KhiCAS, le moteur formel Giac/Xcas pour toutes TI-Nspire

[Ti64CLi++]

Dans le logiciel l'extension .tns n'est pas nécessaire pour apparaitre, uniquement pour transférer

[Adriweb]

A propos de la brightness sur CX II :

13:59:30 <@Vogtinato> 0x90130018 is 0xFF when it's on and 0xFFFF when off
13:59:37 <@Vogtinato> 0x90130014 has the actual brightness.

https://hackspire.org/index.php?title=M ... s_on_CX_II contient désormais un peu plus d'infos qu'avant.

[parisse]

je viens d'essayer de mettre 0x90130014 a 225, dans un premier temps ca semble marcher, mais ensuite il y a plein de petits points qui s'allument et ca finit par ne plus etre noir du tout...

[Ti64CLi++]

je viens d'essayer de mettre 0x90130014 a 225, dans un premier temps ca semble marcher, mais ensuite il y a plein de petits points qui s'allument et ca finit par ne plus etre noir du tout...

J'avais eu le même comportement en faisant quelques tests sur la luminosité, mais je ne suis pas sur que ce soit très bon pour l'écran. Ca me faisait un "point chaud", qui devenait de plus en plus noir...

[Vogtinator]

I just played around with 0x90130014 and -18 and writing values between 0 and 225 to 0x90130014 works as expected. Doing the same to 0x90130018 has a similar effect, so they're most likely both involved in the brightness, probably some PWM parameters. I don't know what's up with the 0xFFFF values which are written there when the OS turns the backlight off yet.

[critor]

Je viens de rajouter un test pour ne pas dupliquer le suffixe .tns, ca devrait donc etre corrige.
Ca serait quand meme tellement plus simple si TI n'imposait pas l'extension tns pour apparaitre dans l'explorateur et le logiciel de transfert.

Merci beaucoup.

[parisse]

I just played around with 0x90130014 and -18 and writing values between 0 and 225 to 0x90130014 works as expected.

If I write 225 then it seems at first that it works (the screen is full black), but after about 20 or 30 seconds, some spots appear and after two ot three minutes the screen is not dark anymore.
BTW, I don't know how you are playing with the I/O registers, I have added some commands in the KhICAS shell to do that:
16=>=> will force hexadecimal output of integers
read32(address) will return the value of a read of 4 bytes (1 dword) at address, e.g. read32(0x90130018)
read32(address,n) will return a list n address/dword (starting at address).
write32(address,value) will write value at address.

[Vogtinator]

I just played around with 0x90130014 and -18 and writing values between 0 and 225 to 0x90130014 works as expected.

If I write 225 then it seems at first that it works (the screen is full black), but after about 20 or 30 seconds, some spots appear and after two ot three minutes the screen is not dark anymore.

Well, 225 is the value which the OS writes for the lowest brightness level. So it's not off and whatever is on the LCD will still be shown. If you turn off the LCD controller, you'll see the usual effect of the LCD losing its state due to missing refresh cycles. Even 255/0xFF isn't fully off, so doing what the OS does (0xFFFF into both, maybe some bits elsewhere?) would be necessary to achieve that.

BTW, I don't know how you are playing with the I/O registers, I have added some commands in the KhICAS shell to do that:

A minimal custom program with basically the exact same functionality as what you have there.

[parisse]

Setting both to 0xffff seems to work, thanks for the suggestion. But there is some flickering after restoring both registers to their initial value, could that damage the hardware? Probably not, but I think it's safer to keep KhiCAS as it is now (i.e. display a warning if the user presses ctrl-on).

[Wistaro]

J'ignore si c'est le comportement souhaité, mais dès que je lance KhiCas, j'ai un écran noir avec juste une bande de pixels blanc à droite qui s'affiche, pendant 30 - 40 secondes.

C'est normal?