----- Original Message -----

From: Jack Purdum via Groups.Io

To: [email protected]

Sent: Tuesday, August 15, 2017 12:03 AM

Subject: Re: [BITX20] Encoder tinning

Also, you do not want to do any Serial.print() or Serial.println() calls in an interrupt service routine. They use their own ISR and you may end up in a blocking situation.

Jack, W8TEE

---

From: John P <j.m.price@...>
To: [email protected]
Sent: Monday, August 14, 2017 8:53 PM
Subject: Re: [BITX20] Encoder tinning

On Mon, Aug 14, 2017 at 05:21 pm, Norberto Modanesi wrote:

Would it work?

One way to find out!
?
--
John - WA2FZW

Hi folks:
What if I add these lines to the Raduino code
#include <Rotary.h>
Rotary r=Rotary(3, 2);

comment out analogtinnig A7

in setup()
Rotary r=Rotary(3, 2);

Then the interrupt routine
ISR(PCINT2_vect) {
unsigned char result = r.process();
if (result == DIR_CW)
counter=counter+1;
if(counter>1020){counter=1020;}
else if (result == DIR_CCW)
counter = counter-1;
if(counter<5){counter=5;}
Serial.println (counter);

}
Finaly, replace every analoRead(analogtunning) with counter.-

Would it work?

73 de LU5DNM

Norberto Modanesi
San Nicolás

OK, now you look again. The two /r /objects are defined at different
scope levels so there is no conflict. The first definition of /r /has
global scope while the second has function (local) scope. Since they are
at different scope levels, the compiler will not throw an error. Prove
it to yourself with this expansion of your post:

#include <Rotary.h>
Rotary r = Rotary(3, 2); // Global scope

int counter;

void setup()
{
Rotary r = Rotary(3, 2); // Local scope
unsigned char result = r.process();
if (result == 0)
counter = 0;
Serial.begin(9600);
}

ISR(PCINT2_vect) {
unsigned char result = r.process();
if (result == DIR_CW)
counter = counter + 1;
if (counter > 1020) {
counter = 1020;
}
else if (result == DIR_CCW)
counter = counter - 1;
if (counter < 5) {
counter = 5;
}
Serial.println (counter);

}
void loop() {
}

The compiler compiles the code without error, as it should.

Jack, W8TEE
------------------------------------------------------------------------
*From:* Vince Vielhaber <vev@...>
*To:* [email protected]
*Sent:* Tuesday, August 15, 2017 12:38 PM
*Subject:* Re: [BITX20] Encoder tinning


Look again:

#include <Rotary.h>
Rotary r=Rotary(3, 2); <----------------------

comment out analogtinnig A7

in setup()
Rotary r=Rotary(3, 2); <----------------------

r is defined more than once.


Vince.



On 08/15/2017 11:40 AM, Jack Purdum via Groups.Io wrote:

He's not redefining the object/r/. He is defining an /unsigned char/
variable named /result/ into which he is assigning the return value from
the /process() /method of the /r/ object. There should be no compiler
error from that.

Jack, W8TEE

------------------------------------------------------------------------
*From:* Vince Vielhaber <vev@... <mailto:vev@...>>
*To:* [email protected] <mailto:[email protected]>
*Sent:* Tuesday, August 15, 2017 10:58 AM
*Subject:* Re: [BITX20] Encoder tinning


You're going to get an error or warning for redefining r. You've
definied it globally right under your include statement, then again in
setup. In setup it's only local but you'll still get an error/warning.

Vince.


On 08/14/2017 08:21 PM, Norberto Modanesi wrote:

Hi folks:
What if I add these lines to the Raduino code
#include <Rotary.h>
Rotary r=Rotary(3, 2);

comment out analogtinnig A7

in setup()
Rotary r=Rotary(3, 2);

Then the interrupt routine
ISR(PCINT2_vect) {
unsigned char result = r.process();
if (result == DIR_CW)
counter=counter+1;
if(counter>1020){counter=1020;}
else if (result == DIR_CCW)
counter = counter-1;
if(counter<5){counter=5;}
Serial.println (counter);

}
Finaly, replace every analoRead(analogtunning) with counter.-

Would it work?

73 de LU5DNM

Norberto Modanesi
San Nicolás

--
Michigan VHF Corp.

