reloadEFS - who understands this behavior?

[ackmann]

Hello cracks,

I know that calling reloadEFS() in main could create a loop
but I don´t understand in the lower example
why it does create a loop after pressing "reload"-button with the variable nLoop=4
but doesn´t create a loop with nLoop=1.

Please could anyone clarify?

function preMain() {
     setPriceStudy(true);
     setShowCursorLabel(false); 
}

debugClear(); 
var nLoop   = 0; 
var bReload = false; 

function main() {     
     trigger();     // sets bReload or not
     if (bReload) {
         bReload     = false;
         nLoop++;
         debugPrintln("  Reloaded Chart  ,     nLoop: "+nLoop);
         if (nLoop==1 || nLoop==4)       reloadEFS();     // !!!!!!!!  works with nLoop==1 but not with nLoop==4
         else {    // Reload-Button
             drawTextRelative(1, 15, " Reload @URL=EFS:reloadexec", null, null, Text.BUTTON | 
                              Text.RELATIVETOLEFT | Text.RELATIVETOBOTTOM, null, 10, "Reload");
         }
     }
     return close(); 
}  

function trigger() {
     var nState = getBarState();
     var nIndex = getCurrentBarIndex();
     if (nIndex == 0 && (nState == BARSTATE_NEWBAR || nState == BARSTATE_ALLBARS)) {
         bReload = true;
     } 
     return;
} 

function reloadexec() {    // Button execute
     debugPrintln("*** Reload ***");
     reloadEFS();
     return; 
} 

[ACM]

Re: reloadEFS - who understands this behavior?

ackmann
Actually it will also not work when nLoop==1 depending on the initial value you assign to nLoop when you declare it
To verify this try setting var nLoop=-1 and you will see that at nLoop==1 you will get the same "nested reloadEFS..." message that you get at nLoop==4 when you set var nLoop=0
This is because when you first load the script it initially reloads the formula at nLoop==1 using only the reloadEFS() command in the main function since you are incrementing the nLoop counter prior to checking its state. Notice in fact that when you first load the script the nLoop counter will display 2. When you subsequently click the Reload button and the nLoop counter is at 3 you are executing both the reloadEFS in the reloadexec and main functions thereby triggering an infinite loop condition that the efs engine recognizes and prevents (see also Steve Hare's explanation in this thread on a similar topic)
To resolve this try replacing reloadEFS() in the reloadexec function with main(). Once you do this you will see that at nLoop==1 and at nLoop==4 the formula executes 2 reloads without however triggering the nested reloadEFS condition
Alex

Originally posted by ackmann
Hello cracks,

I know that calling reloadEFS() in main could create a loop
but I don´t understand in the lower example
why it does create a loop after pressing "reload"-button with the variable nLoop=4
but doesn´t create a loop with nLoop=1.

Please could anyone clarify?

function preMain() {
     setPriceStudy(true);
     setShowCursorLabel(false); 
}

debugClear(); 
var nLoop   = 0; 
var bReload = false; 

function main() {     
     trigger();     // sets bReload or not
     if (bReload) {
         bReload     = false;
         nLoop++;
         debugPrintln("  Reloaded Chart  ,     nLoop: "+nLoop);
         if (nLoop==1 || nLoop==4)       reloadEFS();     // !!!!!!!!  works with nLoop==1 but not with nLoop==4
         else {    // Reload-Button
             drawTextRelative(1, 15, " Reload @URL=EFS:reloadexec", null, null, Text.BUTTON | 
                              Text.RELATIVETOLEFT | Text.RELATIVETOBOTTOM, null, 10, "Reload");
         }
     }
     return close(); 
}  

function trigger() {
     var nState = getBarState();
     var nIndex = getCurrentBarIndex();
     if (nIndex == 0 && (nState == BARSTATE_NEWBAR || nState == BARSTATE_ALLBARS)) {
         bReload = true;
     } 
     return;
} 

function reloadexec() {    // Button execute
     debugPrintln("*** Reload ***");
     reloadEFS();
     return; 
} 

Re: reloadEFS - who understands this behavior?

cracks?

Your question in the post seems to contradict the embedded statement in your efs. One refers to creating a loop when nLoop=4 and the other refers to it working with nLoop==1.

Try starting off with nLoop set to '-1' instead of zero, and try using more debug statements is my advice.

Originally posted by ackmann
Hello cracks,

I know that calling reloadEFS() in main could create a loop
but I don´t understand in the lower example
why it does create a loop after pressing "reload"-button with the variable nLoop=4
but doesn´t create a loop with nLoop=1.

