How to read from console window in visual studio?

Karthik Agarwal

15y2.9k

I have some text like
T_411_A/T_414_B,T_412_A/T_415_A
in which every one of them i.e T_411_A, T_414_B so on are all defined as macro's as follows:

#define T_411_A 0x2341

every one of them has a unique value. 0x2341 is a unique and the max size of the value which macro may consist of is 4 bytes.

in the above example i need to read those T_411_A or any other macro's from console window as then get the value of the macro.

example: if i give T_411_A in the command prompt then i should get the value of that 0x2341 according to the above example.
Note: this is C code where i am working on visual studio 2008.

22 Replies

Know the answer? Post it — somebody with the same question will find it here.

Vulpes15y ago
Accepted answer
As macros are a compile time concept (the C compiler replaces the macro with its value in the source code), it's not possible to read in the macro from the console at runtime and for it then to automatically be replaced by its value.

What you could do instead is to create two arrays:

* A string array containing the macros (as string constants).

* An int array containing the corresponding values of the macros.

You'd then need to iterate through the first array until you found the macro and then read off its value by accessing the element at the same index in the second array.

I'm assuming here that the total number of macros is so small that access speed would not be an issue. If it were, then you'd need to implement some sort of hashtable instead - not much fun in C.
0
Sam Hobbs14y ago
The issue here is that Vulpes said "you'd need to implement some sort of hashtable instead - not much fun in C" so I provided two versions of code that shows it is not that much of a problem; one in C and another in C++. I just wanted Karthik to see that it is possible.
0
Sam Hobbs14y ago
Then there is a problem in the software. You are supposed to get notifications after a thread has been marked answered. I do and so should you.

For what it is worth, the reason I get frustrated is because Karthik's initial questions are seldom clear and then Karthik and Vulpes go through many discussions and the problem changes; the problem often becomes something entirely different. I know I can help but Karhik does not explain things. I try to help anyway but Karthik won't explain what Karthik needs well enough for me to help.

It would be one thing if Karthik understood what I suggest and chooses not to do it. It is frustrating for me when Karthik does not understand what I am saying and does not try to understand.
0
Karthik Agarwal14y ago
No sam, think once how will i be able to know if you post it here? As the answer for this post was already accepted i didn't even get mail so that i can look into this.
0
Sam Hobbs15y ago
So are you ignoring everything I did?
0
Sam Hobbs15y ago
Also note that some of the code that I posted is missing; I think all the parts (due to HTML) that are missing are parts that use angle brackets such as in "". So you need to use the code in the zip file if you are going to use the code, or even just look closely.
0
Sam Hobbs15y ago
This is an equivalent program that uses the C++ standard classes. I assume you will need help understanding it, but I also hope you can see that it is smaller.

#include "stdafx.h"

#define T_411_A 0x2341
#define T_414_B 0x2342
#define T_412_A 0x2343

static const char Filename[] = "Data.txt";

class macroitemsmap : public std::map {};

void ProcessRecord(std::string &line, macroitemsmap ¯oitems) {
std::string::const_iterator at_i(line.begin());
std::string::const_iterator begin_i;
std::string token;
if (line.length() == 0)
return;
while (at_i != line.end()) {
begin_i = at_i;
for (; at_i != line.end() && *at_i != '/' && *at_i != ','; ++at_i);
if (at_i != begin_i) {
token.assign(begin_i, at_i);
std::cout << token << ' ' << macroitems[token] << '\n';
}
if (at_i != line.end())
++at_i;// past the token
}
}

