Stopping a while loop with user input ?

E

ern

I have a program that runs scripts. If the user types "script
myScript.dat" the program will grab commands from the text file, verify
correctness, and begin executing the script UNTIL...

I need a way to stop the execution with user input. I was thinking
something like this:

while(user hasn't pressed 'any key'){
keepExecutingScript();
}

How can I do the "user hasn't pressed 'any key' " of this logic?
Running Linux Red Hat V3.

Thanks,
 
M

Michael Mair

ern said:
I have a program that runs scripts. If the user types "script
myScript.dat" the program will grab commands from the text file, verify
correctness, and begin executing the script UNTIL...

I need a way to stop the execution with user input. I was thinking
something like this:

while(user hasn't pressed 'any key'){
keepExecutingScript();
}

How can I do the "user hasn't pressed 'any key' " of this logic?
Running Linux Red Hat V3.

This cannot done in standard C; try comp.unix.programmer

Cheers
Michael
 
D

Dave Vandervies

I have a program that runs scripts. If the user types "script
myScript.dat" the program will grab commands from the text file, verify
correctness, and begin executing the script UNTIL...

I need a way to stop the execution with user input. I was thinking
something like this:

while(user hasn't pressed 'any key'){
keepExecutingScript();
}

If you're allowed to choose which key the user uses to interrupt the
program, there's a good chance you can do this without going beyond what
the standard defines.
(This may or may not be something you'd want to do instead of using
something system-specific instead.)

Most implementations have a way to cause SIGINT to be raised in a running
program, so you can just install a handler that sets a flag indicating
it's time to stop, and check that in your main loop.
(Note that the way you arrange for SIGINT to be raised asynchronously
is entirely dependent on your system and there may not be a way at all.
(All of the systems that used nontrivially use ctrl-C by default, but
check your documentation.) So you're not actually gaining portability
to systems that don't support user-caused SIGINT, you're just gaining
not having to change the code to move between systems that do.)

--------
#include <signal.h>

volatile sig_atomic_t time_to_stop=0;

void sigint_handler(int ignored)
{
time_to_stop=1;
/*Revert to the default signal handler (usually immediate
termination) for if the user gets impatient and interrupts
again
*/
signal(SIGINT,SIG_DFL);
}

void do_main_loop(some_args args)
{
void (*old_sighandler)(int);

old_sighandler=signal(SIGINT,sigint_handler);
if(old_sighandler==SIG_IGN)
{
/*Assume it was ignored for a good reason, re-ignore*/
signal(SIGINT,SIG_IGN);
}
if(old_sighandler==SIG_ERR)
{
report_signal_installation_error();
abort_if_we_dont_want_to_continue_uninterruptible();
}

while(time_to_stop==0)
{
keep_executing_script(args);
}
do_any_cleanup_required();

/*Depending on what comes next, we may want to restore the original
handler:
*/
if(old_sighandler!=SIG_ERR)
signal(SIGINT,old_sighandler);
}
--------

Note that the things you can do in a handler for an asynchronous signal
are seriously limited. Storing a value in an object of type volatile
sig_atomic_t is guaranteed to do what you'd expect (store the value
completely and correctly and become available the next time it's checked
outside the signal handler), and using signal() to install a new handler
for the signal you're handling is also well-defined. Calling most library
functions isn't safe, examining and/or modifying most objects isn't safe,
and there's really not a whole lot that you can do without trying to do
one or the other of those.


dave
 
M

Mabden

ern said:
I have a program that runs scripts. If the user types "script
myScript.dat" the program will grab commands from the text file, verify
correctness, and begin executing the script UNTIL...

I need a way to stop the execution with user input. I was thinking
something like this:

while(user hasn't pressed 'any key'){
keepExecutingScript();
}

How can I do the "user hasn't pressed 'any key' " of this logic?
Running Linux Red Hat V3.

getch() might work.
 
J

Jordan Abel

getch() might work.

however, you need curses for that. The only "portable" way to do
this would be to use SIGINT and limit the user's ability to stop
the "script" to causing SIGINT (in an implementation-defined way:
on any unix system this is controlled by the tty driver and is
caused by default by hitting ctrl-c - it's not unreasonable to
use ctrl-c and/or ctrl-break on windows/dos, etc.)

your signal handler could set a global variable that is then
checked in the while loop.

it's also a lot easier than using curses, and a lot more in
keeping with how other unix-ish applications do things.
 
