[Bug] Keeps freezing

[Tugdual]

Again and again. I think there must be an issue with the screen management.
Some reported the slow code editor but I think the issue is everywhere. the Prime keeps freezing all the time, in the middle of a function exploration while scrolling or zooming, in home, in CAS. When I On-Symb it, I have lost all my last 3mn work (I mean everything, screen buffers, variables, everything). Providing that 3mn before I was already recovering from a freeze, it becomes hardly not possible to work with the Prime.

Luckily, I don't depend too much on the Prime but if I were a young student having invested a fair amount of money in this calculator and quite dramatically depending on it for my studies, I guess I would start writing pretty strong coloured words...

Seriously, this seems to be a major issue that needs to be addressed. It is quite visible, increasing the chances to get it fixed.

[Helge Gabert]

Perhaps you encountered some errors while updating. Go back to rev 5447, that would be my suggestion (that is what I did).

Make sure everything works properly with rev 5447, then do the update to rev 6030 again.

[Tugdual]

I tried that... didn't work.
Here is how I kill the Prime, seems to be reproducible all the time:
1. go in CAS and make sure you're in radians
2. enter

Code:

series(tan(x),x=0,20)

You should now see $$x+\frac{{x}^3}{3}+2*\frac{{x}^{5}}{15}+17*\frac{{x}^{7}}{315}+62*\frac{{x}^{9}​}{2835}+1382* \frac{{x}^{11}}{155925}+21844*\frac{{x}^{13}}{6081075}+929569*\frac{{x}^{15}}{63​8512875}+6404582*\frac{{x}^{17}}{10854718875}+443861162*\frac{{x}^{19}}{18561569​27625}+x^{21}*order\_ size(x)$$
3. Copy the former line and change the end so that it reads:
$$x+\frac{{x}^3}{3}+2*\frac{{x}^{5}}{15}+17*\frac{{x}^{7}}{315}+62*\frac{{x}^{9}​}{2835}+1382* \frac{{x}^{11}}{155925}+21844*\frac{{x}^{13}}{6081075}+929569*\frac{{x}^{15}}{63​8512875}+6404582*\frac{{x}^{17}}{10854718875}+443861162*\frac{{x}^{19}}{18561569​27625}▶f(x)$$
4. Press enter and now move into the Function App
5. Symb, enter F1(X)=f(X)
6. Plot
7. Mess around with fingers like zoom, scroll, after less than 5 secs, the Prime is frozen.
8. Now On-Symb it, go back to CAS, all is gone... Even f(x) is set back to an earlier definition which is pretty weird because it means the Prime is doing backups!

Note: This is just an example demonstrating the issue, as I mentioned I keep getting it a lot.

[Helge Gabert]

I believe you cannot define your function that way (as you displayed in step 3), because you are defining a CAS function in terms of the built-in variable Ans.

Instead, try in CAS

f(x):= x + x^3/3 + . . .

Then, in Function App, enter f(X) in the F1 field.

Does it still freeze?

[Tugdual]

(06-01-2014 04:37 PM)Helge Gabert Wrote:  I believe you cannot define your function that way (as you displayed in step 3), because you are defining a CAS function in terms of the built-in variable Ans.

Instead, try in CAS

f(x):= x + x^3/3 + . . .

Then, in Function App, enter f(X) in the F1 field.

Does it still freeze?

Hi Gabert, I think the two syntaxes are identical and since I see the curve, I don't think the problem is here. I suspect that calling CAS from the App consumes a lot of CPU to the detriment of the new kinetic scrolling which seems to crash.

[Helge Gabert]

I don't believe the two methods are identical.

With your method (see step 3), try, say f(8). What do you get? f(Ans)?

With the method I mentioned, f(8) gives me a value, and my Prime doesn't freeze, when scrolling around.

Try it!

[parisse]

I don't know why it freezes, it's certainly a bad CAS/non CAS interaction.
By the way, you can define f(x) without editing by the following command:

Code:

f:=unapply(convert(series(tan(x),x=0,20),polynom),x)

[Tugdual]

(06-01-2014 05:01 PM)Helge Gabert Wrote:  I don't believe the two methods are identical.