int _tmain(int argc, _TCHAR* argv[])
{
std::ifstream stream;
std::string line;
macroitemsmap macroitems;
macroitemsmap::iterator Iterator;
macroitems["T_411_A"] = T_411_A;
macroitems["T_414_B"] = T_414_B;
macroitems["T_412_A"] = T_412_A;
//for (Iterator=macroitems.begin(); Iterator!=macroitems.end(); ++Iterator)
//std::cout << Iterator->first << ' ' << Iterator->second << '\n';
stream.open(Filename);
if (!stream.is_open()) {
std::cout << "Open error\n";
return 0;
}
while (getline(stream, line))
ProcessRecord(line, macroitems);
stream.close();
return 0;
}
0
Sam Hobbs15y ago
Here you go. I think this will work using the existing #defines. Hopefully you can use this console program to test the code. I assume that for testing purposes you can create a test input file that only has a subset of the ids; something that contains the type of data that you put in your initial question here. You will need to add one line for each id for the macroitems array; since that will be thousands of lines, that is why I suggest you try a subset of those codes for testing purposes. If this code works for you, then you will need to create a macroitems array with one line for each macro.

#include "stdafx.h"

// http://www.c-sharpcorner.com/Forums/Thread/147992/how-to-read-from-console-window-in-visual-studio.aspx

#define T_411_A 0x2341
#define T_414_B 0x2342
#define T_412_A 0x2343

static const char Filename[] = "Data.txt";

struct macroitem {
char *id;
int value;
};

int compare(char **arg1, char **arg2)
{
macroitem *pitem1 = (macroitem *)arg1;
macroitem *pitem2 = (macroitem *)arg2;
return _strcmpi( *arg1, *arg2 );
}

void ProcessRecord(char *line, macroitem *macroitems, int itemcount) {
char seps[] = "/,";
char *token = NULL;
char *next_token = NULL;
macroitem *pmacroitem;
void *p;
int n;
n = strlen(line);
if (n == 0)
return;
// usually there will be a newline at the end
if (line[n-1] == '\n')
line[n-1] = 0;
puts(line);
token = strtok_s(line, seps, &next_token);
while ((token != NULL)) {
p = bsearch(&token, (char *)macroitems, itemcount,
sizeof(macroitem), (int (*)(const void*, const void*))compare);
if (p == NULL)
printf("\t%s not found\n", token);
else {
pmacroitem = (macroitem *)p;
printf("\t%s %s %i\n", token, pmacroitem->id, pmacroitem->value);
}
token = strtok_s(NULL, seps, &next_token);
}
}

int _tmain(int argc, _TCHAR* argv[])
{
FILE *stream;
char line[100];
macroitem macroitems[]= {
{"T_411_A", T_411_A},
{"T_414_B", T_414_B},
{"T_412_A", T_412_A}
};
int itemcount = sizeof(macroitems)/sizeof(macroitem);
qsort((void *)macroitems, itemcount, sizeof(macroitem), (int (*)(const void*, const void*))compare);
//for (int i=0; i//printf("%s %i\n", macroitems[i].id, macroitems[i].value);
if (fopen_s(&stream, Filename, "r") != 0 ) {
puts("Open error");
return 0;
}
fgets(line, 100, stream);
while (!ferror(stream) && !feof(stream)) {
ProcessRecord(line, macroitems, itemcount);
fgets(line, 100, stream);
}
fclose(stream);
return 0;
}
0
Sam Hobbs15y ago
No, I am not suggesting that you change the way the macros are defined. My suggestion uses the macros as they exist. Since there are more than a thousand, I would do something similar to what Vulpes suggests except I would write a program that reads the #define statements in, parses each one to get the names only, and then generate the lines that I show in my sample I posted above. Or I might write a Visual Studio macro to do the conversion. However I am not suggesting that you replace the macros as they exist; you can still use the macros and if they are modified in the future then for every one #define modified there would be only one or two additional lines needed to be modified to do what we are discussing here.

Another possibility is to write a macro that generates the additional code, and then you could do a one-time conversion and then in the future when an id is added then there would be only one line that needed changing. I assume that that is not clear, so I will show you an example later.

0
Karthik Agarwal15y ago
Even i too feel the same but i what to check out the possibilities of that so i came up with it. Anyways thanks for both(Sam and Vulpes) of your contribution. When try implementing this i will catch you up.
0
Vulpes15y ago
It's the first time I can remember seeing the question asked but, as I said earlier, I don't see how it is possible to read the macros from the console and have them automatically replaced by their values.

