dev-cpp-users

[Dev-C++] c on unix

[Juliano P. B. C. <jp...@bo...>]

    Hello to everybody. I know this isn't a pure c/c++ programming list (=
it
s more a c++ plus dev-c++ programming list, right?) but maybe some of you
could help me with this:=0D
1-Why does gcc for unix tell me "gets" is a dangerous function?=0D
2-What exactly is the difference between the <ENTER> key stroke of unix a=
nd
of windows? I often have problems with it when I read two consecutive val=
ues
 for example:=0D
    for(i=3D0;i<5;i++)=0D
        scanf("%c",&a);=0D
    In the example above, I've tried to use "fflush(stdin), but it seemed=
 to
have no effect. The one thing that worked fine was something like:=0D
    for(i=3D0;i<5;i++)=0D
    {=0D
        scanf("%c",&a);=0D
        getchar(); //strange getchar(). Would it be to read some extra
character of the Unix <ENTER>?=0D
    }=0D
    I hope my second doubt isn't too confusing.=0D
=0D
    Thanks to all.=0D
    Juliano=0D
 =0D
***********************=0D
 Visite tamb=E9m meu site: =0D
 http://www.jpavel.kit.net =0D
=20

RES: [Dev-C++] c on unix

[Daniel K. O. <dan...@ya...>]

 >  1-Why does gcc for unix tell me "gets" is a dangerous function?


Simple answer: buffer overflow. Look at fgets.



 >  2-What exactly is the difference
 >  between the <ENTER> key stroke of unix and
 >  of windows? I often have problems with it when I read two
 >  consecutive values
 >   for example:     for(i=0;i<5;i++)         scanf("%c",&a);
 >  In the example above, I've tried to use "fflush(stdin), but it seemed to
 >  have no effect. The one thing that worked fine was something
 >  like:     for(i=0;i<5;i++)     {         scanf("%c",&a);
 >   getchar(); //strange getchar(). Would it be to read some extra
 >  character of the Unix <ENTER>?     }


Ioannis said to me once: "fflush empties output buffers, not input". sscanf
doesn't read the remaining '\n'. You should use fgets to read a line
(including the '\n', no need to fflush anything), then sscanf what you need
inside the string.





Daniel K. O.


It's OK to be ignorant; it's not OK to play stupid.
http://www.catb.org/~esr/faqs/smart-questions.html

RE: [Dev-C++] c on unix

[Ioannis V. <iv...@em...>]

> -----Original Message-----
> From: dev...@li...=20
> [mailto:dev...@li...] On Behalf=20
> Of Juliano P=E1vel Brasil Cust=F3dio
> Sent: Thursday, April 03, 2003 4:18 PM
> To: Dev-Cpp-Users (dev...@li...
> Subject: [Dev-C++] c on unix
>=20
>=20
>     Hello to everybody. I know this isn't a pure c/c++=20
> programming list (it
> s more a c++ plus dev-c++ programming list, right?) but maybe=20
> some of you
> could help me with this:
>=20
> 1-Why does gcc for unix tell me "gets" is a dangerous function?


Because you can't limit the input to the characters that the provided =
buffer
can hold. Example:


#include <stdio.h>

/* ... */


chas s[10];

/* imagine that you enter 50 characters... */
gets(s);


>=20
> 2-What exactly is the difference between the <ENTER> key=20
> stroke of unix and
> of windows? I often have problems with it when I read two=20
> consecutive values
>  for example:
>=20
>     for(i=3D0;i<5;i++)
>=20
>         scanf("%c",&a);
>=20
>     In the example above, I've tried to use "fflush(stdin),=20
> but it seemed to
> have no effect.=20



fflush(stdin); is not portable (is not guaranteed to work). fflush() is
intended for *output* streams.



The one thing that worked fine was something like:
>=20
>     for(i=3D0;i<5;i++)
>=20
>     {
>=20
>         scanf("%c",&a);
>=20
>         getchar(); //strange getchar(). Would it be to read some extra
> character of the Unix <ENTER>?
>=20
>     }


getchar() reads a character including '\n' (*after* '\n' has been =
entered).


In the above code, '\n' remains in the stream (scanf() doesn't read the
'\n').



>=20
>     I hope my second doubt isn't too confusing.