With your method (see step 3), try, say f(8). What do you get? f(Ans)?

With the method I mentioned, f(8) gives me a value, and my Prime doesn't freeze, when scrolling around.

Try it!

I tried and I don't see any difference.
Can you show me some screenshots showing the difference?
To me me ▶ (STO) and ":=" are identical features.

Also ':=' on the prime requires 4 keys while ▶ requires 2 unless we're not talking about the same thing?

(06-01-2014 05:12 PM)parisse Wrote:  I don't know why it freezes, it's certainly a bad CAS/non CAS interaction.
By the way, you can define f(x) without editing by the following command:

Code:

f:=unapply(convert(series(tan(x),x=0,20),polynom),x)

Does it also freeze on yours?

[Helge Gabert]

Weird . . . your method works now. I defined g(x) with the solid triangle arrow, right after copying the output of series(tan(x),x=1,20) . . . , and in the memory manager, it is displayed correctly, under CAS vars, function. So yes, the two methods are identical, I agree!

But before, all I can say is that f(x) displayed incorrectly, showing f(Ans) when I tried to calculate f(0.8)! I wished I could produce a screenshot, but I overwrote f right away with the definition of the series obtained by my method.

However, when I monkey around the screen, zooming, using the touchscreen, etc., my unit doesn't freeze - - and it doesn't freeze (or, at least, it hasn't frozen yet) whether I show just g(x), or g(x) and tan(x) simultaneously.

[Tugdual]

(06-01-2014 10:47 PM)Helge Gabert Wrote:  But before, all I can say is that f(x) displayed incorrectly, showing f(Ans) when I tried to calculate f(0.8)! I wished I could produce a screenshot, but I overwrote f right away with the definition of the series obtained by my method.

I do get a result both in Home and CAS.

(06-01-2014 10:47 PM)Helge Gabert Wrote:  However, when I monkey around the screen, zooming, using the touchscreen, etc., my unit doesn't freeze - - and it doesn't freeze (or, at least, it hasn't frozen yet) whether I show just g(x), or g(x) and tan(x) simultaneously.

I tried something more drastic this morning.
I just took the f(x) definition and replaced all 'x' with 'X' and put it straight into F1(X) defintion. Same problem! We can now eliminate CAS.

Just out of curiosity, did you also have the battery drain issue after update?
Looks like some people didn't get it (or didn't notice it?) and I wonder if there are not already multiple hardwares with multiple issues...

[jebem]

We could clarify this issue by posting here what are our serial numbers.
If the offending hardware belongs to the same batch productions (meaning, the same weeks) then maybe we have something to work with.

[Tugdual]

(06-02-2014 09:14 AM)jebem Wrote:  We could clarify this issue by posting here what are our serial numbers.
If the offending hardware belongs to the same batch productions (meaning, the same weeks) then maybe we have something to work with.

May be.
However I haven't heard HP denying the issue or requiring this infomation.
Actually I would be interested in getting feeback from other users. Am I the only one having this problem?

[Miguel Toro]

(06-02-2014 12:53 PM)Tugdual Wrote:  

(06-02-2014 09:14 AM)jebem Wrote:  We could clarify this issue by posting here what are our serial numbers.
If the offending hardware belongs to the same batch productions (meaning, the same weeks) then maybe we have something to work with.

May be.
However I haven't heard HP denying the issue or requiring this infomation.
Actually I would be interested in getting feeback from other users. Am I the only one having this problem?

I tried the example to see if I could reproduce the issue, and sure I did: my unit froze after some touchscreen manipulations (zooming, scrolling, etc.)

HP-Prime revision A 4CY3430757.

Cheers,

Miguel

[Helge Gabert]

[/quote]
I tried something more drastic this morning.
I just took the f(x) definition and replaced all 'x' with 'X' and put it straight into F1(X) defintion. Same problem! We can now eliminate CAS.

Just out of curiosity, did you also have the battery drain issue after update?
Looks like some people didn't get it (or didn't notice it?) and I wonder if there are not already multiple hardwares with multiple issues...
[/quote]

Yes, I also had the battery drain issue, which has been resolved.