The macro names have been replaced by their values before the program has even been built.

You therefore need to write code to do this and, if there are a lot of defines, Sam's approach looks best.

However, to avoid typing in all the names/values again, you could use my approach to load them from a text file into the struct array when the program starts.
0
Karthik Agarwal15y ago
Surely this is one way of doing but internally in the coding part can't we do anything? Probably i might be the only one who's trying to read macro's from console window. Sounds too odd but it's my requirement what can i do? If i feel that it can't be done in anyway then i will look how to change the structure of the macro's.
0
Vulpes15y ago
For future reference, a practical way to do this would be:

1. To copy all the defines 'en bloc' to a text file.

2. Globally replace all #define symbols with nothing.

3. Read in the file, when the program starts, and split the strings using the space between the name and value as the delimiter.

4. Populate the array of structs from that, parsing the values to UINT32's as you go.
0
Karthik Agarwal15y ago
Probably it's looks like you are asking me change the way the macro's are defined. Is that's what your intention is then may be i can do it in the future when i get time to change the whole structure of the file. But as if now i can't do that.

What i want know exactly is that?
Can i read the name of the macro name from the console window and the get the value of the same defined in the code? As of knowledge i think it's not possible. Sam tell me straight away tell me whether it can be done or not?
0
Sam Hobbs15y ago
I am not sure what you need. I will write something that satisfies what you ask for as best as I understand it but I won't have it before about 12 hours. So until then, if you can tell me if the following can be done with your data then I can do more later:

struct macroitem {
char *id;
int value;
}

struct macroitem macroitems[3];

macroitems[0].id = "T_411_A";
macroitems[0].value = T_411_A;
macroitems[1].id = "T_414_B";
macroitems[1].value = T_414_B;
macroitems[2].id = "T_412_A";
macroitems[2].value = T_412_A;

If it is possible to build an array that maps your ids (macro) to numbers (macro value), then it is possible to sort using qsort. I can show you how to do that but not right now. Then you can read each line in and separate the text that is delimited by a slash or comma and then do a lookup of each id (macro) using bsearch. I can show you how to do that too but not right now.
0
Karthik Agarwal15y ago
If i had to be precise 1575 is the number of id's which i use, but it may go higher day by day. And about the comma and slash they are not just used to delimit the id's, i was to just delimit i would just had used either slash or comma but i have used both.

If the id's are just separated by slash then they mean that the corresponding strings to those will be concatenated in the sequence which id's are present. if they are just separated by comma's then they mean that they are individual strings which can selected from by user from the drop down. But it can combination of both slash and comma.

Anyways i guess there isn't a solution for the question i asked, as macro's will be replaced with their respective value a compile time. Now what i wanted was reading the name of the macro at run time from console window and get it's respective value. Is that possible Sam? If not probably we should end this post here.
0
Sam Hobbs15y ago
That is not what I remember you said in the thread that got very large. In that thread you said that you had to use void pointers due to some other program, but you were vague about what that other program is.

I know very well that software sometimes gets very large and that management does not want to invest the time to re-write it.

I see that your data consists of two pairs and between the pairs is a comma and between the ids in the pairs is a slash; so in other words:

id/id,id/id

So what is the significance of the slash and what is the significance of the comma? If there is no significance other than being a delimiter between ids then an explanation is not so important but if the ids are to be considered as pairs then that affects the logic.

