Wizzup | Does anyone know how fremantle rotates its input device (touchscreen) when the device orientation changes? | 11:25 |
---|---|---|
Maxdamantus | Wizzup: my guess would be using randr. | 12:30 |
Maxdamantus | Wizzup: `xrandr --output LCD --rotate right` seems to work for me. | 12:30 |
Wizzup | Maxdamantus: right, but what component does that? | 12:30 |
Maxdamantus | Oh, hildon-desktop iirc | 12:30 |
Wizzup | That rotates the screen, I don't think that rotates input | 12:30 |
Wizzup | That is usually done via xinput | 12:30 |
Maxdamantus | The input seems to be rotated when I do that too. | 12:31 |
Wizzup | yes, but this is not 'normal' on Linux | 12:31 |
Maxdamantus | (I've disabled the rotation functionality in hildon-desktop) | 12:31 |
Maxdamantus | What do you mean? xrandr is how I've always rotated screens in Linux. | 12:32 |
Wizzup | yes, but it won't rotate your input devices for you, at least not touch screens | 12:32 |
Maxdamantus | haven't really used any other devices with touchscreens, but I would expect Xorg will know to transform the touchscreen inputs appropriately. | 12:32 |
Wizzup | It doesn't, I can assure you | 12:33 |
Wizzup | that's why stuff like this exists https://gist.github.com/mildmojo/48e9025070a2ba40795c | 12:33 |
Wizzup | Maybe they patched X to do that or something, but it doesn't work like this on X in that last 10 years | 12:36 |
Wizzup | (I've dealt with lots of touchscreen devices, and they're separate entities from the screen, and there is not necessarily a way to tell X that they belong together | 12:36 |
Wizzup | So usually they have to be rotated separately | 12:36 |
* Wizzup back later | 12:37 | |
sixwheeledbeast | The way I remember it all windows have an extra flag that permits rotation and the whole window is redrawn by hildon. It's not a rotation as such everything is flexible and just redrawn. There wasn't much to enable it from the CSSU guys back in the day, Nokia had purposely not made it a thing as not everything would be able to play nice. | 13:41 |
Wizzup | sixwheeledbeast: right, that is for windows, but not the input layer | 13:46 |
Wizzup | I know how that works, but that's not what I mean | 13:46 |
Wizzup | basically, if you take any modern linux, and a device with a touchscreen, if you rotate the screen with xrandr, the touch device won't follow implicitly | 13:47 |
Wizzup | so all input will be wrong until you make the touch input device match the screen | 13:47 |
Wizzup | (this is what 'xinput map-to-output' does) | 13:47 |
sixwheeledbeast | i don't really understand it in depth it's just what I recall. if the window is just redrawn instead of "rotated" then why does input layer matter? to the touchscreen nothing is different? | 13:51 |
Wizzup | the window isn't drawn rotated, the screen is rotated | 13:52 |
Wizzup | and the window properties do not relate here: they just tell hildon is it needs to force a certain screen rotation | 13:53 |
Wizzup | (I know, because I just finished porting them to leste qt 5) | 13:53 |
sixwheeledbeast | oh not sure then | 14:12 |
sixwheeledbeast | Maybe something in the xserver-xomap metapackage has some answers? that was a package modified by the community for rotation support | 14:16 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!