C47 Bug Reports

This area is for discussion about these families of custom high-end Scientific Calculator applications for SwissMicros devices.
MaxGaspa
Posts: 3
Joined: Wed Sep 06, 2017 11:39 am

Re: C43 Bug Reports

Post by MaxGaspa »

Hi, I'm writing to report an issue I got installing the C43.

Steps to reproduce

1) setup
2) system
3) enter system nenu
4) reset to dmcp menu
5) load program

then I selected DM42-3.22.pgm

6) enter to continue


hash ok
erasing
writing
loading ok

7) press enter

the about display appears

8) exit

Checking DMCP content

then the DM42 FW starts correctly.


OK now for the C43

1,2,3,4,5 as above

the I select 108_08c_c43...

6) enter to continue

hash ok
erasing
writing
loading ok

7) press enter

the about display appears

8) exit

the program info appears

9) enter


...now I got the DMCP menu again.

I pressed the rear reset button but I got the about menu ...exit...I got the program info... enter ...DMCP menu...endlessy.


To exit from this bricked status I have to re-load the DM42-3.22.pgm and the DM42 is again working.

Any idea? It may be related to the HW version of my DM42, I have the DMCP 3.24 (IFC 0315) BASE , DM42 3.22 , the diagnostic menu is showing HW-ID : 1

Any help?
HP15-LE DM16L HP48G HP48SX DM42
User avatar
Jaymos
Posts: 1634
Joined: Sun Nov 04, 2018 7:03 pm
Location: Cape Town

Re: C43 Bug Reports

Post by Jaymos »

MaxGaspa wrote:
Wed Nov 30, 2022 10:45 pm
Hi, I'm writing to report an issue I got installing the C43.

Steps to reproduce

1) setup
2) system
3) enter system nenu
4) reset to dmcp menu
5) load program

then I selected DM42-3.22.pgm

6) enter to continue


hash ok
erasing
writing
loading ok

7) press enter

the about display appears

8) exit

Checking DMCP content

then the DM42 FW starts correctly.


OK now for the C43

1,2,3,4,5 as above

the I select 108_08c_c43...

6) enter to continue

hash ok
erasing
writing
loading ok

7) press enter

the about display appears

8) exit

the program info appears

9) enter


...now I got the DMCP menu again.

I pressed the rear reset button but I got the about menu ...exit...I got the program info... enter ...DMCP menu...endlessy.


To exit from this bricked status I have to re-load the DM42-3.22.pgm and the DM42 is again working.

Any idea? It may be related to the HW version of my DM42, I have the DMCP 3.24 (IFC 0315) BASE , DM42 3.22 , the diagnostic menu is showing HW-ID : 1

Any help?
It seems the step for loading the C43qspi.bin file was left out. Please try that just before you load 108_08c_c43. The interaction of the qspi file may cause a different ’brick’ mode which is discussed somewhere else on these forums. If you do get this, keep F1 pressed while activating the rear reset button to access the DMCP menu again.

Let us know here of your progress.
Jaco Mostert
Elec Eng, South Africa
https://47calc.com C47 (s/n 03818 & 06199), WP43 (0015). In box: HP42S, HP32Sii, WP34S&C, HP28C, HP35s, EL-506P, EL-W506, PB700; ex: FX702P, 11C, HP67 & HP85; iOS: 42s Byron, Free42+, WP31S/34S, HCalc.
burkhard
Posts: 43
Joined: Fri Mar 16, 2018 10:15 pm

Re: C43 Bug Reports

Post by burkhard »

This seems a bug. It occurs for me in both the hardware (reflashed DM42) and simulator.

I find that if I place a number into x (but do not press ENTER) and then long-press a function key, the option for the f-function gets skipped when it cycles through the functions on that key.

EXAMPLE
Suppose I want to take the arcsin of 0.5 when in degree mode. The answer should be 30 obviously, with ARCSIN invoked by f-SIN.

0.5 (don't press ENTER)
long press of SIN key yields SIN → GTO → NOP instead of the expected SIN → ARCSIN → GTO → NOP.

•Note you can get to the g-function fine; it's just the f-function that is skipped.
•The problem can be avoided by closing the 0.5 entry with ENTER, but it seems this shouldn't be required.
•Actually pressing the physical yellow shift and then the SIN key to achieve ARCSIN works fine. It's just the long-press cycling that doesn't work for the f-functions when the operand hasn't been "closed".
•This isn't specific to just ARCSIN... it seems like a common problem with all f-functions on long-press if the operand has not been closed.

While I've mentioned before that I'm not a fan of the single shift key, I do very much like this feature of long-pressing to cycle through the shifted functions. That is a great idea which (except as reported above) works really well, and does make the single shift key much more "livable" for me. It's a very nice innovation which I wish were on all calculators!
User avatar
Jaymos
Posts: 1634
Joined: Sun Nov 04, 2018 7:03 pm
Location: Cape Town

Re: C43 Bug Reports

Post by Jaymos »

burkhard wrote:
Thu Dec 01, 2022 6:34 pm
This seems a bug. It occurs for me in both the hardware (reflashed DM42) and simulator.