<><>

<><>

<>

<>

<>

Virus-free. www.avast.com

<>

<#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>

--
Michigan VHF Corp. <>
<>

<>

I agree that the duplicate definition is unnecessary. What I was
disagreeing with was your statement that the definition would cause a
compiler error; it won't. The most you might get is a warning if /r/
goes unused at that scope level. Also, the definition of /r /in
/setup()/ really doesn't waste any SRAM because, as soon as execution
leaves /setup()/, the stack space allocated to/r /is reclaimed. It's
variables defined in the heap space that don't get reclaimed and
permanently chew up SRAM space.

Jack, W8TEE
------------------------------------------------------------------------
*From:* Vince Vielhaber <vev@...>
*To:* [email protected]
*Sent:* Tuesday, August 15, 2017 2:21 PM
*Subject:* Re: [BITX20] Encoder tinning

The one in the local scope is still redefining r. It's also unneeded.
You're setting a global variable (counter) based on r.process()'s value.
You're not changing anything to do with r, so while it's shadowing the
global r, it's also wasting precious program space.

Because the compiler (wrongly, IMO) doesn't complain, it's no reason to
do it.

Vince.



On 08/15/2017 02:05 PM, Jack Purdum via Groups.Io wrote:

OK, now you look again. The two /r /objects are defined at different
scope levels so there is no conflict. The first definition of /r /has
global scope while the second has function (local) scope. Since they are
at different scope levels, the compiler will not throw an error. Prove
it to yourself with this expansion of your post:

#include <Rotary.h>
Rotary r = Rotary(3, 2); // Global scope

int counter;

void setup()
{
Rotary r = Rotary(3, 2); // Local scope
unsigned char result = r.process();
if (result == 0)
counter = 0;
Serial.begin(9600);
}

ISR(PCINT2_vect) {
unsigned char result = r.process();
if (result == DIR_CW)
counter = counter + 1;
if (counter > 1020) {
counter = 1020;
}
else if (result == DIR_CCW)
counter = counter - 1;
if (counter < 5) {
counter = 5;
}
Serial.println (counter);

}
void loop() {
}

The compiler compiles the code without error, as it should.

Jack, W8TEE
------------------------------------------------------------------------
*From:* Vince Vielhaber <vev@... <mailto:vev@...>>
*To:* [email protected] <mailto:[email protected]>
*Sent:* Tuesday, August 15, 2017 12:38 PM
*Subject:* Re: [BITX20] Encoder tinning


Look again:

#include <Rotary.h>
Rotary r=Rotary(3, 2); <----------------------

comment out analogtinnig A7

in setup()
Rotary r=Rotary(3, 2); <----------------------

r is defined more than once.


Vince.



On 08/15/2017 11:40 AM, Jack Purdum via Groups.Io wrote:

He's not redefining the object/r/. He is defining an /unsigned char/
variable named /result/ into which he is assigning the return value from
the /process() /method of the /r/ object. There should be no compiler
error from that.

Jack, W8TEE

------------------------------------------------------------------------
*From:* Vince Vielhaber <vev@... <mailto:vev@...>

<mailto:vev@... <mailto:vev@...>>>

*To:* [email protected] <mailto:[email protected]>

<mailto:[email protected] <mailto:[email protected]>>

*Sent:* Tuesday, August 15, 2017 10:58 AM
*Subject:* Re: [BITX20] Encoder tinning


You're going to get an error or warning for redefining r. You've
definied it globally right under your include statement, then again in
setup. In setup it's only local but you'll still get an error/warning.

Vince.


On 08/14/2017 08:21 PM, Norberto Modanesi wrote:

Hi folks:
What if I add these lines to the Raduino code
#include <Rotary.h>
Rotary r=Rotary(3, 2);

comment out analogtinnig A7

in setup()
Rotary r=Rotary(3, 2);

Then the interrupt routine
ISR(PCINT2_vect) {
unsigned char result = r.process();
if (result == DIR_CW)
counter=counter+1;
if(counter>1020){counter=1020;}
else if (result == DIR_CCW)
counter = counter-1;
if(counter<5){counter=5;}
Serial.println (counter);

}
Finaly, replace every analoRead(analogtunning) with counter.-

Would it work?

73 de LU5DNM

Norberto Modanesi
San Nicolás

--
Michigan VHF Corp.

<><>

<><><>

<>

<>

<>

<>

Virus-free. www.avast.com

<>

<#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>

--
Michigan VHF Corp.