Also note that Vulpes said "I'm assuming here that the total number of macros is so small that access speed would not be an issue". You should have answered the implied question; how many of these ids (macros) are there? Not in the data itself; how many unique values, such as how many #defines are there.
0
Karthik Agarwal15y ago
i don't understand why you are thinking like this. Actually i do name Vulpes many of the times may be bcoz of that you mistook me i guess. I never ever underestimated your capabilities. In the starting days when i joined this forum when i was desperate for some help as a fresher to work on VC++, he use to answer many of my posts. Hence i started following him much but not like i neglected you. And if it comes to why i am using C, I don't know much about that as a successor. But the fact is that it's a very huge file which have many interfaces and people over here took many days to write that and the procedure is still continued when we need some new thing we just extend the same file with new API's and use the instead start writing it from the scratch.
If you remember one day when you posted about the usage of the void pointers, i said that if you guide me i am ready to take up the challenge. If i won't have trust on you i would not have said that word. And one more thing is, if you ask me to change the whole structure i am following and change the language from c to c++ which i can never ever do as it's being very early stage for me to propose something new being a fresher.
0
Sam Hobbs15y ago
Karthik, please see CodeGuru Forums - View Profile: Sam Hobbs; that is me, kind of a previous life of mine. I was the first member to ever get 10,000 posts in that web site, and more than 99% of them were efforts to help others, and many of them was with the C++ language. Most was in the Visual C++ Programming forum, which is not strictly the C++ language but was questions about programming using C++.

Also see My Profile in the MSDN forums; most of the "answers" I provided there was for C++.

I say that because you seem to think I can't help. It is very frustrating for me to not be able to help.

The first Windows program I wrote was a device driver (back in the days before 32-bit Windows existed) that sent data from a SCSI host adapter to a HP LaserJet printer. Most of it was written in C, except I used assembler to rotate a bitmap. I could have gotten rich from that but about three months after I started, HP introduced the LaserJet IV and that made our software unnecessary. My program definitely did some bit-twiddling (working at the bit level) to send the data.

A few years ago I wrote an application that uses Detours to hook into an accounting package so management could audit what their employees were doing. I could have gotten rich from that except the person that I was working with at the time is a fool (I could use other terms but there is no need for that here). I used at least one technique that the CodeGuru C++ experts considered to be too difficult to do.

I can do C and C++. I can do bit-twiddling. I can work with pointers when appropriate. I know that it is better to avoid pointers. The one thing I don't know how to do is to explain to you that you don't need to do the things you are doing.

Please explain why you cannot use C++ to do what you need to do to read the data. C++ has classes to make it easier. If there is a good reason why you cannot use C++ then we can use C, but I don't see a reason why.
0
Vulpes15y ago
Here's some code to do it using the approach I suggested in my first post:

char *macros[] = {"T_411_A", "T_414_B", "T_412_A", "T_415_A"};
UINT32 values[] = {0x2341, 0x2351, 0x2361, 0x2371}; // I've just made these up!
//...
char m[10];
printf("\nInput a macro : ");
scanf("%s", &m);
int index = -1;

for(int i = 0; i < 4; i++)
{
if(!strcmp(macros[i], m))
{
index = i;
break;
}
}

if (index == -1)
{
printf("The macro %s doesn't exist\n", m);
}
else
{
printf("The value of the macro %s is 0x%04x\n", m, values[index]);
}
0
Vulpes15y ago
No, casting the macro string to a UINT32 won't work - you'll just get back the address of the string.

If you want to try it for yourself, then here's some code:

char m[10];
printf("\nInput a macro : ");
scanf("%s", &m);
printf("The value of the macro is %d\n", (UINT32)m);

It's many years since I learnt C and that was from a book (K & R).

However, there seems to be plenty of online C tutorials. Check google for a list:

http://www.google.co.uk/#hl=en&sugexp=ppwl&cp=4&gs_id=e&xhr=t&q=c+tutorial&pf=p&sclient=psy-ab&site=&source=hp&pbx=1&oq=c+tu&aq=0&aqi=g4&aql=&gs_sm=&gs_upl=&bav=on.2,or.r_gc.r_pw.,cf.osb&fp=b037f854561b5e39&biw=1366&bih=643
0
Karthik Agarwal15y ago
Can't i read it as string from console window and type cast it to UINT32 something like that!
Also refer me a site where i can learn pointers may be whole C also will do the job. These days i lost control on whole C.

Also tell me how to read a string from console window and convert to UINT32 as well. i will try that out.
0