Please could anyone clarify?

function preMain() {
     setPriceStudy(true);
     setShowCursorLabel(false); 
}

debugClear(); 
var nLoop   = 0; 
var bReload = false; 

function main() {     
     trigger();     // sets bReload or not
     if (bReload) {
         bReload     = false;
         nLoop++;
         debugPrintln("  Reloaded Chart  ,     nLoop: "+nLoop);
         if (nLoop==1 || nLoop==4)       reloadEFS();     // !!!!!!!!  works with nLoop==1 but not with nLoop==4
         else {    // Reload-Button
             drawTextRelative(1, 15, " Reload @URL=EFS:reloadexec", null, null, Text.BUTTON | 
                              Text.RELATIVETOLEFT | Text.RELATIVETOBOTTOM, null, 10, "Reload");
         }
     }
     return close(); 
}  

function trigger() {
     var nState = getBarState();
     var nIndex = getCurrentBarIndex();
     if (nIndex == 0 && (nState == BARSTATE_NEWBAR || nState == BARSTATE_ALLBARS)) {
         bReload = true;
     } 
     return;
} 

function reloadexec() {    // Button execute
     debugPrintln("*** Reload ***");
     reloadEFS();
     return; 
} 

[ackmann]

Thank you very much for the fast reply!

Alex, Steve
thank you both very much for that fast reply.
Alex, your detailed explanation already helped a lot. Thanks again. I will try the replacement reloadEFS() -> main this evening.

I understood that the reloadEFS produced a loop but I think my question is,
why it doesn´t produce a loop in the initial load ?
Because in the initial load all bars are loaded two times.
Is there any possibility to load all bars automatically (without button) more than two times (-or does this the reloadEFS() -> main replacement do)?
nLoop = 1 as start value was intended by me.
acki

[ACM]

Re: Thank you very much for the fast reply!

acki
Assuming I understood correctly what you are asking that is because even though you initially set nLoop to 0 [when you declare the variable] you then increment the value of that variable by 1 at nLoop++ so on initial load the value of nLoop is already 1 hence executing the reloadEFS() command inside the condition if(nLoop==1...).
As both Steve and I suggested try setting the variable nLoop to -1 when you declare it and you will see that the script will not reload immediately after the initial load. Also, as Steve has already suggested, try using more debug statements as that will help you better understand what is occurring in your script
Alex

Originally posted by ackmann
Alex, Steve
thank you both very much for that fast reply.
Alex, your detailed explanation already helped a lot. Thanks again. I will try the replacement reloadEFS() -> main this evening.

I understood that the reloadEFS produced a loop but I think my question is,
why it doesn´t produce a loop in the initial load ?
Because in the initial load all bars are loaded two times.
Is there any possibility to load all bars automatically (without button) more than two times (-or does this the reloadEFS() -> main replacement do)?
nLoop = 1 as start value was intended by me.
acki

[ackmann]

Alex,

I think we have a missunderstanding based on some missing information from me (sorry):
That is I tested the EFS-script not in realtime but in the evening when no new data come in.

Of course if the initial value of nLoop=-1 there is no initial reload.
That was clear before. Some debugPrint-statements didn´t help me. Unfortunately also the suggested relaodEFS -> main replacement didn´t bring the desired result.

But if you try the EFS-script when no new data come in you will see that the initial reloadEFS()-command (in the example at nLoop=1) will execute but all following reloadEFS()-command (at nLoop=4) will not (regardless the start value of nLoop).

And what goes not in my head (just tell me if I´m stupid! ) is,
why the initial reloadEFS doesn´t produce an infinitive loop while the following reloadEFS do?
I am sorry that I was not able to make my question clear for you, but hopefully now my problem is understandable.

acki

[ACM]

acki
Try the enclosed revision of your script and look at the sequence of events listed in the Formula Output window (a sample is also enclosed below)
You will see that after the script has reloaded the global variable bReload is false which means that all the commands included in the if (bReload) condition are not executed hence no infinite loop.
Keep in mind [and this may be where your confusion arises] that the relaodEFS() command does not reset global variables (as explained in this KB article on the reloadEFS() function) so unless you reset them at BARSTATE_ALLBARS they will maintain the state prior to the reload.
Keep also in mind that in the condition
if (nIndex == 0 && (nState == BARSTATE_NEWBAR || nState == BARSTATE_ALLBARS))
you are setting bReload to true ONLY when nIndex==0 and nState==BARSTATE_NEWBAR. This is because the condition nIndex==0 and nState==BARSTATE_ALLBARS will never be true by definition as BARSTATE_ALLBARS occurs on the oldest bar index (ie the first time main is called)
Hope this helps
Alex

function preMain() {
     setPriceStudy(true);
     setShowCursorLabel(false); 
}

var nLoop   = 0; 
var bReload = false; 

