A popping question

[mdh]

Could someone help me understand one aspect of an exercise in K&R II;
(page 76)

Given function:

double pop( void){

....return a double from a stack if present

}

how does this work?

/* adds 2 doubles*/

case '+' :

push ( pop() + pop()) ;

I would have thought you would have to do something like

double firstpop, secondpop;

firstpop=pop();
secondpop=pop();
push ( firstpop+secondpop);

ie where do the doubles go if they are not assigned. It obviously goes
somewhere and it does work, it's just that I do not see how.

Thanks in advance.

[Ben Pfaff]

"mdh" <md**@comcast.netwrites:

how does this work?

....

push ( pop() + pop()) ;

I would have thought you would have to do something like

double firstpop, secondpop;

firstpop=pop();
secondpop=pop();
push ( firstpop+secondpop);

ie where do the doubles go if they are not assigned. It obviously goes
somewhere and it does work, it's just that I do not see how.

It doesn't matter. It's the compiler's job to reserve space for
temporaries used in expressions. The programmer doesn't have to
worry about it.
--
"Programmers have the right to be ignorant of many details of your code
and still make reasonable changes."
--Kernighan and Plauger, _Software Tools_

[mdh]

"mdh" <md**@comcast.netwrites:

ie where do the doubles go if they are not assigned. It obviously goes
somewhere and it does work, it's just that I do not see how.

Ben Pfaff wrote:

It doesn't matter. It's the compiler's job to reserve space for
temporaries used in expressions. The programmer doesn't have to
worry about it.

So is it fair to assume that "pop" is in a sense replaced by it's
value, and any assignment is "additonal" but not necessary to the
execution of "pop"?

[Ben Pfaff]

"mdh" <md**@comcast.netwrites:

>"mdh" <md**@comcast.netwrites:

ie where do the doubles go if they are not assigned. It obviously goes
somewhere and it does work, it's just that I do not see how.

Ben Pfaff wrote:

>It doesn't matter. It's the compiler's job to reserve space for
temporaries used in expressions. The programmer doesn't have to
worry about it.

So is it fair to assume that "pop" is in a sense replaced by it's
value, and any assignment is "additonal" but not necessary to the
execution of "pop"?

Do you have an idea that the result of a function must be
immediately assigned to a variable, or that otherwise it is lost?
That is how I interpret your words. This idea is wrong. All of
the following are allowed in C, given an appropriate function foo
and variable x:
foo(); /* Throws away return value. */
x = foo(); /* Assigns return value. */
x = foo() + 1; /* Assigns return value plus 1. */
x = foo() + foo(); /* Calls foo twice, assigns the sum. */
--
int main(void){char p[]="ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuv wxyz.\
\n",*q="kl BIcNBFr.NKEzjwCIxNJC";int i=sizeof p/2;char *strchr();int putchar(\
);while(*q){i+=strchr(p,*q++)-p;if(i>=(int)sizeof p)i-=sizeof p-1;putchar(p[i]\
);}return 0;}

[mdh]

Ben Pfaff wrote:
Do you have an idea that the result of a function must be

immediately assigned to a variable, or that otherwise it is lost?
That is how I interpret your words. This idea is wrong.

no....I did not have an fixed idea...although it may have come across
that way...but your example clarifies it for me.

thanks.

[CBFalconer]

Ben Pfaff wrote:

"mdh" <md**@comcast.netwrites:

>how does this work?

...

>push ( pop() + pop()) ;

I would have thought you would have to do something like

double firstpop, secondpop;

firstpop=pop();
secondpop=pop();
push ( firstpop+secondpop);

ie where do the doubles go if they are not assigned. It obviously goes
somewhere and it does work, it's just that I do not see how.

It doesn't matter. It's the compiler's job to reserve space for
temporaries used in expressions. The programmer doesn't have to
worry about it.

Fine for + and *. Change it to - or / and things are different.

--
Some informative links:
news:news.announce.newusers
http://www.geocities.com/nnqweb/
http://www.catb.org/~esr/faqs/smart-questions.html
http://www.caliburn.nl/topposting.html
http://www.netmeister.org/news/learn2quote.html

--
Posted via a free Usenet account from http://www.teranews.com

[Ben Pfaff]

CBFalconer <cb********@yahoo.comwrites:

Ben Pfaff wrote:

>"mdh" <md**@comcast.netwrites:

>>how does this work?