J

Jordan Abel

Curses? Unixish? what's that? My 'puter speaks DOS. My compiler knows
getch() and so does my bible, the K&R.

there is a getch() that exists on DOS and one that exists on unix.
they do not have the same semantics. K&R uses the name, but it is
for neither of the two commonly-existing functions of that name, but
rather of its own example of how one might write a getc-ish function
that could play nice with an ungetc-ish one. Moreover, the getch in
K&R, the one in unix/curses under some circumstances, and the dos
one, will ALL block until the user actually presses a key - none of
them [except the curses one under SOME sets of options] will return
immediately if there is no key pressed by the user.
Can't he read a line of script and then check if a key was hit? I mean,
the one line of the script may take a little while to finish, but at
least the whole script doesn't have to finish before the user gets
control.

he _can_ do it, using the curses getch()
SIGINT is fine, but seems overkill for processing a script, since you
can just poll after every line. I use SIGINT for trapping Ctrl-C to stop
something lengthy like parsing a directory tree, but a script file of
presumably short-lived commands could be halted by just polling the
keyboard buffer with getch(). Unless the script was something like:

the problem is that getch() will not return if there is _not_ a
character. [well, you can do it with curses, but it's tricky]
 
M

Mabden

Jordan Abel said:
however, you need curses for that. The only "portable" way to do
this would be to use SIGINT and limit the user's ability to stop
the "script" to causing SIGINT (in an implementation-defined way:
on any unix system this is controlled by the tty driver and is
caused by default by hitting ctrl-c - it's not unreasonable to
use ctrl-c and/or ctrl-break on windows/dos, etc.)

your signal handler could set a global variable that is then
checked in the while loop.

it's also a lot easier than using curses, and a lot more in
keeping with how other unix-ish applications do things.

Curses? Unixish? what's that? My 'puter speaks DOS. My compiler knows
getch() and so does my bible, the K&R.

Can't he read a line of script and then check if a key was hit? I mean,
the one line of the script may take a little while to finish, but at
least the whole script doesn't have to finish before the user gets
control.

SIGINT is fine, but seems overkill for processing a script, since you
can just poll after every line. I use SIGINT for trapping Ctrl-C to stop
something lengthy like parsing a directory tree, but a script file of
presumably short-lived commands could be halted by just polling the
keyboard buffer with getch(). Unless the script was something like:

1 cure cancer
2 cure Michael Jackson
3 fix Madben
 
M

Mabden

Jordan Abel said:
the problem is that getch() will not return if there is _not_ a
character. [well, you can do it with curses, but it's tricky]

Oops.. so That's why I use SIGINT... I looked back at some DOS programs
and there were more of them than I remembered. But they all seemed to be
for Ctrl-C handling.

Did the OP say they wanted to trap "The Any Key" or a specific key? Does
SIGINT flag, say a space bar? I forget, since I used it to exit
gracefully from a program when ctrl-c or ctrl-break was hit. What about
the old ctrl-d ctrl-s combo for pausing and restarting a program, maybe
that might work for the OP.

I guess I need to read the first post. ;-)
 
J

Joe Wright

Mabden said:
getch() might work.
Just when I thought you were learning.. getch() simply does not exist in
Standard C, the topic of this newsgroup. Various implementations may
provide it (an others) as extensions but it isn't C.
 
M

Mabden

Joe Wright said:
Just when I thought you were learning.. getch() simply does not exist in
Standard C, the topic of this newsgroup. Various implementations may
provide it (an others) as extensions but it isn't C.

Awww.. how sweet, you thought...

getch() is on page 78-79 of the only C standard that matters, K&R2.

I've never compiled a program using getch() and had it fail. But as I
said, I was surprized to find that I actually used SIGINT more than I
thought, and getch() less than you think.

So, while you are not wrong - you're wrong! ;-)
 
J

Jordan Abel

Awww.. how sweet, you thought...

getch() is on page 78-79 of the only C standard that matters,
K&R2.

I've never compiled a program using getch() and had it fail. But
as I said, I was surprized to find that I actually used SIGINT
more than I thought, and getch() less than you think.

