New firmware BUG found on Solve app related to digit grouping

[alexzkter]

Hi, first of all thanks for the new firmware, I sucessfully updated the calculator on Win 8.1 64bits enterprise lightning fast.

Problem comes when I realized all stored formulas are partially gone. Creating all user variables again made it appear most of all the stored formulas. The ones showing an error were recovered by editing them (deleting one of the variables in the formula and inserting it back again to be more precise)

So here's the real problem:
when Digit grouping is the standard 123456.789 , in other words...with no extra space or symbol in between the 3&4 numbers, I get a proper solution, as shown below:


for the Mechanical Energy Balance, which is the following simple equation:


But...when you choose some of the rest of options as default digit grouping, for example, the 123 456.789 grouping, you get a "Bad guess(es)" solution:


Take a look at the density value (ρ=1000 shown as 1 000)
changing the value to 100 will make it throw a proper answer.

Second ¿bug?:

I created the V1 and V2 variables, but they're not showing up on user variables list, see:
I'm guessing they're system reserved variables that have to do with vectors so not a bug.

Third:
On Home settings, kinetic scrolling is available too, which makes very difficult to scroll page without skipping pages. Kinetic scrolling here should be disabled, I just checked on the emulator as well, and it's difficult too not to go from page 1 to 3.



And last but not least:
I find the inclusion of undo/redo softkey buttons very convenient. I know it's been requested before, but I experienced the need of such option when accidentally deleting formulas. Or at least a confirmation window asking whether I'm sure I want to delete a formula or not (and a way to disable what I just mentioned)

[ndzied1]

Just Re-Did the update per Tim's instructions and this bug is still there.
I have it for sure when I select the number grouping with decimal point for the fraction separator and space for the 1000's separator.

[Tim Wessman]

#1 Thanks, will put on the list to investigate.

#2 That is actually the adv grapher variables. V1-V0

#3 Seems quite easy to me here. Guess I just go slower maybe? Anyone else have any thoughts?

[ndzied1]

Here is a clue as to what is going on.

I have the following equation in my solver:


If I do a show with the digit grouping standard I get what I would expect:


But if I use the digit grouping with the space in it, and do a Show I see this:


Looks like the parsing in the solver app is assuming the space makes the 1 and 456 different numbers.

[Tim Wessman]

There are basically two different numerical representations for a number - the "edit" option which has no digit grouping, and the "non-edit" one that does. Looks like this is a case where the wrong one is getting used. The fact that it impacts the solver there though is quite surprising.

[Didier Lachieze]

A similar issue related to digit grouping is occuring in program debug mode when entering input data.

For example, with this simple program:

Code:

EXPORT TEST(N)
BEGIN
MSGBOX(N);
END;

If you launch it with “Debug” and enter a number >999 with a space in the digit grouping you get something like:

   

Then after OK there is an error message: “Undefined result”:

   

There is no error if there is no space in the digit grouping.

[Michael Lopez]

I've also experienced similar problems with the SOLVE application since the firmware upgrade :-(.

Cheers,

MIchael