...

>>push ( pop() + pop()) ;

I would have thought you would have to do something like

double firstpop, secondpop;

firstpop=pop();
secondpop=pop();
push ( firstpop+secondpop);

ie where do the doubles go if they are not assigned. It obviously goes
somewhere and it does work, it's just that I do not see how.

It doesn't matter. It's the compiler's job to reserve space for
temporaries used in expressions. The programmer doesn't have to
worry about it.

Fine for + and *. Change it to - or / and things are different.

The OP's question was about memory. You are talking about order
of evaluation. The K&R2 page in question mentions that - and /
will not (reliably) do what we want for order of evaluation, so
there was no reason for me or the OP to bring it up.
--
Just another C hacker.

[mdh]

Ben Pfaff wrote:

>
The OP's question was about memory. You are talking about order
of evaluation. The K&R2 page in question mentions that - and /
will not (reliably) do what we want for order of evaluation, so
there was no reason for me or the OP to bring it up.

Hi Ben,
Correct....I probably did think that any value had to be assigned hence
the question. I think, even in C, one has to take certain things as
"given" as your example clarified. And yes...I am aware of the order of
things.

[Barry Schwarz]

On 21 Sep 2006 15:36:21 -0700, "mdh" <md**@comcast.netwrote:

>
Ben Pfaff wrote:

>>
The OP's question was about memory. You are talking about order
of evaluation. The K&R2 page in question mentions that - and /
will not (reliably) do what we want for order of evaluation, so
there was no reason for me or the OP to bring it up.

Hi Ben,
Correct....I probably did think that any value had to be assigned hence
the question. I think, even in C, one has to take certain things as
"given" as your example clarified. And yes...I am aware of the order of
things.

I hope you meant to say "the lack of order".
Remove del for email

[mdh]

Ben Pfaff wrote:

>
The OP's question was about memory. You are talking about order
of evaluation. The K&R2 page in question mentions that - and /
will not (reliably) do what we want for order of evaluation, so
there was no reason for me or the OP to bring it up.

Hi Ben,
Correct....I probably did think that any value had to be assigned hence
the question. I think, even in C, one has to take certain things as
"given" as your example clarified. And yes...I am aware of the order of
things.

Barry Schwarz wrote:

I hope you meant to say "the lack of order".

Clearly carrying some deeper meaning, deeper than my knowledge. Care
to explain?

[Barry Schwarz]

On 21 Sep 2006 18:46:39 -0700, "mdh" <md**@comcast.netwrote:

>Correct....I probably did think that any value had to be assigned hence
the question. I think, even in C, one has to take certain things as
"given" as your example clarified. And yes...I am aware of the order of
things.

Barry Schwarz wrote:

>I hope you meant to say "the lack of order".

Clearly carrying some deeper meaning, deeper than my knowledge. Care
to explain?

The order of evaluation is at best implementation specified.
Consequently, for a non commutative operator such as - or /, you don't
know which operand will be evaluated first. In the code in question,
the operand was the result of a call to a stack popping function.
pop_1 - pop_2 produces a different result than pop_2 - pop_1. You
claimed to be aware of this but your terminology raised a doubt in my
mind.
Remove del for email

[mdh]

Barry Schwarz wrote:

The order of evaluation is at best implementation specified.
Consequently, for a non commutative operator such as - or /, you don't
know which operand will be evaluated first. In the code in question,
the operand was the result of a call to a stack popping function.
pop_1 - pop_2 produces a different result than pop_2 - pop_1. You
claimed to be aware of this but your terminology raised a doubt in my
mind.

And I see by my reply how this could have happened. My initial question
related to the issue of how "it worked" when and expression like

push( pop()+pop()) was evaluated.( I chose that example to avoid the
exact situation you showed.)

The answer I got explained that in the C language, it is permissible
to call a function and
--not doing anything with the result
--assign it to a variable
--assign the result to another function...which is still a little
puzzling to me..although I will have to accept the fact that the
compiler simply deals with it and it "just works".

[suresh]

mdh wrote:

The answer I got explained that in the C language, it is permissible
to call a function and
--not doing anything with the result
--assign it to a variable
--assign the result to another function...which is still a little
puzzling to me..although I will have to accept the fact that the
compiler simply deals with it and it "just works".

