Hacker News new | past | comments | ask | show | jobs | submit login

> However, the framebuffer driver has been extracted from the GUI service

Extracted? Last time I look at it, they were inserting a patch in the main (and only) binary of the proprietary GUI via LD_PRELOAD, and then all other programs which wanted to draw _anything_ to the screen would basically _directly alter the memory_ of the proprietary GUI binary in order to modify its internal framebuffer. You have to do that whether you want to integrate with the main GUI or not -- since you need this hack to draw anything to the screen.

Literally every other Remarkable update so far has killed all of this.

> Most components are FOSS.

The usual thing: most 3rd party components which they are forced to open anyway are FOSS. Anything that may even be remotely interesting is NOT FOSS. I do not understand why people try to claim this as a positive, cause it isn't.




I wouldn't look at the mouth of a gifted horse. Their closed source binary is not huge.

Also I don't worry about updates. Why would I connect my notebook to the Internet anyway? I just don't install them.

I guess it's a matter of time before fully open source display drivers are available and I can do my part too to make it happen.

When it happens I will have an ownable very nice, thin and responsible e-notebook. It will never be connected to the Internet and I will patch it myself if I will happen to want a new feature.

Maybe you are right and it could be more FOSS than it is, but in a world where you don't really own anything valuable anymore I am so elated about the hackability of this device that I'll take a bit of tinkering with binary files.


You can disable automatic updates, too. This way you can still connect.


xochitl (the rM GUI) doesn't need to be run when other GUI programs are running. I don't fully understand the way this works but for anyone interested, here is the link for the framebuffer workaround: https://github.com/ddvk/remarkable2-framebuffer

On rM1 it was possible to write to the framebuffer directly, on rM2 it is not exposed as a device file directly.


The link you are sharing explicitly mentions you need xochitl running with the LD_PRELOAD hack, which is what becomes "the framebuffer server". In fact,

> To do their job, both the client and the server need to know the address at which a number of functions reside in the Xochitl binary. These addresses change from one release to the next.

This is absurdly fragile. In fact, it's even worse than I was guessing :/.


Huh, I really had the wrong idea... Thanks.

I'm still very happy with the device but I'll keep that in mind when recommending it to others.




Consider applying for YC's Spring batch! Applications are open till Feb 11.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: