Wp34s: word size lost?

[Cristian Arezzini]

With version 946 (svn updated a few minutes ago to 955, and I flashed the newly-updated calc.bin, but on the real calc it shows 946... Is that because only the emulator was updated with the latest builds?)
Anyway: if I go to integer mode, Hex, and set word size to 16 ("un 16" displayed); then I go back to decimal mode with H.d; then go back to integer Hex mode; the word size is reset to 64 bit ("un 64").
Is this a regression or has it always been like this? I can't remember now...

Cristian

[Marcus von Cube, Germany]

The revision number is strange. The most recent build should now be at 956.

I can't help with the word size question. We have to wait for Pauli or Walter on this one. I assume that the word size is reset when switching back to floating point. IIRC Pauli has simplified the switch between integer and floating point by just rounding the stack values instead of mimicking the 16C behavior (which I still don't understand).

[Cristian Arezzini]

Hm... I myself don't care how the stack is modified when switching back and forth (for me integer and floating are totally separate sessions), as long as it remembers my settings... I use the integer mode almost exclusively in 16 bit unsigned mode, so I'd like to avoid having to re-set it everytime I need integer! :)

Cristian

[Marcus von Cube, Germany]

I've modified Pauli's mode switching code so as not to reset the word size. I hope this did not break anything in floating point mode. I doubt it because it had always been possible to switch the word size when not in integer mode.

The MODE catalogue is now context sensitive, just like X.FCN is. :-)

Edited: 25 May 2011, 10:55 a.m.

[Marcus von Cube, Germany]

Quote:

---

Is this a regression or has it always been like this? I can't remember now...

---

I checked it with my 16C. It does the same and I know why. :-)

In integer mode, the 16C reallocates the registers depending on the word size. When switching back to floating point, the size must be adjusted to at least the width of a floating point number.

[Cristian Arezzini]

I just checked 960, and it seems to do what I wanted! Thanks! :)

Quote:

---

I checked it with my 16C. It does the same and I know why. :-)

In integer mode, the 16C reallocates the registers depending on the word size. When switching back to floating point, the size must be adjusted to at least the width of a floating point number.

---

Well... I didn't remember that because I only exclusively use the 16c in integer mode... But, then this is an occasion where "total 16C compatibility" isn't the best choice, in my opinion! :) Anyway, thanks for implementing this feature, and I hope it didn't break anything anywhere else!

Cristian