Change in sequence due to a newline

[arunj]

Hi!

I am trying to run the following code.

#include <stdio.h>
#include <errno.h>


int main()
{
        char s[10];
        int nread;
        FILE* fp;
        char c;

        fp = fopen("abc", "w+");
        printf("error: %d, error_string: %s", errno, strerror(errno));

        fprintf(fp, "Hello\n");
        fseek(fp, 0, SEEK_SET);

        while((c = fgetc(fp)) != EOF)
                write(1, &c, 1);

        fclose(fp);
        return 0;
}

It compiles fine.
(none):~/prog # gcc -o err err.c
(none):~/prog #

But when I try to run it I am getting -

(none):~/prog # ./err
Hello
error: 0, error_string: Success(none):~/prog #

If I put a newline after printf like below -

printf("error: %d, error_string: %s\n", errno, strerror(errno));

I am getting the following output -

(none):~/prog # ./err
error: 0, error_string: Success
Hello
(none):~/prog #

Why does this change in sequence happen?

Thanks and regards,

Arun

[zacs7]

I hope you're not actually running / compiling it as root, that'd be silly.

Depending on how you've got output buffering set-up, it will flush at the end of every line or when the buffer fills up. Think of it as so,

Scenario 1.
out = "error: 0, error_string: Success(none):"
out ="hello\n" # flush

Scenario 2.
out = "error: 0, error_string: Success(none):\n" #flush
out ="hello\n" #flush

If you want to change buffering use setvbuf or call fflush().

[arunj]

Hi!

Thanks for the reply. What is silly if it is run as root?

Thanks,

Arun

[Salem]

> What is silly if it is run as root?
Because of the amount of extra damage you can do if you get it wrong.

Say for example you had a buggy 'rm' command and you tried to remove /
Running as root will hang you out to dry.
Running as a mortal will tell you to take a hike.

Likewise, reading email as root, surfing the web as root.
The root account is for necessary admin only, not a convenience for day to day work. It's not like it's that hard to leave another terminal window open if you need to do root work.

[arunj]

Hi!

Which method in this piece of code is comparable to # rm /?

Thanks,

Arun

[arunj]

or # rm -rf /?

[Salem]

I was merely highlighting a possible future scenario.

Or are you suggesting that you'll do the right thing when it matters (because I won't believe you).

Or are you going to wait for someone to "own" your box because you've become overly complacent about running as root?

[brewbuck]

Hi!

Which method in this piece of code is comparable to # rm /?

If you have bugs in your code, anything is possible.

Back in the days of DOS, where everything basically "ran as root," I completely blitzed my entire system MORE THAN ONCE by making simple programming errors. It can happen to you.

[matsp]

If you have bugs in your code, anything is possible.

Back in the days of DOS, where everything basically "ran as root," I completely blitzed my entire system MORE THAN ONCE by making simple programming errors. It can happen to you.

I once type "delete *.*" on "the wrong disk" (meaning do delete a:*.*, but I ended up deleting c:*.* instead). Not a good idea.

If you are not developing system code, where you have to run as root to be able to run the code in the first place, I suggest you do "adduser yourname" and then use the new account for code development. You can always set up a "sudo" account for "yourname" and then do "sudo some-command" to perform commands as root.

--
Mats

[brewbuck]

I once type "delete *.*" on "the wrong disk" (meaning do delete a:*.*, but I ended up deleting c:*.* instead). Not a good idea.

One of my boo-boos was trying to access a structure member through a far NULL pointer. A far NULL pointer, of course, points to the interrupt vector table. So I ended up smashing the vector for one of DOS's disk write routines, and it happened inside a TSR, right in the middle of DOS trying to update something in the FAT. Imagine yanking the rug out from under somebody trying to carefully paint a china lamp.

The FAT was hosed, that was a reinstall on that one.

[Kennedy]

To be fair, though, it is sometimes needful to develop as root -- Like when you are attempting to control hardware without drivers. Cannot do it without being root, or at least chmod +S the file -- which is the same thing.

EDIT: Though, I will say that I did do the following command(s) and really screwed myself:
modprobe mtdcore
modprobe jffs2
modprobe mtdram total_size=63488 erasesize=12
modprobe mtdchar
modprobe mtdblock
dd if=fs.img of=/dev/mtd0
mount -t jffs2 /dev/mtdblock0 mnt
cd mnt/mnt/etc
cp * /etc

oops!!!!!! CRAP!!!!, I mean
cp * ../../etc
SNOT BUBBLES!!!!!!

Dumb dumb di dumb, Dumb di dumdum dumb di dumb. <-- meant to be sung not read.

[arunj]

I was merely highlighting a possible future scenario.

Or are you suggesting that you'll do the right thing when it matters (because I won't believe you).

Or are you going to wait for someone to "own" your box because you've become overly complacent about running as root?

Hi!

It often happens that if I don't do the wrong thing and learn how to rectify it again I'll never be able to grab the correct concept.
And I think comments like "because I won't believe you" and "you've become overly complacent about running as root" etc. are a bit personal. Are these lines really necessary to answer a question like " What is the problem if I run it as a root"?

Thanks and reagrds,

Arun