<><>

<><>

<>

<>

--
Michigan VHF Corp. <>
<>

<>

Part of the reason for this exchange we're having is that I don't want anyone to be discouraged from trying to experiment with programming an Arduino. To you and me, error/warning is no big deal, but to a newcomer to programming, I've actually seen it literally paralize a student. They think they "broke" something. An error means the compile stops and the upload does not happen. While the duplicate definition of the object you pointed out shouldn't be done, it is not an error. Indeed, setting the proper flags to the compiler and it won't even issue a warning. Warnings are often semantic (not syntax) errors that are not sufficiently bad to cease the compile. Unused variables are an examlpe. The mistake you first pointed out is not an error, but will generate a warning. I don't want warnings to discourage new programmers from slogging through those warnings.

Starting with (I think) IDE version 1.8.0, the gcc parser was tightened up, which caused it to increase the number of warning significantly. The thing that irritates me most is that they did not take the time to "fix" the standard library files to the point where no warnings are issues. Newcomers see these warnings and may wonder if it is "safe" to proceed. Generally, if the IDE completes the upload, you're good to go. If any warnings from the IDE pertain to your own code, you should take the time to fix them. Otherwise, I leave them alone. I'm pretty sure that I could fix the standard library warnings, but then I would be out of sync with everyone else, so I don't. I need to see the same problems anyone else may see.

Namespace collisions are pretty rare. Names in other files are private to those files so the namespace doesn't flow between files unless you want it to (e.g., using the extern keyword). Include files can be nettlesome because some do define global variables. Still, those are pretty obvious since your INO file had a #include in it. You're right about the compile-time statistics; they cannot reflect runtime memory use other than functions like freeRam(). As a rule, if SRAM starts to approach 70% usage, I start to get nervous since parameter use via function calls isn't directly available. If your code starts to act "funny", quite often it's because you've run out of runtime SRAM even though the compiletime stats say you have a third left.

To any beginner who wants to work at removing warning or error messages from their own code, ususally a Google search using the format: "Arduino " + plus the error/warning message will give you enough information to isolate the error. Also, the IDE often points you to the proper line number. Sometimes the error is many lines earlier before the compiler throws up its hand and gives up, so if the line of the error message looks ok, look at earlier lines for the error.

Jack, W8TEE

---

From: Vince Vielhaber <vev@...>
To: [email protected]
Sent: Tuesday, August 15, 2017 10:58 PM
Subject: Re: [BITX20] Encoder tinning

Actually I said "error/warning".? Arduino seems to let you get away with
that, gcc and g++ typically do not 'cuze I see that regularly when
compiling someone else's code or when I forget and reuse a global's name
- which isn't hard when you're dealing with a program spread across a
couple dozen files with 100s of thousands of lines of code.

The wasted space I was referring to is the compiled code, not the
running code.

Vince.



