Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.c > #86420

Re: getopt() function

Show all headers | View raw

Geoff <geoff@invalid.invalid> writes:
> It's a horrible function, don't use it. You are much better off
> examining command line arguments and parsing them in your program
> explicitly.

Whatever its flaws, getopt() is standard (POSIX, not ISO C), and using
means that your program's handling of command-line arguments is likely
to be consistent with that of other programs.  If each program reinvents
the wheel by implementing its own option handling, you're likely to have
gratuitous inconsistencies.

> The example given on that web page is flawed and doesn't print Usage
> unless an unexpected option is used. It should print usage when no
> options are used as well.

Not providing any *options* is valid (that's why they're called
"options"), and shouldn't result in an error message.  Not providing any
*arguments* is an error for this particular program, and I agree that it
should result in a usage message rather than just an error message.

>                           Furthermore, it doesn't properly handle -t
> since it prints error "expected argument after options" whether you
> use -t, -t3, -t 3 on command line.

The program requires zero or more *options* followed by at least one
argument that is not an option.  (It could be a file name, for example.)
The "expected argument after options" message is about the lack of an
argument after the options, not about the lack of an argument to the -t
option.

If you run
    ./getopt_demo -t
it complains that "-t" requires an argument.  If you run
    ./getopt_demo -t arg
it takes "arg" as the argument to the "-t" option, and then complains
that there's no argument after the options.  If you run
    ./getopt_demo -t42 arg
or
    ./getopt_demo -t 42 arg
it takes 42 as the argument to the "-t" option, and "arg" as the
final non-option program argument.

The program is a bit terse (for example the use of "tfnd" to
indicate whether the "-t" option was found on the command line).
It's also a bit weak on error checking (the argument to -t should
be an integer but it doesn't check that), but I suggest that would
clutter the code.

If you want to improve the example, you can submit a suggestion to
the Linux man-pages project, <https://www.kernel.org/doc/man-pages/>.
I might do so myself.

> If the man page examples can't use it correctly, how can you expect to
> use it correctly?
>
> I don't think Richard Heathfield was joking. I don't think the man
> page author(s) understand getopt either.

I think you've misunderstood the example in the man page.  I agree that
the example could be improved.

There's another usage example in the POSIX documentation:
http://pubs.opengroup.org/onlinepubs/9699919799/functions/getopt.html

-- 
Keith Thompson (The_Other_Keith) kst-u@mib.org  <http://www.ghoti.net/~kst>
Working, but not speaking, for JetHead Development, Inc.
"We must do something.  This is something.  Therefore, we must do this."
    -- Antony Jay and Jonathan Lynn, "Yes Minister"

Back to comp.lang.c | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread

Thread

getopt() function Alla _ <modelling.data@gmail.com> - 2016-04-15 05:05 -0700
  Re: getopt() function Richard Heathfield <rjh@cpax.org.uk> - 2016-04-15 13:09 +0100
    Re: getopt() function Alla _ <modelling.data@gmail.com> - 2016-04-15 05:25 -0700
    Re: getopt() function gpietsch08618@gmail.com - 2016-04-26 22:39 -0700
      Re: getopt() function Alla _ <modelling.data@gmail.com> - 2016-04-27 00:09 -0700
        Re: getopt() function gpietsch08618@gmail.com - 2016-05-31 05:44 -0700
  Re: getopt() function Geoff <geoff@invalid.invalid> - 2016-04-15 06:58 -0700
    Re: getopt() function Alla _ <modelling.data@gmail.com> - 2016-04-15 07:54 -0700
      Re: getopt() function Richard Heathfield <rjh@cpax.org.uk> - 2016-04-15 15:58 +0100
      Re: getopt() function Geoff <geoff@invalid.invalid> - 2016-04-15 08:24 -0700
    Re: getopt() function Keith Thompson <kst-u@mib.org> - 2016-04-15 08:48 -0700
      Re: getopt() function Geoff <geoff@invalid.invalid> - 2016-04-15 09:29 -0700
        Re: getopt() function Keith Thompson <kst-u@mib.org> - 2016-04-15 11:20 -0700
    Re: getopt() function Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-04-16 16:51 -0700
      Re: getopt() function janus <emekamicro@gmail.com> - 2016-05-31 22:23 -0700
        Re: getopt() function Malcolm McLean <malcolm.mclean5@btinternet.com> - 2016-05-31 23:27 -0700
          Re: getopt() function janus <emekamicro@gmail.com> - 2016-06-01 00:27 -0700
  Re: getopt() function Geoff <geoff@invalid.invalid> - 2016-04-15 08:37 -0700
    Re: getopt() function Geoff <geoff@invalid.invalid> - 2016-04-15 08:46 -0700
      Re: getopt() function Keith Thompson <kst-u@mib.org> - 2016-04-15 12:11 -0700
        Re: getopt() function Geoff <geoff@invalid.invalid> - 2016-04-15 13:43 -0700
  Re: getopt() function Ian Collins <ian-news@hotmail.com> - 2016-04-16 08:34 +1200
    Re: getopt() function Alla _ <modelling.data@gmail.com> - 2016-04-18 08:40 -0700
      Re: getopt() function Randy Howard <rhoward.mx@EverybodyUsesIt.com> - 2016-04-18 15:20 -0500
  Re: getopt() function boon <fred900rbc@gmail.com> - 2016-04-18 06:26 -0700

csiph-web