function main() {  
     debugPrintln("bar index "+getCurrentBarIndex()+" line10 nLoop is "+nLoop+" bReload is "+bReload);
     trigger();     // sets bReload or not
     if (bReload) {
         debugPrintln("bar index "+getCurrentBarIndex()+" line13 nLoop is "+nLoop+" bReload is "+bReload)
         bReload     = false;
         debugPrintln("bar index "+getCurrentBarIndex()+" line15 nLoop is "+nLoop+" bReload is "+bReload);
         nLoop++;
         debugPrintln("bar index "+getCurrentBarIndex()+" line17 nLoop is "+nLoop)
         if (nLoop==1 || nLoop==4) {
            reloadEFS();     // !!!!!!!!  works with nLoop==1 but not with nLoop==4
            debugPrintln("bar index "+getCurrentBarIndex()+" line20 EFS reloaded by main reloadEFS command")
            debugPrintln("bar index "+getCurrentBarIndex()+" line21 nLoop is "+nLoop)
         }else {    // Reload-Button
             drawTextRelative(1, 15, " Reload @URL=EFS:reloadexec", null, null, Text.BUTTON | 
                              Text.RELATIVETOLEFT | Text.RELATIVETOBOTTOM, null, 10, "Reload");
         }
     }
     return close(); 
}  

function trigger() {
     var nState = getBarState();
     var nIndex = getCurrentBarIndex();
     if (nIndex == 0 && (nState == BARSTATE_NEWBAR || nState == BARSTATE_ALLBARS)) {
         bReload = true;
     } 
     return;
} 

function reloadexec() {    // Button execute
     debugPrintln("*** Reload ***");
     reloadEFS();
     return; 
} 

Originally posted by ackmann
Alex,

I think we have a missunderstanding based on some missing information from me (sorry):
That is I tested the EFS-script not in realtime but in the evening when no new data come in.

Of course if the initial value of nLoop=-1 there is no initial reload.
That was clear before. Some debugPrint-statements didn´t help me. Unfortunately also the suggested relaodEFS -> main replacement didn´t bring the desired result.

But if you try the EFS-script when no new data come in you will see that the initial reloadEFS()-command (in the example at nLoop=1) will execute but all following reloadEFS()-command (at nLoop=4) will not (regardless the start value of nLoop).

And what goes not in my head (just tell me if I´m stupid! ) is,
why the initial reloadEFS doesn´t produce an infinitive loop while the following reloadEFS do?
I am sorry that I was not able to make my question clear for you, but hopefully now my problem is understandable.

acki

[ackmann]

Last attempt

Alex,

thank you for so much work and for your last answer which allows me to formulate my question the last time (I promise!!! ) as precise as I can:
I took your version of the code and ran it without data coming in (I unplugged the network). In the attached screenshots you see the results at nLoop=1 and nLoop=4.

Everything is similar:
1. bReload is the same
2. The processed lines are the same
3. In both cases all bars were loaded before
4. nLoop=1 and nLoop=4 both lead to the main reloadEFS() command

But why is then the "nLoop=4 reloadEFS() command" nested whereas the "nLoop=1 reloadEFS() command" is not.

Where is the difference between the two situations (nLoop=1 and nLoop=4) which I don´t see so far?

acki

[ACM]

Re: Last attempt

acki
They are not the same as I see it. The difference is in the fact that at nLoop==3 you are clicking the Reload button to reload the efs hence using the reloadEFS() command in the reloadexec() function which sets nLoop to 4 and executes the reloadEFS() command in the main function thereby triggering the nested reloadEFS condition.
At nLoop==1 instead which occurs as you initially load the efs in the chart you are not reloading the efs using the command in the reloadexec() function but using only the one in the main() function
If instead of clicking the button you were to run this efs with real time data you would see that at the third new real time bar [ie at nLoop==4] the efs would reload using ONLY the reloadEFS() command in the main() function and you would not get a nested reloadEFS error message.
Alex

Originally posted by ackmann
Alex,

thank you for so much work and for your last answer which allows me to formulate my question the last time (I promise!!! ) as precise as I can:
I took your version of the code and ran it without data coming in (I unplugged the network). In the attached screenshots you see the results at nLoop=1 and nLoop=4.

Everything is similar:
1. bReload is the same
2. The processed lines are the same
3. In both cases all bars were loaded before
4. nLoop=1 and nLoop=4 both lead to the main reloadEFS() command

But why is then the "nLoop=4 reloadEFS() command" nested whereas the "nLoop=1 reloadEFS() command" is not.

Where is the difference between the two situations (nLoop=1 and nLoop=4) which I don´t see so far?

acki

[ackmann]

Thanks again for the explanation!

Alex,

Thanks again for the explanation!

acki

[ACM]

Re: Thanks again for the explanation!

acki
You are most welcome
Alex

Originally posted by ackmann
Alex,

Thanks again for the explanation!

acki