Another "Undefined Reference to..."

I cannot for the life of my figure this out, probably have been staring at it too long. I have been getting this error

Code:

---

main.c:(.text+0x90): undefined reference to `lreceive'

---

Here is what I typically use to compile it:

Code:

---

gcc -o a main.c linsock.c -Wall -pthread -pedantic

---

I also tried compiling both of them into separate object files and then linking them, but to no avail...
What I am doing is creating a thread, but the function the thread uses is in another source file. Would that make a difference?

main.c

Code:

---

pthread_create(&(info.tid), NULL, lreceive, (void *)&info);

---

linsock.c

Code:

---

void *lreceive (void *params)
{
        struct linfo *info = (struct *linfo)params;
        int result = 1;
        char *buffer = malloc(2048 * sizeof(char));

        while (result)
        {
                result = recv(info->fsock, buffer, 2048, 0);
                if (result)
                {
                        printf("%d bytes received\n");
                }
                else
                        break;
        }

        return NULL;
}

---

Any words of wisdom?

Your main.c needs to include the header where the function is declared.

It does, sorry for not mentioning it.

Quote:

---

Originally Posted by carrotcake1029

It does, sorry for not mentioning it.

---

Then you don't get that error. See for instance:
main.c

Code:

---

include "linsock.h"

int main() {
    void *(*f)(void*);
    f = lreceive;
    return 0;
}

---

and linsock.c

Code:

---

#include <stdio.h>

void *lreceive(void *params) {
    printf("Hello from linsock!\n");
    return params;
}

---

(with linsock.h being what you expect) which gives the following errors/warnings:

Code:

---

$ gcc -o a main.c linsock.c -Wall -pedantic
$

---

I tried that and it compiled for me, but mine still doesn't...

I will post some trimmed down files for you, because it sounds like I am overlooking something here:

Sorry for the long post.

main.c

Code:

---

#include <stdio.h>
#include <stdlib.h>
#include <string.h>

#include "linsock.h"

int main (void)
{
        struct linfo info;
        memset(&info, 0, sizeof(struct linfo));
       
        info.host = "google.com";
        info.port = 80;
        info.timeout = 5;

        printf("Attempting to connect to %s on port %d\n", info.host, info.port);
        lconnect(&info);
        if (info.fsock)
        {
                printf("Connected\n");
                pthread_create(&(info.tid), NULL, lreceive, (void *)&info);
                pthread_join(info.tid, NULL);
        }
        else
        {
                printf("%s\n", strerror(errno));
        }

        close(info.fsock);
        return 0;
}

---

linsock.c

Code:

---

#include "linsock.h"

void lconnect (struct linfo *info)
{
    /*
        Yadda Yadda Yadda
    */
}

void *lreceive (void *params)
{
        struct linfo *info = (struct *linfo)params;
        int result = 1;
        char *buffer = malloc(2048 * sizeof(char));

        while (result)
        {
                result = recv(info->fsock, buffer, 2048, 0);
                if (result)
                {
                        printf("%d bytes received\n");
                }
                else
                        break;
        }

        return NULL;
}

void lresolve (struct linfo *info)
{
    /*
        Yadda Yadda Yadda
    */
}

---

linsock.h

Code:

---

#ifndef _LINSOCK_H_
#define _LINSOCK_H_

#include <stdio.h>
#include <string.h>
#include <stdlib.h>

#include <fcntl.h>

#include <sys/time.h>
#include <sys/types.h>

#include <sys/types.h>
#include <sys/socket.h>

#include <pthread.h>

#include <sys/socket.h>
#include <unistd.h>
#include <netinet/in.h>
#include <netdb.h>
#include <arpa/inet.h>

#include <errno.h>

struct linfo
{
        int fsock;
        char *host;
        char *ip;
        short port;
        int timeout;
        struct sockaddr_in service;
        pthread_t tid;
};

void lresolve (struct linfo *info);
void lconnect (struct linfo *info);
void *lreceive (void *params);

#endif

---

After I move the * where it belongs (as in "struct linfo*") on line 12, I get:

Code:

---

$ gcc -o a main.c linsock.c -Wall -lpthread -pedantic
linsock.c: In function ‘lreceive’:
linsock.c:21: warning: too few arguments for format

---

and that's it.