Note that the following is valid in C.

int *func1();
int func2();

*func1() = func2();

But I don't think this is what you meant by "assigning the result to
another function".
..

[Nishu]

mdh wrote:

>

-snip-

push( pop()+pop()) was evaluated.( I chose that example to avoid the
exact situation you showed.)

The answer I got explained that in the C language, it is permissible
to call a function and
--not doing anything with the result
--assign it to a variable
--assign the result to another function...which is still a little
puzzling to me..although I will have to accept the fact that the
compiler simply deals with it and it "just works".

It works because this is how you have defined your function prototype.
If pop returns int and push accepts int argument, it makes lots of
sense to use it in this manner. It removes unnecessary declarations of
int type temp variables and other assignment statements.

-N

[mdh]

mdh wrote:

push( pop()+pop()) was evaluated.( I chose that example to avoid the

exact situation you showed.)

The answer I got explained that in the C language, it is permissible
to call a function and
--not doing anything with the result
--assign it to a variable
--assign the result to another function...which is still a little
puzzling to me..although I will have to accept the fact that the
compiler simply deals with it and it "just works".

Nishu wrote:

It works because this is how you have defined your function prototype.
If pop returns int and push accepts int argument, it makes lots of
sense to use it in this manner. It removes unnecessary declarations of
int type temp variables and other assignment statements.

I guess what I have been asking is this?

function:

push( pop()+pop());

If you are an experienced C programmer, how do you comprehend this?

Do you say...

1) OK...that's defined as such ( as Ben described--below)

All of the following are allowed in C, given an appropriate function
foo
and variable x:
foo(); /* Throws away return value. */
x = foo(); /* Assigns return value. */
x = foo() + 1; /* Assigns return value plus 1. */
x = foo() + foo(); /* Calls foo twice*/

To me as a C programmer, that's all I worry about...I leave the rest to
the compiler and know it will work

or

2) I wonder what happens to the value of "pop()" before it is
"pushed()" to the stack. Is there a working stack that the "pops" are
pushed onto, then added then pushed back ....or is this merely someone
like me trying understand something that should not or cannot easily
be known...or is dependent upon the flavor of C being used hence (1) is
what I am concerned about.

Well, I hope I am not driving people nuts...but, thats the way I try to
comprehend things.

Thanks for all your patience.

And I am slowly seeing why definitions in C are so important!!!! :-)

[Simon Biber]

Barry Schwarz wrote:

The order of evaluation is at best implementation specified.
Consequently, for a non commutative operator such as - or /, you don't
know which operand will be evaluated first. In the code in question,
the operand was the result of a call to a stack popping function.
pop_1 - pop_2 produces a different result than pop_2 - pop_1. You
claimed to be aware of this but your terminology raised a doubt in my
mind.

<OT>

You'd be best not to use "doubt" in its original meaning any more, since
apparently over a billion "English" speakers have now replaced its
meaning by that of "question".

</OT>

--
Simon.

[Barry Schwarz]

On 21 Sep 2006 23:59:48 -0700, "mdh" <md**@comcast.netwrote:

>

>mdh wrote:

push( pop()+pop()) was evaluated.( I chose that example to avoid the

exact situation you showed.)

The answer I got explained that in the C language, it is permissible
to call a function and
--not doing anything with the result
--assign it to a variable
--assign the result to another function...which is still a little
puzzling to me..although I will have to accept the fact that the
compiler simply deals with it and it "just works".

Nishu wrote:

>It works because this is how you have defined your function prototype.
If pop returns int and push accepts int argument, it makes lots of
sense to use it in this manner. It removes unnecessary declarations of
int type temp variables and other assignment statements.

I guess what I have been asking is this?

function:

push( pop()+pop());

If you are an experienced C programmer, how do you comprehend this?

Do you say...

1) OK...that's defined as such ( as Ben described--below)

All of the following are allowed in C, given an appropriate function
foo
and variable x:
foo(); /* Throws away return value. */
x = foo(); /* Assigns return value. */
x = foo() + 1; /* Assigns return value plus 1. */
x = foo() + foo(); /* Calls foo twice*/

And uses the result of the two calls as the operands of the + operator
which will then attempt to compute the sum and assign that value to x.

