About "program" and "command"

About "program" and "command"

[Xue Fuqiao]

The doc string of `call-process' says:

  Call PROGRAM synchronously in separate process.

The doc string of `call-process-shell-command' says:

  Execute the shell command COMMAND synchronously in separate process.

What's the different of these two functions?  In other words, what's the
difference of "program" and "command" here?  After searching the
archives, I found two threads[1][2], but they don't help.  So can I get
a pointer to information on these two concepts?


[1] http://lists.gnu.org/archive/html/help-gnu-emacs/2010-09/msg00009.html
[2] http://lists.gnu.org/archive/html/help-gnu-emacs/2007-10/msg00059.html

-- 
Xue Fuqiao
http://www.gnu.org/software/emacs/

Re: About "program" and "command"

[Andreas Röhler]

Am 24.03.2013 13:32, schrieb Xue Fuqiao:
> The doc string of `call-process' says:
>
>    Call PROGRAM synchronously in separate process.
>
> The doc string of `call-process-shell-command' says:
>
>    Execute the shell command COMMAND synchronously in separate process.
>
> What's the different of these two functions?  In other words, what's the
> difference of "program" and "command" here?  After searching the
> archives, I found two threads[1][2], but they don't help.  So can I get
> a pointer to information on these two concepts?
>
>
> [1] http://lists.gnu.org/archive/html/help-gnu-emacs/2010-09/msg00009.html
> [2] http://lists.gnu.org/archive/html/help-gnu-emacs/2007-10/msg00059.html
>

Hi,

the commander, that's you. I.e. a command is a function designed for interactive use.
While a program is every recipe directed at computer, wherefrom wherever.

Probably most commanders here also are programmers :)

Andreas

Re: About "program" and "command"

[Yuri Khan]

On Sun, Mar 24, 2013 at 7:32 PM, Xue Fuqiao <xfq.free@gmail.com> wrote:
> The doc string of `call-process' says:
>
>   Call PROGRAM synchronously in separate process.
>
> The doc string of `call-process-shell-command' says:
>
>   Execute the shell command COMMAND synchronously in separate process.
>
> What's the different of these two functions?  In other words, what's the
> difference of "program" and "command" here?

I would expect that “program” refers to only the binary being invoked,
without its command-line arguments; while “command” refers to the
whole command including the binary and the arguments, and, with the
explicit qualification as a “shell command”, probably also with such
shell-specific constructs as environment variables, input/output
redirection, piping, etc.

Of course, my expectations could be completely wrong.

Re: About "program" and "command"

[Barry Margolin]

In article <mailman.22746.1364128338.855.help-gnu-emacs@gnu.org>,
 Xue Fuqiao <xfq.free@gmail.com> wrote:

> The doc string of `call-process' says:
> 
>   Call PROGRAM synchronously in separate process.
> 
> The doc string of `call-process-shell-command' says:
> 
>   Execute the shell command COMMAND synchronously in separate process.
> 
> What's the different of these two functions?  In other words, what's the
> difference of "program" and "command" here?  After searching the
> archives, I found two threads[1][2], but they don't help.  So can I get
> a pointer to information on these two concepts?
> 
> 
> [1] http://lists.gnu.org/archive/html/help-gnu-emacs/2010-09/msg00009.html
> [2] http://lists.gnu.org/archive/html/help-gnu-emacs/2007-10/msg00059.html

A program is just a single external executable. A shell command is a 
whole shell command line -- it can run multiple commands (separated by 
';') and use shell metacharacters (wildcard filenames, backticks, I/O 
redirection and piping, etc.).

It's like the difference between execl() and system() in C programming.

-- 
Barry Margolin, barmar@alum.mit.edu
Arlington, MA
*** PLEASE post questions in newsgroups, not directly to me ***

Re: About "program" and "command"

[Pascal J. Bourguignon]

Xue Fuqiao <xfq.free@gmail.com> writes:

> The doc string of `call-process' says:
>
>   Call PROGRAM synchronously in separate process.
>
> The doc string of `call-process-shell-command' says:
>
>   Execute the shell command COMMAND synchronously in separate process.
>
> What's the different of these two functions?  In other words, what's the
> difference of "program" and "command" here?  After searching the
> archives, I found two threads[1][2], but they don't help.  So can I get
> a pointer to information on these two concepts?

Some shell commands are not programs, but built-in routines in the
shell.  Or also, aliases, functions, etc.

Compare:

(let ((buffer (get-buffer-create (generate-new-buffer-name "*test*"))))
   (call-process-shell-command "set" nil  buffer)
   (switch-to-buffer buffer))

with:

(let ((buffer (get-buffer-create (generate-new-buffer-name "*test*"))))
   (call-process "set" nil  buffer)
   (switch-to-buffer buffer))


In some cases, there are both a command and a program so this may be
confusing.  For example, there is /bin/echo and the builtin echo.  They
may behave differently!

Again, compare:

(let ((buffer (get-buffer-create (generate-new-buffer-name "*test*"))))
   (call-process-shell-command "echo" nil  buffer nil "--help")
   (switch-to-buffer buffer))

with:

(let ((buffer (get-buffer-create (generate-new-buffer-name "*test*"))))
   (call-process "echo" nil  buffer nil "--help")
   (switch-to-buffer buffer))


Bash has a builtin named 'command' that skip looking for functions or
aliases, but still runs the built-in command if there's one.

Bash also has a builtin named 'enable' which lets you enable or disable
a builtin command, so it doesn't shadow the programs.

Also since call-process-shell-command uses the shell to interpret the
'command', one can pass several commands separated with semicolon:

(let ((buffer (get-buffer-create (generate-new-buffer-name "*test*"))))
   (call-process-shell-command "echo a ; echo b" nil  buffer)
   (switch-to-buffer buffer))

(let ((buffer (get-buffer-create (generate-new-buffer-name "*test*"))))
   (call-process-shell-command "echo --help ; enable -n echo ; echo --help" nil  buffer)
   (switch-to-buffer buffer))

or of course, you can always give the path to the program to avoid the
builtin command:

(let ((buffer (get-buffer-create (generate-new-buffer-name "*test*"))))
   (call-process-shell-command "echo --help ; /bin/echo --help" nil  buffer)
   (switch-to-buffer buffer))

-- 
__Pascal Bourguignon__                     http://www.informatimago.com/
A bad day in () is better than a good day in {}.