One other thought: I noticed from your screen shot that you were in Geometry app when you defined f. Does it make any difference if you are in, say, function app?

[Tugdual]

(06-02-2014 03:45 PM)Helge Gabert Wrote:  One other thought: I noticed from your screen shot that you were in Geometry app when you defined f. Does it make any difference if you are in, say, function app?

The screenshot was done on the emulated Prime but the real one is the same and the application doesn't seem to impact the result. Very weird that we don't get the same!?
Did you try the emulated version? At least we're 100% certain that the hardwared will not be involved on the emulated version and you should get identical results.

[Helge Gabert]

Hi,

I tried the emulator, and it appears to work fine with the CAS definition of the series in the function app. The only problem I can see is that the slope does not get computed (returns NaN - - not a number). Other features work as expected, e.g., root, or area between two x values.

[Tugdual]

(06-02-2014 07:48 PM)Helge Gabert Wrote:  Hi,

I tried the emulator, and it appears to work fine with the CAS definition of the series in the function app. The only problem I can see is that the slope does not get computed (returns NaN - - not a number). Other features work as expected, e.g., root, or area between two x values.

Well, then it doesn't make sense at all
I'm lost with your issue... are you sure you copied the whole formula without any "spelling" mistake while removing the end of the formula? I remember I copied it once with an error and it couldn't calculate but wouldn't return any specific error message.

[Helge Gabert]

Attached are my emulator screenshots:

   

   

   

Works fine!
S/N 7382000058004 (replacement unit, directly from HP)

[jte]

(06-02-2014 07:46 AM)Tugdual Wrote:  I tried something more drastic this morning.
I just took the f(x) definition and replaced all 'x' with 'X' and put it straight into F1(X) defintion. Same problem! We can now eliminate CAS.

I've just tried scrolling, zooming etc. in Function's plot view after defining F1(X) as f(X) in Function's symb view with f defined in the CAS screen. After a minute or so I didn't experience any crashes or lock-ups. The plotting speed could certainly be improved.

Have you tried the above after starting with a clean slate? Perhaps there are some corrupted data structures in your calculator's memory. (While not giving you a completely clean slate, issuing the CAS "restart" command [WARNING: this will purge your CAS definitions] followed by shift+Esc and then a full reboot (shift + On [you might want to wait a tiny bit to ensure that the calculator state has been saved] and then On + Symb) will get you a good part of the way there.)

[Tugdual]

(06-02-2014 09:55 PM)jte Wrote:  I've just tried scrolling, zooming etc. in Function's plot view after defining F1(X) as f(X) in Function's symb view with f defined in the CAS screen. After a minute or so I didn't experience any crashes or lock-ups. The plotting speed could certainly be improved.

Did you try to get some inertia on scroll by pulling rapidly to the left or the right?
When I zoom with center of axis remaining at the center it looks ok. As soon as I give it an impulsion, I see the drawing scrolling, then blank area (curve barely or no longer visible) and after 2 or 3 very slow refreshs it is frozen.
If you check the Num table, you'll see that there are a number of NaN or huge values. This could be another reason while the redraw fails.

(06-02-2014 09:55 PM)jte Wrote:  Have you tried the above after starting with a clean slate? Perhaps there are some corrupted data structures in your calculator's memory. (While not giving you a completely clean slate, issuing the CAS "restart" command [WARNING: this will purge your CAS definitions] followed by shift+Esc and then a full reboot (shift + On [you might want to wait a tiny bit to ensure that the calculator state has been saved] and then On + Symb) will get you a good part of the way there.)

I just tried all what you said. I did more:
Using On + Apps + Esc as documented by Cyrille -> no improvement
Using the CFO menu, I "formatted C:" funny name, definitely a DOS legacy :-). This supposedly totally cleaned the memory.
I found the calculator a little faster after that but still crashing.
Anyway, as I mentioned in an earlier post, I replaced the f(X) call by the actual definition of the function so that CAS is no longer involved. Same result...

This is becoming mental, to me this is a bug, I didn't have the issue on the former version of the OS.