 --> 05:34.640 In this case, we're going to get the zero. 05:34.640 --> 05:38.660 And in this case, we're going to get the -12. 05:38.780 --> 05:42.920 And in this case, we're going to get a 16 here. 05:44.500 --> 05:52.810 So since the minus one is used to indicate that the end of the list, it means that the sum of zero 05:52.810 --> 05:54.790 number of parameters. 05:54.790 --> 06:00.910 So you have to pass at least one parameters, as you can see here. 06:00.910 --> 06:02.080 So. 06:03.450 --> 06:06.060 And that is the minus one here. 06:06.060 --> 06:10.310 In addition, the second line shows the here's the second line. 06:10.320 --> 06:19.050 This shows the, um, the second line shows that you have a problem if you are passing a list of negative 06:19.050 --> 06:20.220 numbers here. 06:21.850 --> 06:27.580 So in this case, the minus one cannot be a parameter here. 06:27.580 --> 06:36.220 So the problem could be addressed in this implementation by choosing another marker value. 06:36.220 --> 06:44.530 So another implementation could skip that the use of marker for the end of the list and instead of first 06:44.530 --> 06:48.760 required argument to give the count of parameters that follow here. 06:48.760 --> 06:57.100 So now while we're going to change, we're going to write it actually from zero to some equals zero. 06:57.130 --> 07:00.550 Var list args v v a start. 07:00.580 --> 07:02.620 We're going to first pass the. 07:03.420 --> 07:04.050 Arguments. 07:04.050 --> 07:11.220 And here we're going to change this name to count and count here. 07:11.220 --> 07:19.650 And while we're going to change this while code here, the while the count is count here. 07:20.470 --> 07:21.580 Integer. 07:29.300 --> 07:34.540 And here integer the var args here. 07:37.880 --> 07:38.750 Integer. 07:40.250 --> 07:41.420 Here, as you can see. 07:41.450 --> 07:42.110 Count. 07:42.260 --> 07:42.920 Count. 07:43.070 --> 07:43.820 Oh, sorry. 07:46.530 --> 07:47.450 Count here. 07:47.460 --> 07:52.860 So and lastly, the sum of E here. 07:53.130 --> 07:56.940 And after that, we're going to firstly end the arguments here. 07:56.970 --> 08:00.870 Arguments and then return the sum. 08:02.080 --> 08:03.570 Return the sum here. 08:03.580 --> 08:05.380 So this time. 08:07.220 --> 08:09.380 Here this time. 08:10.560 --> 08:12.360 The first value here. 08:12.390 --> 08:17.060 The first value is the number of arguments that follows. 08:17.100 --> 08:23.970 And so the routine will extract this number of the integers of the stack and sum them here. 08:23.970 --> 08:31.230 So this code is called when we call this this code, we're going to get the zero if we call this. 08:32.050 --> 08:34.840 And then I get -21. 08:35.020 --> 08:39.910 And if we call this, we're going to get the 60 here. 08:39.910 --> 08:42.970 So let's make it run. 08:42.970 --> 08:46.450 And as you can see here, oops, we got an error here. 08:46.450 --> 08:54.040 So this is no convention for how to handle the determining how many parameters have been passed here. 08:54.040 --> 08:59.080 So the return assumes here, the routine assumes we have to see count. 08:59.740 --> 09:00.370 Okay. 09:01.400 --> 09:03.650 Here and let's run it again. 09:04.670 --> 09:05.150 Here. 09:13.480 --> 09:14.170 Here. 09:14.170 --> 09:20.140 So the routine assumes that the ever items on the stack is an integer. 09:20.140 --> 09:25.180 But there is no information about this in the prototype of the function here. 09:25.180 --> 09:33.430 So the compiler cannot do type checking on the parameters actually used to the call function here. 09:33.430 --> 09:39.340 If the caller provides a parameter of different type, the wrong number of bytes should be read off 09:39.340 --> 09:40.240 the stack here. 09:40.240 --> 09:42.700 So for example, the sum. 09:43.830 --> 09:46.020 And, uh, like this, for example. 09:46.050 --> 09:48.240 See out some. 09:49.000 --> 09:49.690 Here. 09:49.720 --> 09:53.570 310 dot and 2030. 09:53.950 --> 09:59.500 So it's easy to press both the comma and period keys here at the same time. 09:59.500 --> 10:03.490 And this has happened after trying the temp parameter. 10:03.490 --> 10:11.350 So the period means that the ten here, this means here ten is a double and this compiler puts a double 10:11.350 --> 10:12.610 value on the stack here. 10:12.610 --> 10:20.020 So in this case, the first is integer and double integer integer here, integer integer here. 10:20.260 --> 10:21.400 So. 10:22.390 --> 10:27.070 The compiler is going to put a double value on the stack. 10:27.070 --> 10:32.430 So when the function reads values of the stack with the var argument. 10:32.660 --> 10:41.830 Var argument macro it will read the eight bytes double as two byte integer values and for code produced 10:41.860 --> 10:49.060 by this compiler, the results in a total sum of just a random number here. 10:49.060 --> 10:55.000 So this illustrates the type unsafe aspect of arguments list here. 10:55.000 --> 11:00.280 So you cannot get the compiler to do the type checks of the parameter passed to the function here. 11:00.280 --> 11:07.690 So if your function has different types passed to it, then you have to implement some mechanism to 11:07.690 --> 11:11.380 determine what those parameters are.