So, while you are not wrong - you're wrong! ;-)

You are incorrect. K&R2 does NOT claim getch() exists or a standard
function. It uses the name in an example, which is in fact an
implicit claim that it is NOT a standard function
 
F

Flash Gordon

Mabden said:
Awww.. how sweet, you thought...

getch() is on page 78-79 of the only C standard that matters, K&R2.

Where it provides an implementation of getch which has different
semantics to the implementation I have on my Linux box and different
semantics to the implementation provided my MS in their C library. If
you look at Appendix B which describes the standard library you will
find absolutely NO reference to getch.

Also, it is the ISO standard that defines C and therefore what is
topical here, K&R2 is merely a very good, but not perfect, book about
the language.
I've never compiled a program using getch() and had it fail.

Which proves nothing apart from your experience being limited.
> But as I
said, I was surprized to find that I actually used SIGINT more than I
thought, and getch() less than you think.

So, while you are not wrong - you're wrong! ;-)

No, Joe is COMPLETELY correct and you are COMPLETELY wrong.

Since you are obviously not improving I'll put you back in the kill file.
 
P

pete

Jordan said:
K&R2 does NOT claim getch() exists or a standard function.

You are correct.
It uses the name in an example, which is in fact an
implicit claim that it is NOT a standard function

You would think so, but there is a counterexample in K&R.

http://cm.bell-labs.com/cm/cs/cbook/2ediffs.html

119-121(§5.11): The qsort discussion needs recasting in several ways.
First, qsort is a standard routine in ANSI/ISO C, so
the rendition here should be given a different name, especially because
the arguments to standard qsort are a bit different: the
standard accepts a base pointer and a count, while this example uses a
base pointer and two offsets.
 
K

Keith Thompson

pete said:
You are correct.


You would think so, but there is a counterexample in K&R.

http://cm.bell-labs.com/cm/cs/cbook/2ediffs.html

119-121(§5.11): The qsort discussion needs recasting in several ways.
First, qsort is a standard routine in ANSI/ISO C, so
the rendition here should be given a different name, especially because
the arguments to standard qsort are a bit different: the
standard accepts a base pointer and a count, while this example uses a
base pointer and two offsets.

But using the name "qsort" in an example *is* an implicit claim that
it's not a standard function. It happens to be an incorrect claim,
which is why it's mentioned in the errata.
 
M

Mabden

Keith Thompson said:
But using the name "qsort" in an example *is* an implicit claim that
it's not a standard function. It happens to be an incorrect claim,
which is why it's mentioned in the errata.

So none of the programs named in K&R (2 in my version) are actually in
The Standard? Is that canonical? Have you checked this yourself, or is
it hearsay?
 
J

Jordan Abel

So none of the programs named in K&R (2 in my version) are actually in
The Standard? Is that canonical? Have you checked this yourself, or is
it hearsay?

No program which contains a definition of a standard function is
permitted on a conforming hosted implementation. With the sole (i
believe) exceptions of the qsort function and the unix-specific
examples in chapter 8, all the examples in K&R conform to the
standard.
 
M

Mark McIntyre

No program which contains a definition of a standard function is
permitted on a conforming hosted implementation.

I believe thats incorrect. My understanding is that the names are only
reserved if you include the header.

7.1.3 Reserved Identifiers
- Each identifier with file scope listed in any of the following
subclauses (including the future library directions) is reserved for
use as a macro name and as an identifier with file scope in the same
name space if any of its associated headers is included.
 
F

Flash Gordon

Mark said:
I believe thats incorrect. My understanding is that the names are only
reserved if you include the header.

7.1.3 Reserved Identifiers
- Each identifier with file scope listed in any of the following
subclauses (including the future library directions) is reserved for
use as a macro name and as an identifier with file scope in the same
name space if any of its associated headers is included.

In n1124, section 7.13, it also says:
— All identifiers with external linkage in any of the following
subclauses (including the future library directions) are always
reserved for use as identifiers with external linkage.157)

So you can never use the names for functions unless you don't include
the relevant header *and* you declare the function as static.
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Members online

No members online now.

Forum statistics

Threads
473,744
Messages
2,569,484
Members
44,903
Latest member
orderPeak8CBDGummies

Latest Threads

Top