On 08/15/2017 03:05 PM, Jack Purdum via Groups.Io wrote:
> I agree that the duplicate definition is unnecessary. What I was
> disagreeing with was your statement that the definition would cause a
> compiler error; it won't. The most you might get is a warning if /r/
> goes unused at that scope level. Also, the definition of /r /in
> /setup()/ really doesn't waste any SRAM because, as soon as execution
> leaves /setup()/, the stack space allocated to/r /is reclaimed. It's
> variables defined in the heap space that don't get reclaimed and
> permanently chew up SRAM space.
>
> Jack, W8TEE
> ------------------------------------------------------------------------
> *From:* Vince Vielhaber <vev@...>
> *To:* [email protected]
> *Sent:* Tuesday, August 15, 2017 2:21 PM
> *Subject:* Re: [BITX20] Encoder tinning
>
> The one in the local scope is still redefining r.? It's also unneeded.
> You're setting a global variable (counter) based on r.process()'s value.
>? You're not changing anything to do with r, so while it's shadowing the
> global r, it's also wasting precious program space.
>
> Because the compiler (wrongly, IMO) doesn't complain, it's no reason to
> do it.
>
> Vince.
>
>
>
> On 08/15/2017 02:05 PM, Jack Purdum via Groups.Io wrote:
>> OK, now you look again. The two /r /objects are defined at different
>> scope levels so there is no conflict. The first definition of /r /has
>> global scope while the second has function (local) scope. Since they are
>> at different scope levels, the compiler will not throw an error. Prove
>> it to yourself with this expansion of your post:
>>
>> #include <Rotary.h>
>> Rotary r = Rotary(3, 2);? ? ? // Global scope
>>
>> int counter;
>>
>> void setup()
>> {
>>? Rotary r = Rotary(3, 2);? ? // Local scope
>>? unsigned char result = r.process();
>>? if (result == 0)
>>? ? counter = 0;
>>? Serial.begin(9600);
>> }
>>
>> ISR(PCINT2_vect) {
>>? unsigned char result = r.process();
>>? if (result == DIR_CW)
>>? ? counter = counter + 1;
>>? if (counter > 1020) {
>>? ? counter = 1020;
>>? }
>>? else if (result == DIR_CCW)
>>? ? counter = counter - 1;
>>? if (counter < 5) {
>>? ? counter = 5;
>>? }
>>? Serial.println (counter);
>>
>> }
>> void loop() {
>> }
>>
>> The compiler compiles the code without error, as it should.
>>
>> Jack, W8TEE
>> ------------------------------------------------------------------------
>> *From:* Vince Vielhaber <vev@... <mailto:vev@...>>
>> *To:* [email protected] <mailto:[email protected]>
>> *Sent:* Tuesday, August 15, 2017 12:38 PM
>> *Subject:* Re: [BITX20] Encoder tinning
>>
>>
>> Look again:
>>
>>>> #include <Rotary.h>
>>>> Rotary r=Rotary(3, 2);? ? <----------------------
>>>>
>>>> comment out? analogtinnig? A7
>>>>
>>>> in setup()
>>>> Rotary r=Rotary(3, 2);? ? <----------------------
>>
>> r is defined more than once.
>>
>>
>> Vince.
>>
>>
>>
>> On 08/15/2017 11:40 AM, Jack Purdum via Groups.Io wrote:
>>> He's not redefining the object/r/. He is defining an /unsigned char/
>>> variable named /result/ into which he is assigning the return value from
>>> the /process() /method of the? /r/ object. There should be no compiler
>>> error from that.
>>>
>>> Jack, W8TEE
>>>
>>> ------------------------------------------------------------------------
>>> *From:* Vince Vielhaber <vev@... <mailto:vev@...>
> <mailto:vev@... <mailto:vev@...>>>
>>> *To:* [email protected] <mailto:[email protected]>
> <mailto:[email protected] <mailto:[email protected]>>
>>> *Sent:* Tuesday, August 15, 2017 10:58 AM
>>> *Subject:* Re: [BITX20] Encoder tinning
>>>
>>>
>>> You're going to get an error or warning for redefining r.? You've
>>> definied it globally right under your include statement, then again in
>>> setup.? In setup it's only local but you'll still get an error/warning.
>>>
>>> Vince.
>>>
>>>
>>> On 08/14/2017 08:21 PM, Norberto Modanesi wrote:
>>>> Hi folks:
>>>> What if I add these lines to the Raduino code
>>>> #include <Rotary.h>
>>>> Rotary r=Rotary(3, 2);
>>>>
>>>> comment out? analogtinnig? A7
>>>>
>>>> in setup()
>>>> Rotary r=Rotary(3, 2);
>>>>
>>>> Then the interrupt routine
>>>> ISR(PCINT2_vect) {
>>>>? unsigned char result = r.process();
>>>>? if (result == DIR_CW)
>>>>? ? counter=counter+1;
>>>>? ? if(counter>1020){counter=1020;}
>>>>? else if (result == DIR_CCW)
>>>>? ? counter = counter-1;
>>>>? ? if(counter<5){counter=5;}
>>>>? ? Serial.println (counter);
>>>>
>>>>? ? }
>>>> Finaly, replace every analoRead(analogtunning) with counter.-
>>>>
>>>> Would it work?
>>>>
>>>> 73 de LU5DNM
>>>>
>>>> Norberto Modanesi
>>>> San Nicolás
>>>>
>>>>
>>>>
>>>
>>> --
>>>? Michigan VHF Corp.?
>> <><>
>>>
> <><><>
>>>? ? ? ? ? ? ? ? ? ? ? ? ?
> <>
>> <>
>>> <>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>
> <>
>>>? ? Virus-free.
>>>
>>
> <>
>>>
>>>
>>> <#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
>>>
>>
>> --
>>? Michigan VHF Corp.?
> <><>
>> <><>
>>? ? ? ? ? ? ? ? ? ? ? ? ?
> <>
>> <>
>>
>>
>>
>>
>>
>>
>
> --
>? Michigan VHF Corp.? <>
> <>
>? ? ? ? ? ? ? ? ? ? ? ? ?
> <>
>
>
>
>
>
>

--
? Michigan VHF Corp.? ?
? ? ? ? ? ? ? ? ? ? ? ? ?