> Hi,

> I've not been using C for quite a while.  I recall that if a programmer
> names a function itoa, the same as the standard C function name,

>I've not been using C for quite a while.  I recall that if a programmer
>names a function itoa, the same as the standard C function name, there

> >Hi,

> >I've not been using C for quite a while.  I recall that if a
programmer
> >names a function itoa, the same as the standard C function name,
there

> There is no standard C function caled itoa (although there is one
called
> atoi).

> >is a way to choose which one the programmer intends to use.

> In general defining a function that clashed with a standard library
function
> results in undefined behaviour, the whole program is effectively
invalid.
> There is no standard way around this other than to rename the
offending
> function.

> >In my case
> >I am working on compiling a fairly large editor that someone wrote
and
> >it has an 'itoa' function that is different from the standard C
> >function as defined in stdlib. Is there a GCC compile switch that I
can
> >use to override the standard C definition of the function named itoa?

> Since itoa is not a standard C function a conforming implementation
> should not be declaring it in stdlib.h or any other standard header.
> Your compiler may have options to "turn off" non-standard extensions
> such as this. However gcc is very often used with system-supplied
> headers so you may have to research the particular headers you are
> using as much as gcc.

> For more information about gcc try asking in gnu.gcc.help, for
information
> about your system headers ask in a newsgroup related to the system
> or library that they are supplied with. I don't normally recommend
> rummaging system headers but in this case it might be worth while to
take
> a look. Non-standard features are often included conditionally through
> the use of #if and #ifdef and control macros. You may find it set up
> to exclude extensions when certain macros are defined (e.g. with -
D... on
> the compiler command line). However things can be more complex than
they
> first appear and if you can find the "official" method in the
> documentation.

> >(without changing the programmer written itoa to my_itoa)

> That would probably be the most sensible approach. If you have the
complete
> source code it would be trivial to search and replace all instances
> of itoa in it.

> Please don't be rude. You expect us to read your question here, you
can
> at least have the basic courtesy to read our responses here. It helps
> you to because an incorrect response will almost certainly be
corrected.

> --
> -----------------------------------------


> -----------------------------------------

>Atoi is defined in the stdlib.h as having 3 arguments but the
>application source codes define atoi as having only 1 argument.  Both
>are of type character pointer.

> Atoi is defined in the stdlib.h as having 3 arguments but the
> application source codes define atoi as having only 1 argument.  Both
> are of type character pointer.

> Well, let me try compiling with MSVC++ just to see if the problem goes
> away.

> Enoch

> ...

> >Thank you very much for the comprehensive answer.  I made a mistake
of
> >typing itoa when I really meant atoi. I'd like to apologize for that.

> >Atoi is defined in the stdlib.h as having 3 arguments but the
> >application source codes define atoi as having only 1 argument.  Both
> >are of type character pointer.

> In that case your code rerally is broken and needs fixing by renaming
> the atoi() function it defines to something else.

> >Well, let me try compiling with MSVC++ just to see if the problem
goes
> >away.

> The problem is going to plague you on various compilers until you fix
it.
> Fixing it shouldn't require a great deal of effort and will be the
> least painful option in the long run.

> --
> -----------------------------------------


> -----------------------------------------

>    [snip]

> > Thank you very much for the comprehensive answer.  I made a mistake
of
> > typing itoa when I really meant atoi. I'd like to apologize for
that.

> > Atoi is defined in the stdlib.h as having 3 arguments but the
> > application source codes define atoi as having only 1 argument.
Both
> > are of type character pointer.

> > Well, let me try compiling with MSVC++ just to see if the problem
goes
> > away.

> > Enoch

> There seems to be some confusion here.  The prototype for the standard
> ANSI/ISO function atoi is:

> int atoi(const char *s);

> The ANSI/ISO standard function accepts only one argument.  Are you
> sure you are not still confusing atoi() and itoa()?  Some compilers
> supply a non-standard extension by that name, and some of them (but
> not all of them) take 3 parameters.

> I have never heard of an implementation of atoi() which takes other
> than the standard single pointer to char.

> Jack Klein
> --
> Home: http://jackklein.home.att.net