You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
About a year ago I forked this project to create a quick-and-dirty modified version which merges two displays into one rather than splitting one into two. Recently I noticed this repository is active again, and I figured it was worth suggesting that this functionality be rolled into the original codebase.
This feature is immensely useful to owners of the IBM T221 monitor (I am one such owner, but there are many others). Originally released in 2001, it was the world's first (almost) 4K retina display. Many people, myself included, still consider it to be the best (it has built-in hardware color management and a 48hz native refresh rate, two features which have yet to be replicated ). Unfortunately, since it was built before any connectors were capable of carrying enough bandwidth, it must be attached to the computer with two separate dual-link DVI connections, or four single link connections. As such, the computer reports it as multiple identical independent monitors. For years it was possible to override this behavior when using the nvidia binary driver by manually specifying the xinerama configuration in Xorg.conf, but nvidia has taken that ability away and doesn't seem likely to give it back, and their equivalent Mosaic system is restricted to those with very expensive quadro cards (not even re-flashed fake quadros will do the trick unfortunately). Thus the only way to merge the two halves of the monitor correctly is to rewrite the output of xrandr. Info on the monitor can be found on wikipedia: https://en.wikipedia.org/wiki/IBM_T220/T221_LCD_monitors
I have found that not all applications receive the right information. Some still detect two separate monitors - in particular, some web video players (perhaps those which require flash?) and some games. I haven't yet determined a firm pattern, but I suspect SDL may be one culprit. As I said, my modification was quick, and done without a thorough understanding of xrandr or xorg. Clearly it is incomplete in some way.
The text was updated successfully, but these errors were encountered:
If your fork is old, it probably does not include a fake xinerama yet, on
which some applications rely. Try to copy the code from the bottom of the
file & create a link to libxinerama. Another possibility is that they use
libxcb, which has an entirely different api for randr; I'm currently
working on that.
As for merging outputs, yes, creating arbitrary outputs on the fb would
certainly be quite useful. I'll think about how to integrate this into the
configurable version!
About a year ago I forked this project to create a quick-and-dirty modified version which merges two displays into one rather than splitting one into two. Recently I noticed this repository is active again, and I figured it was worth suggesting that this functionality be rolled into the original codebase.
This feature is immensely useful to owners of the IBM T221 monitor (I am one such owner, but there are many others). Originally released in 2001, it was the world's first (almost) 4K retina display. Many people, myself included, still consider it to be the best (it has built-in hardware color management and a 48hz native refresh rate, two features which have yet to be replicated ). Unfortunately, since it was built before any connectors were capable of carrying enough bandwidth, it must be attached to the computer with two separate dual-link DVI connections, or four single link connections. As such, the computer reports it as multiple identical independent monitors. For years it was possible to override this behavior when using the nvidia binary driver by manually specifying the xinerama configuration in Xorg.conf, but nvidia has taken that ability away and doesn't seem likely to give it back, and their equivalent Mosaic system is restricted to those with very expensive quadro cards (not even re-flashed fake quadros will do the trick unfortunately). Thus the only way to merge the two halves of the monitor correctly is to rewrite the output of xrandr. Info on the monitor can be found on wikipedia:
https://en.wikipedia.org/wiki/IBM_T220/T221_LCD_monitors
Additionally, I understand some new 4k displays have the same issue:
http://www.phoronix.com/scan.php?page=news_item&px=MTU3NDU
There may be other use cases as well, such as tiled projector setups or video walls.
My version of fakexrandr can be found here, though it's probably not of any use:
https://bitbucket.org/jackdoerner/fakexrandr
I have found that not all applications receive the right information. Some still detect two separate monitors - in particular, some web video players (perhaps those which require flash?) and some games. I haven't yet determined a firm pattern, but I suspect SDL may be one culprit. As I said, my modification was quick, and done without a thorough understanding of xrandr or xorg. Clearly it is incomplete in some way.
The text was updated successfully, but these errors were encountered: