writing a comparison the long way and getting unforeseen results

hebdemnobad · 17-Oct-2014 11:55:04

hello

I am trying to write a comparison operator the long way to read a txt document with strextr.
Here are the contents of the txt file:

name 6.3.1
update 2014-07-31
version d3da4b00

here is how I am trying to get the line whose first word is "version":

fun getversion()=

 let strextr _getpack _checkpack "strextr/strextr.txt" -> linelist in
 let sizelist linelist -> size in
 let nil-> version in
 let 0 -> i in
 (
 while i < size do
 		(
 			let nth_list linelist i -> element in
 			let hd element ->first_word in
 			(
 			if ((strcmp   "version" first_word) ==0)  then nil else
 			set version = hd tl element;	
 			set i=i+1; 
 			);
 		
 		);	
 	_showconsole; 	
 _fooS "the version is"; 
 _fooS version;

 );;

Instead of returning:

The version is
d3da4b00

in the console, the function is returning

The version is
2014-07-31

so strcmp "version" first_word returns 0 when the first word in a line is "update"
I don't know why....it's something obvious that I'm missing unfortunately.
thx for your help

iri · 17-Oct-2014 12:50:21

Hi,

First, the better way for that is :

fun getversion()=
  let strextr _getpack _checkpack "strextr/strextr.txt" -> linelist in
  let hd switchstr linelist "version" -> version in
  (
  _fooS "the version is"; 
  _fooS version;
  );;

switchstr : http://www.scolring.org/files/doc_html/switchstr.html
I use "hd switchstr ..." because i want the first word of the result only.

Another way is a recursive function (but in this case switchstr is faster)

fun searchVersion (list)=
	if list == nil then
		"not found"
	else
		let hd hd list -> firstWord in
		if (!strcmp firstWord "version") then
			hd tl hd list // if you wish an explanation tell me
		else
			searchVersion tl list;;

fun getversion()=
  let strextr _getpack _checkpack "strextr/strextr.txt" -> linelist in
  let searchVersion linelist  -> version in
  (
  _fooS "the version is"; 
  _fooS version;
  );;

Your solution is wrong (right on the principle) because when you check the string you set the variable with the wrong result.

Look at your code :

if ((strcmp   "version" first_word) ==0)  then // TRUE : first_word = "version"
  nil // it is TRUE and you do nothing !!
else  // FALSE : first_word is different than "version"
  set version = hd tl element;	// you set the variable ... So, the last "WRONG" line is "update 2014-07-31" ...

Another remark :
to compare two strings, use the logical negation "!" instead of your "== 0". The result is the same but it's more readable.

if ((strcmp   "version" first_word) ==0)  then

is the same than

if (!strcmp "version" first_word) then

Thus, your function should be :

fun getversion()=

 let strextr _getpack _checkpack "strextr/strextr.txt" -> linelist in
 let sizelist linelist -> size in
 let nil-> version in
 let 0 -> i in
 (
 while i < size do
 		(
 			let nth_list linelist i -> element in
 			let hd element ->first_word in
 			(
 			if (!strcmp "version" first_word)  then
				set version = hd tl element
 			else
				nil;
 			set i=i+1; 
 			);
 		);	
 _fooS "the version is"; 
 _fooS version;
 );;

hebdemnobad · 17-Oct-2014 13:29:21

Thx Iri! So

if ((strcmp   "version" first_word) ==0)  then nil

Causes the VM to skip the following line

set version = hd tl element;

And the VM jumps to

	
set i=i+1;

Until i == size?

iri · 17-Oct-2014 13:43:49

The VM skips this line because you wrote

else

if ((strcmp   "version" first_word) ==0)  then nil else
 			set version = hd tl element;

In words, this is :

IF the first word is "version" THEN do nothing ELSE set my variable to the second word

Thus, the variable is set ONLY if the first word is NOT "version".

So, with your text :

name 6.3.1
update 2014-07-31
version d3da4b00

At the first loop, the line is :

name 6.3.1

"name" is not "version", so the VM skips to ELSE : set version = "6.3.1" !!

At the second loop, the line becomes :

update 2014-07-31

"update" is not "version", so the VM skips to ELSE : set version = "2014-07-31" !!

At the third loop, the line is :

version d3da4b00

"version" = "version", so the VM does nothing (your nil statment). So, "set version" keep the previous value.

_fooS "the version is"; 
_fooS version;

version is "2014-07-31", so "2014-07-31" is displayed !

iri · 17-Oct-2014 13:45:12

hebdemnobad wrote:

Until i == size?

In your code, yes.

All lines are browsed even if the "good" value has been found.

hebdemnobad · 17-Oct-2014 13:55:37

thx, I think I understand (most) of your explanations. fwiw, this works too:

fun getversion()=

 let strextr _getpack _checkpack "strextr/strextr.txt" -> linelist in
 let sizelist linelist -> size in
 let nil-> version in
 let 0 -> i in
 (
 while (i < size) && (version == nil) do
 		(
 			let nth_list linelist i -> element in
 			let hd element ->first_word in
 			(
 			//	else if ((strfindi "idlex_5_24cv" message 0) != nil) then
 			if ((strcmp   "version" first_word) ==1)  then nil else
 			set version = hd tl element;	
 			set i=i+1; 
 			);
 		
 		);	
 	_showconsole; 	
 _fooS "the version is"; 
 _fooS version;

 );;

this way, if "version" is not the first word of the list of strings in a line, the vm is forced to look at the following line. once the loop finds a line where the first word is "version", it returns the remainder of the line and and the loop no longer executes since version != nil. So that is the same as writing

(!strcmp firstWord "version")

your first example though is the simplest, thx for letting me know about that function (again, someday I'll know them from memory.) this is a great help in beginning to learn file reading and prepare to parse files in .xml

iri · 17-Oct-2014 14:28:02

if ((strcmp   "version" first_word) ==1)  then

In this code, it is WRONG and DANGEROUS !

Read the doc about strcmp : http://www.scolring.org/files/doc_html/strcmp.html

strcmp can return 0, 1 or -1.

So, if you want check if a string is equal another one, use ALWAYS :

if (!strcmp anyString "a string") then
  // TRUE, they are equal
else
  // FALSE, they are NOT equal

or, if you prefer without the negation (the negation of 0 is 1, the negation of 1 is 0) :

if (0 == strcmp anyString "a string") then
  // TRUE, they are equal
else
  // FALSE, they are NOT equal

Use another way ONLY to know if a string is greater (return 1) or lesser (return -1) than another one.

hebdemnobad wrote:

your first example though is the simplest, thx for letting me know about that function (again, someday I'll know them from memory.) this is a great help in beginning to learn file reading and prepare to parse files in .xml

Yes, my first version is the best and fastest !

fun getversion()=
	_fooS hd switchstr strextr _getpack _checkpack "strextr/strextr.txt" "version";;

(it is my previous version in one line only)

To parse a xml file, you are already an API : http://redmine.scolring.org/projects/sx … psxml.html

hebdemnobad · 17-Oct-2014 17:07:11

so if (strcmp anyString "a string") is 1 or -1, both are the same as: (!strcmp anyString "a string")

iri · 17-Oct-2014 17:54:43

_fooId strcmp "name" "name" ==>> 0 because "name" is the same string than "name". Right ?

_fooId strcmp "name" "version" ==>> -1 because "name" is lesser than "version" (in the alphabetic order, n is lesser than v). Right ?

_fooId strcmp "version" "name" ==>> 1 because "version" is greater than "name" (in the alphabetic order, v is greater than n). Right ?

_fooId !0 ==>> 1 because the negation ( ! exclamation mark) of 0 is 1

_fooId !1 ==>> 0 because the negation of all integer different than 0 is 1

So :
strcmp "name" "name" ==> 0 so, the negation of this expression i.e. !strcmp "name" "name" (exclamation mark following by strcmp "name" "name") is 1. Right ?

When i write :

if A then // blabla else // blabla

implicitely, i wrote : if A == 1 then // blabla else // blabla

Right ?

When i write :

if !A then // blabla else // blabla

implicitely, i wrote :

if A == 0 then // blabla else // blabla

Thus, when i write :

if !strcmp anyString "a string" then // EQUAL else // NOT EQUAL

I'm certain that anyString is strictly equal at "a string" because :
- the negation of strcmp anyString "a string" is 1 so anyString "a string" returns 0.

The doc is clear : strcmp returns 0 only if the two strings are equals.

hebdemnobad wrote:

so if (strcmp anyString "a string") is 1 or -1, both are the same as: (!strcmp anyString "a string")

So, wrong.

hebdemnobad · 17-Oct-2014 21:07:27

Thx for the detailed reply....at the moment I feel a bit but I think I'll figure it out.

Last edited by hebdemnobad (17-Oct-2014 22:26:01)

hebdemnobad · 17-Oct-2014 23:35:30

I think I get it....I had forgotten that strcmp returns 0 when the strings are equal but had thought they return 1 when they are equal, so I had it backwards, what I should have said was. in my coding I've usually used 1 for true and 0 for false in setting interger flags in javascript.

so...

if (strcmp anyString "a string") is 1 or -1, both are the same as: (strcmp anyString "a string")

that is correct, no?

arkeon · 18-Oct-2014 01:24:21

to simplify just take the habbit to use
(!strcmp A B) to check if a string is equal
and (strcmp A B) to check if they are different

hebdemnobad · 18-Oct-2014 02:04:23

It's all JavaScript's fault..http://stackoverflow.com/questions/12236271/is-true-1-and-false-0-in-javascript.

arkeon · 18-Oct-2014 11:22:15

no this is the same in scol, this is just that since strcmp do not return a boolean value but a size difference

hebdemnobad · 18-Oct-2014 12:10:27

you are right I was confusing string comparison with Boolean values in javascript where true == 1 and false == 0.

Last edited by hebdemnobad (18-Oct-2014 12:10:59)

arkeon · 18-Oct-2014 12:12:08

yes strcmp is different of ==
== on string in scol compare the string pointer so it's different of string comparison

iri · 19-Oct-2014 11:01:42

== : in memory

Scolring - Forum

#1 17-Oct-2014 11:55:04

writing a comparison the long way and getting unforeseen results

#2 17-Oct-2014 12:50:21

Re: writing a comparison the long way and getting unforeseen results

#3 17-Oct-2014 13:29:21

Re: writing a comparison the long way and getting unforeseen results

#4 17-Oct-2014 13:43:49

Re: writing a comparison the long way and getting unforeseen results

#5 17-Oct-2014 13:45:12

Re: writing a comparison the long way and getting unforeseen results

#6 17-Oct-2014 13:55:37

Re: writing a comparison the long way and getting unforeseen results

#7 17-Oct-2014 14:28:02

Re: writing a comparison the long way and getting unforeseen results

#8 17-Oct-2014 17:07:11

Re: writing a comparison the long way and getting unforeseen results

#9 17-Oct-2014 17:54:43

Re: writing a comparison the long way and getting unforeseen results

#10 17-Oct-2014 21:07:27

Re: writing a comparison the long way and getting unforeseen results

#11 17-Oct-2014 23:35:30

Re: writing a comparison the long way and getting unforeseen results

#12 18-Oct-2014 01:24:21

Re: writing a comparison the long way and getting unforeseen results

#13 18-Oct-2014 02:04:23

Re: writing a comparison the long way and getting unforeseen results

#14 18-Oct-2014 11:22:15

Re: writing a comparison the long way and getting unforeseen results

#15 18-Oct-2014 12:10:27

Re: writing a comparison the long way and getting unforeseen results

#16 18-Oct-2014 12:12:08

Re: writing a comparison the long way and getting unforeseen results

#17 19-Oct-2014 11:01:42

Re: writing a comparison the long way and getting unforeseen results

Board footer