Ioannis Vranos
=20
* Programming pages: http://www.noicys.freeurl.com
* Alternative URL: http://run.to/noicys


Anti-racism movement: http://www.positive-youthfoundation.org
Iraq Peace Pledge: http://www.peacepledge.org

RE: [Dev-C++] c on unix

[Ioannis V. <iv...@em...>]

> -----Original Message-----
> From: dev...@li...=20
> [mailto:dev...@li...] On Behalf=20
> Of Juliano P=E1vel Brasil Cust=F3dio
> Sent: Thursday, April 03, 2003 4:18 PM
> To: Dev-Cpp-Users (dev...@li...
> Subject: [Dev-C++] c on unix
>=20
>=20
>     Hello to everybody. I know this isn't a pure c/c++=20
> programming list (it
> s more a c++ plus dev-c++ programming list, right?) but maybe=20
> some of you
> could help me with this:
>=20
> 1-Why does gcc for unix tell me "gets" is a dangerous function?


Because you can't limit the input to the characters that the provided =
buffer
can hold. Example:


#include <stdio.h>

/* ... */


chas s[10];

/* imagine that you enter 50 characters... */
gets(s);


>=20
> 2-What exactly is the difference between the <ENTER> key=20
> stroke of unix and
> of windows? I often have problems with it when I read two=20
> consecutive values
>  for example:
>=20
>     for(i=3D0;i<5;i++)
>=20
>         scanf("%c",&a);
>=20
>     In the example above, I've tried to use "fflush(stdin),=20
> but it seemed to
> have no effect.=20



fflush(stdin); is not portable (is not guaranteed to work). fflush() is
intended for *output* streams.



The one thing that worked fine was something like:
>=20
>     for(i=3D0;i<5;i++)
>=20
>     {
>=20
>         scanf("%c",&a);
>=20
>         getchar(); //strange getchar(). Would it be to read some extra
> character of the Unix <ENTER>?
>=20
>     }


getchar() reads a character including '\n' (*after* '\n' has been =
entered).


In the above code, '\n' remains in the stream (scanf() doesn't read the
'\n').



>=20
>     I hope my second doubt isn't too confusing.





Ioannis Vranos
=20
* Programming pages: http://www.noicys.freeurl.com
* Alternative URL: http://run.to/noicys


Anti-racism movement: http://www.positive-youthfoundation.org
Iraq Peace Pledge: http://www.peacepledge.org

Re: [Dev-C++] c on unix

[Luigi S. <gi...@li...>]

On Thu, Apr 03, 2003 at 10:18:22AM -0300, Juliano P=E1vel Brasil Cust=F3d=
io wrote:
> [..]
> 1-Why does gcc for unix tell me "gets" is a dangerous function?

You already got an answer to this question

> 2-What exactly is the difference between the <ENTER> key stroke of unix=
 and
> of windows? I often have problems with it when I read two consecutive v=
alues
>  for example:
>     for(i=3D0;i<5;i++)
>         scanf("%c",&a);
>     In the example above, I've tried to use "fflush(stdin), but it seem=
ed to
> have no effect. The one thing that worked fine was something like:
>     for(i=3D0;i<5;i++)
>     {
>         scanf("%c",&a);
>         getchar(); //strange getchar(). Would it be to read some extra
> character of the Unix <ENTER>?
>     }
> [..]

The newline character is translated (by the editor you are using) into wh=
at=20
is appropriate for the underlying OS, for instance: LF on Unix/Linux,=20
CRLF on Dos/Windows.

However, I think that this is not the point here, in fact you seem to be
concerned with reading characters from a stream. Probably the following
may be of help.

When you use the format "%c" *all* characters found on the stream are
read (even white-spaces, for instance '\n'), if instead you wish to skip
white-spaces, the format "%1s" should be used.

Example:=20
Suppose that the file "data" contains the following 2 rows:
a b
c	d

Remarks: the command "od -cb data" outputs the following 3 rows:
0000000   a       b  \n   c  \t   d  \n
        141 040 142 012 143 011 144 012
0000010

Now if you read the file "data" using:

  ...
  char c;

  while (scanf("%c", &c) !=3D EOF)
     print("'%c'\n", c);
  ...

You will see echoed all characters (one on each line), including
the space (octal code 040), the newlines and the tab.=20
Instead if you use:

  ...
  char s[] =3D "\0"; /* 2 chars, both are '\0' */

  while (scanf("%1s", s) !=3D EOF)
     print("'%c'\n", *s);
  ...

You will only see the letters "a, b, c, d" (one per line).
--=20
>> gisan >>