Alternately, the + operator takes to operands. Each operand is an
expression. The two expressions must be evaluated before the +
operator can do its thing. Whichever operand is evaluated first
happens to result in a call to foo. In the course of this evaluation,
the result returned by foo replaces the call as the expression is
evaluated. Ditto for the other operand which is evaluated second.

Consider
x = foo() + (foo() * 2);
When the second + operand is evaluated, the * expression requires both
foo() and 2 to be evaluated and then their product formed. We don't
know whether this call to foo is the first or second because we don't
know the order in which the + operands are evaluated.

>
To me as a C programmer, that's all I worry about...I leave the rest to
the compiler and know it will work

or

2) I wonder what happens to the value of "pop()" before it is
"pushed()" to the stack. Is there a working stack that the "pops" are
pushed onto, then added then pushed back ....or is this merely someone
like me trying understand something that should not or cannot easily
be known...or is dependent upon the flavor of C being used hence (1) is
what I am concerned about.

pop() was a function you introduced to start the discussion. From the
code you've shown, there is no push(). How the compiler chooses to
handle intermediate results while evaluating expressions is an
implementation detail that will vary from system to system. My system
doesn't have a hardware stack and intermediate results are stored in
temporary variables which the compiler keeps track of (and attempts to
reuse efficiently).
Remove del for email

[Barry Schwarz]

On Fri, 22 Sep 2006 14:01:54 GMT, Simon Biber <ne**@ralmin.ccwrote:

>Barry Schwarz wrote:

>The order of evaluation is at best implementation specified.
Consequently, for a non commutative operator such as - or /, you don't
know which operand will be evaluated first. In the code in question,
the operand was the result of a call to a stack popping function.
pop_1 - pop_2 produces a different result than pop_2 - pop_1. You
claimed to be aware of this but your terminology raised a doubt in my
mind.

<OT>

You'd be best not to use "doubt" in its original meaning any more, since
apparently over a billion "English" speakers have now replaced its
meaning by that of "question".

</OT>

Substituting disbelief, uncertainty, or question for doubt in this
sentence would still get the same point across.
Remove del for email

[Keith Thompson]

Barry Schwarz <sc******@doezl.netwrites:
[...]

pop() was a function you introduced to start the discussion. From the
code you've shown, there is no push(). How the compiler chooses to
handle intermediate results while evaluating expressions is an
implementation detail that will vary from system to system. My system
doesn't have a hardware stack and intermediate results are stored in
temporary variables which the compiler keeps track of (and attempts to
reuse efficiently).

Out of curiosity, what system are you using that doesn't have a
hardware stack? It would be good to have a specific example in the
recurring argument about whether C defines a "stack".

--
Keith Thompson (The_Other_Keith) ks***@mib.org <http://www.ghoti.net/~kst>
San Diego Supercomputer Center <* <http://users.sdsc.edu/~kst>
We must do something. This is something. Therefore, we must do this.

[mdh]

mdh wrote:

I guess what I have been asking is this?

If you are an experienced C programmer, how do you comprehend this?

Do you say...

1) OK...that's defined as such ( as Ben described--below)

or

2) I wonder what happens to the value of "pop()" before it is
"pushed()" to the stack.

And I am slowly seeing why definitions in C are so important!!!! :-)

Barry Schwarz wrote:
pop() was a function you introduced to start the discussion. How the
compiler chooses to
handle intermediate results while evaluating expressions is an
implementation detail that

>>>>>>>

Mdh replies:
Thanks for our reply. That helps clear it up for me.

[Barry Schwarz]

On Sat, 23 Sep 2006 00:10:26 GMT, Keith Thompson <ks***@mib.org>
wrote:

>Barry Schwarz <sc******@doezl.netwrites:
[...]

>pop() was a function you introduced to start the discussion. From the
code you've shown, there is no push(). How the compiler chooses to
handle intermediate results while evaluating expressions is an
implementation detail that will vary from system to system. My system
doesn't have a hardware stack and intermediate results are stored in
temporary variables which the compiler keeps track of (and attempts to
reuse efficiently).

Out of curiosity, what system are you using that doesn't have a
hardware stack? It would be good to have a specific example in the
recurring argument about whether C defines a "stack".

An IBM Multiprise 2003 S/390 running OS/390 2.10.
Remove del for email