I find that if I place a number into x (but do not press ENTER) and then long-press a function key, the option for the f-function gets skipped when it cycles through the functions on that key.

EXAMPLE
Suppose I want to take the arcsin of 0.5 when in degree mode. The answer should be 30 obviously, with ARCSIN invoked by f-SIN.

0.5 (don't press ENTER)
long press of SIN key yields SIN → GTO → NOP instead of the expected SIN → ARCSIN → GTO → NOP.

•Note you can get to the g-function fine; it's just the f-function that is skipped.
•The problem can be avoided by closing the 0.5 entry with ENTER, but it seems this shouldn't be required.
•Actually pressing the physical yellow shift and then the SIN key to achieve ARCSIN works fine. It's just the long-press cycling that doesn't work for the f-functions when the operand hasn't been "closed".
•This isn't specific to just ARCSIN... it seems like a common problem with all f-functions on long-press if the operand has not been closed.

While I've mentioned before that I'm not a fan of the single shift key, I do very much like this feature of long-pressing to cycle through the shifted functions. That is a great idea which (except as reported above) works really well, and does make the single shift key much more "livable" for me. It's a very nice innovation which I wish were on all calculators!

Thanks for the catch. I can't yet say if it was, or became broken, but I can confirm it is. Repeatable, therefore a bug that I will arrest and deal with ;-)
Jaco Mostert
Elec Eng, South Africa
https://47calc.com C47 (s/n 03818 & 06199), WP43 (0015). In box: HP42S, HP32Sii, WP34S&C, HP28C, HP35s, EL-506P, EL-W506, PB700; ex: FX702P, 11C, HP67 & HP85; iOS: 42s Byron, Free42+, WP31S/34S, HCalc.
MaxGaspa
Posts: 3
Joined: Wed Sep 06, 2017 11:39 am

Re: C43 Bug Reports

Post by MaxGaspa »

Jaymos wrote:
Thu Dec 01, 2022 6:33 am

It seems the step for loading the C43qspi.bin file was left out. Please try that just before you load 108_08c_c43. The interaction of the qspi file may cause a different ’brick’ mode which is discussed somewhere else on these forums. If you do get this, keep F1 pressed while activating the rear reset button to access the DMCP menu again.

Let us know here of your progress.
Yes. I wasn't aware that I need to load the qspi file. I didn't find this step in the docs...but it should be my fault.

Anyway loading the qspi I got the C43 working......I would say...I like it very much...

I really need the bezel now...

Thank you for your advice.

Max
HP15-LE DM16L HP48G HP48SX DM42
burkhard
Posts: 43
Joined: Fri Mar 16, 2018 10:15 pm

Re: C43 Bug Reports

Post by burkhard »

Jaymos wrote:
Thu Dec 01, 2022 8:59 pm
Thanks for the catch. I can't yet say if it was, or became broken, but I can confirm it is. Repeatable, therefore a bug that I will arrest and deal with ;-)
Excellent, glad you reproduced it OK.

Playing with this, I think I found a couple more mini-bugs relating to long-press cycling through the functions:
Neither of these require un-ended entries into the stack. They are present all the time.

1) Long-pressing x↔y yields: x↔y → LASTx → NOP instead of x↔y → LASTx → R↓ →NOP
2) Long-pressing ENTER yields: ENTER → COMPLEX → NOP instead of ENTER → COMPLEX → CPX → NOP

In both cases the missing g-functions work fine when the yellow button is double-pressed, so it's just a minor UI thing on long-press.

Thanks!
User avatar
Jaymos
Posts: 1634
Joined: Sun Nov 04, 2018 7:03 pm
Location: Cape Town

Re: C43 Bug Reports

Post by Jaymos »

burkhard wrote:
Thu Dec 01, 2022 11:36 pm
Jaymos wrote:
Thu Dec 01, 2022 8:59 pm
Thanks for the catch. I can't yet say if it was, or became broken, but I can confirm it is. Repeatable, therefore a bug that I will arrest and deal with ;-)
Excellent, glad you reproduced it OK.

Playing with this, I think I found a couple more mini-bugs relating to long-press cycling through the functions:
Neither of these require un-ended entries into the stack. They are present all the time.

1) Long-pressing x↔y yields: x↔y → LASTx → NOP instead of x↔y → LASTx → R↓ →NOP
2) Long-pressing ENTER yields: ENTER → COMPLEX → NOP instead of ENTER → COMPLEX → CPX → NOP

In both cases the missing g-functions work fine when the yellow button is double-pressed, so it's just a minor UI thing on long-press.

Thanks!
The bug was found and fixed. I will update/provide the download link soon, probably this weekend if all goes well.

I'm glad you are finding the longpress add-on useful - you are not the first saying this but it is still good to hear. Either way, this week we have internally specified configuration options where the longpress add-on performance can be altered and can switched out (in the MODE menu). That may take a while to implement but the spec is there.

As a matter of interest, 2 issues you mentioned have the same cause (issue 0 in the previous post and issue 1 here). Issue 2 is planned and correct as menus are not included. The purpose of the longpress on the top rows (non-menu keys) is an add-on to access the math functions and since most keys to the right and below EEX are actioned without delay, as such most menus are inaccessible, therefore menus are also excluded.
Jaco Mostert
Elec Eng, South Africa
https://47calc.com C47 (s/n 03818 & 06199), WP43 (0015). In box: HP42S, HP32Sii, WP34S&C, HP28C, HP35s, EL-506P, EL-W506, PB700; ex: FX702P, 11C, HP67 & HP85; iOS: 42s Byron, Free42+, WP31S/34S, HCalc.
User avatar
rudi
Posts: 413
Joined: Wed Nov 03, 2021 9:03 am
Location: Denmark
Contact:

Re: C43 Bug Reports

Post by rudi »

While testing the new DT->J and J->DT in C43 108_08g, I found an issue with the ->TIME and TIME-> functions.

Sometimes, J->DT returned a date and the time as 24:00:00, which is ok, yet akward, it just means the next date at 00:00:00.
After some investigation, I found that it happens when the time is between 24:00:00 and 23:59:59.5

To reproduce:

Code: Select all

[blue] [CLK]
DATE
TIME
DT->J
[yellow] [PART] IP
0,4999999
+
[blue] [CLK]
J->DT
This will show the time in x register as 24:00:00 !

Trying TIME-> and then ->TIME causes error message "Bad time or date input"

My C43's TDISP is set to 0 (default?) Setting TDISP to 6 will return time as 24:00:00.000 in the above example, and TIME-> returns
z: 24
y: 0
x: 0

I would have expected
z: 23
y: 59
x: 59.99136
/Rudi

DM-42 (s/n 06999), HP-42S, HP-35s, HP-11c, HP-32SII (ex HP-41CV, ex HP-75C, ex HP-48G + a lot, really lot of a accessories)
Denmark
User avatar
Jaymos
Posts: 1634
Joined: Sun Nov 04, 2018 7:03 pm
Location: Cape Town

Re: C43 Bug Reports

Post by Jaymos »

rudi wrote:
Mon Dec 05, 2022 6:23 am
While testing the new DT->J and J->DT in C43 108_08g, I found an issue with the ->TIME and TIME-> functions.

Sometimes, J->DT returned a date and the time as 24:00:00, which is ok, yet akward, it just means the next date at 00:00:00.
After some investigation, I found that it happens when the time is between 24:00:00 and 23:59:59.5

To reproduce:

Code: Select all

[blue] [CLK]
DATE
TIME
DT->J
[yellow] [PART] IP
0,4999999
+
[blue] [CLK]
J->DT
This will show the time in x register as 24:00:00 !

Trying TIME-> and then ->TIME causes error message "Bad time or date input"

My C43's TDISP is set to 0 (default?) Setting TDISP to 6 will return time as 24:00:00.000 in the above example, and TIME-> returns
z: 24
y: 0
x: 0

I would have expected
z: 23
y: 59
x: 59.99136

Let me analyse it out loud:

After working through your example and verifying the code, I agree that you point out a coding error. Somehow in the deliberate rounding of the decimal JDN (which btw is done by rounding the sexagesimal seconds to 0.01), I made an error in the rounding calculation, which meant that it previously rounded a number which is precisely 24 times larger, hence it was rounding up too soon. With this fixed, the last result that you show above is now:

Code: Select all

z: 23
y: 59
x: 59.990 (instead of 59.99136)
The difference of 0.00136 seconds indicated above is now according to the planned rounding and not to a bug.


About TIME-> and ->TIME, that should work and it gets stumped by the fact that hours are defined to be up to 23. However, I feel that this restriction is unneccesary as the time type can store much larger times, and TIME-> works for that. I will have internal discussions to relax this limitation. I also do not like that the input to ->TIME are all truncated. I will also internally submit for a change to TIME: surely 0 0.5 1 ->TIME should convert to 00:00:31 and similarly 0 61 1 ->TIME should convert to 01:01:01. After all ->h.ms already does these gymnastics, i.e. 0.6101 ->h.ms results in 1:01:01.
Jaco Mostert
Elec Eng, South Africa
https://47calc.com C47 (s/n 03818 & 06199), WP43 (0015). In box: HP42S, HP32Sii, WP34S&C, HP28C, HP35s, EL-506P, EL-W506, PB700; ex: FX702P, 11C, HP67 & HP85; iOS: 42s Byron, Free42+, WP31S/34S, HCalc.
User avatar
rudi
Posts: 413
Joined: Wed Nov 03, 2021 9:03 am
Location: Denmark
Contact:

Re: C43 Bug Reports

Post by rudi »

Sorry for the rounding error in my example, had only had one mug of coffee at the time of writing :roll:

And yes, I can deal with time values with hour value >= 24.
But if You call a function and then it's inverse, I wouldn't expect an error message ;-)

PS: If I could just be able to compile the git repository, I would be more than happy to help out with the coding. But that's for another thread I guess.
/Rudi

DM-42 (s/n 06999), HP-42S, HP-35s, HP-11c, HP-32SII (ex HP-41CV, ex HP-75C, ex HP-48G + a lot, really lot of a accessories)
Denmark
Post Reply