Edit: And it runs, says it is trying to get to google.com on port 80, and then dies with unknown error 0.

Well crap man, why is my gcc screwing up. I'm using gcc version 4.3.2. I highly doubt the fact that my gcc is about a week out of date makes a difference...

Well, I'm running 4.0 so I doubt they broke it since then. Tell you what: can you call the function lreceive? As in just "lreceive(&info)" and see what happens. I doubt you would have fixed a spelling in copy-paste, but I'm starting to run out of ideas.

Edit: Oh, and something else that shouldn't make a difference: you can try &lreceive in the lpthread_create call and see if that helps. (I don't see how it could, but hey.)

I cannot. I'm out of ideas too.

If you can't call it, then either (a) the header isn't there, somehow, or (b) the function name is mysteriously spelled differently in the header or (c) something else I haven't thought of yet. If you're not getting errors for anything else, then (a) is right out. Edit: And since it's a link error and not a compile error, (b) is right out too.

Yes, currently that is the only error I am getting. I think I agree with part (c).

Alright, well, what's your command line? Are you using a makefile?

I've tried both command line and Makefile. I have tried compiling both sources into object files first and then compiling them. Both of them ultimately end up with the same error.

Silly idea: put linsock.c first on the command line before main.c (does ld still need functions found before use? I thought they fixed that).

Not getting anything new. It could be that I am possibly using a bad compile of gcc. It is doubtful, I know, but I'll compile the newest one tomorrow maybe.

gcc -E linsock.c > linsock.i
Then look at linsock.i in a text editor (it'll be very large).

What you're looking at is what the compiler sees after the pre-processor has had it's turn.
What you should look for is
- whether the code for the function even exists - perhaps it was in a conditional block, or a comment
- whether the function still has the same name, perhaps a macro substitution gone wrong.

Well, I ran that and sure enough I could not find my function. I found the other two. All I found of lreceive was a declaration, but no function. I also searched for something similar but with a different name, but there wasn't any.

If anyone wants to see the file, it is attached.

So I'm guessing it's in a big fat comment then?

Put a single open comment /* or close comment */ inside the function, and see what sort of error messages you get out of it.

Looking at your .i file is pointless, try the .c file.

Ok. Strange update. I went to sleep and woke up, decided to give it another shot. I opened linsock.c, and the function was gone. I put it back in and now get the same warning tabstop did.

I wonder what happened last night. I am using gedit. Could it possibly not have been saving it correctly?

Code:

---

warning: too few arguments for format

---

Anyone know how to fix this. I can't find a clear cut definition from my google queries.

When you print a number with %d, you need to have a number to print.

Haha, wow. Thank you.
Thanks for your perseverance. I have no idea why I was having those issues. Last night I even tried closing and reopening gedit but I still had that same problem. It ended up that doing a restart caused me to realize that function was not being saved, for whatever reason it was.

Most likely it was being saved elsewhere (subdirectory? parent directory? who knows?)

Ah, the old "not compiling what you forgot to save", or "compiling the wrong copy" sketch.
Seen it many times before, it's a classic.

Quote:

---

Originally Posted by Salem

Ah, the old "not compiling what you forgot to save", or "compiling the wrong copy" sketch.
Seen it many times before, it's a classic.

---

Yup, you haven't been a programmer for long if you haven't ever made that mistake. Of course, modern IDE's help a whole lot by not really bothering with where things are saved, autosaving, and such things.

I've also been known to wonder why my code isn't printing my debug outputs and such things, only to figure out that I'm compiling for release-mode or that I've actually not copied the new code to the target system.... It's really embarrasing when it takes one of your colleagues to figure it out, when you should know better! A trick that I've used a few times is to put "sdjhasdk6&%$" somewhere in the code I'm editing, just to get an error - that way, I know I'm compiling the code I think I'm compiling [where I work now, we tend to have "drives" created with Windows/DOS subst that contain different projects - sometimes I've been working on the same bit of code in two different projects - it can get REALLY confusing at times].

--
Mats