I think that most developers are like myself, in that we don't really worry or care about precisely where a pop-up shows up.

Speaking from a developer's perspective, I'd rather the software application decide which screen to use, rather than the window manager. For example, I use MyPaint with the extra control dialogs opening up on a secondary screen. This leaves the entire main screen available for drawing space. By letting the software application control this, it can intelligently utilize the benefits of multiple screens.

So, my preference is to fix this issue on an application level, rather than a window manager level.

Seems odd that this does not bother more people to make it a bigger priority, it is enough to start making me consider figuring out what it would take to design a graphical interface for Linux but I imagine that is a HUGE project. It is extremely infuriating and it does indeed surprise me that more devs don't have 3+ monitor setups to get annoyed by this. I can't imagine trying to code on just one monitor, it's just not doable without constantly having to shuffle between stuff and it is super inefficient.

I have tried the Synergy route, in fact I just got off that. The problem with that route is that it is super buggy. Synergy locks up/crashes constantly, if you accidentally try to copy and paste across multiple screens, it locks up the whole PC, the browser on the Raspberry Pi (in my case that's what I used for side monitors - did not want to use power sucking regular PCs) keeps crashing... the setup requires constantly tending and can't just be left alone for more than a few hours. To make matters worse when synergy does crash I need to do a kill -9, because it stays stuck. I then tried to install the latest version to see if it would work better, but it requires too many dependencies that I could not get to work, then I gave up and re-attempted a single PC solution, and here I am. Basically back to square one. (I just remembered the Raspberry PIs are actually still running lol)

My current setup is already set so the middle one is set as primary that's not the issue, but it does not stop some apps from still sending dialogs all over the place. Heck, even menus, tool tips and other elements that have absolutely zero reason to be detached from where they were spawned.


Is there a way to run the primary monitor as local/native OS, then have some kind of VM setup that dedicates the other two so the host OS does not even see them? That could be a viable solution as at least graphic acceleration would work for the middle one, while the other two would simply be running a basic VM locally but the host would not know they exist.

It's really too bad even separate X sessions don't work properly. That has always been my plan B, but even that is shot.

Since I don't play games on this machine I could technically get away with using the nouveau driver. Anything specific to that which might help me? I can always switch to it.

What versions of synergy are you using? I use Debian stable and XFCE4 across the board, with the version of synergy in the Debian stable repositories (1.4.16). It works, and never crashes. Cut-and-paste works great. It gets kind of laggy sometimes for reasons I don't really understand, but that's why I made some taskbar shortcuts to restart the clients/server in a few clicks.

The bottom line is that you can get synergy to work reliably and lag free.

You would rather open an application on monitor A then, clicking on part of a project (for example) it opens on monitor B so, then you decide to explore files on monitor B and find something you're interested in and, when opened, it opens on monitor A in the corner maximised?
That's how Linux multi-monitor works -- how it doesn't work I'm sure people will let you know.

I don't really notice which monitor something initially opens up the first time. I just move it to where I want it, and the application remembers where I wanted it. Not all applications are necessarily so well behaved, of course, but I guess I don't use badly behaved ones often enough to really care.

That strikes me as very odd too:-

It's possible that they only run open-src drivers but I highly doubt they limit themselves to just one monitor.

I personally know a developer and he is paid exceptionally well so purchasing one or 2 brand new monitors and a good GPU to drive them should be a reasonable expense that wouldn't break their bank account.

I think (or at least what it looks like) is developers are not being paid to do this and I'm beginning to wonder if this is why Linux DE's still have this issue with multi-monitoring.

---------------- -------------------- --------------------- ---------------------
It looks like nVidia supports multi-monitoring but I'm not sure about AMD/ATI.
http://www.nvidia.com/object/nvs-graphics-cards.html

How to set up multiple monitors with nVidia:
http://nvidia.custhelp.com/app/answe...tiple-monitors

Why is it that not all application are behaved so well?

Could it be that the API is lacking something?

I would think it's because the developers of those badly behaved ones don't care to fix it and/or they have their own ideas about how it's supposed to behave and/or they honestly don't have multi-monitor setups. I don't know, like I said I don't really notice it at all.

If I could afford to spend money on it, I'd much rather have a single large 4K monitor than a cluster of lower resolution monitors. A single 50" 4K monitor gives as much space as a cluster of four 25" 1920x1080 monitors, but it's more flexible and it doesn't have black bars in the way. The sheer size would make it impractical for me to utilize more than one of them effectively...I'd have to twist my neck around like crazy. Heck, I'd probably do better with a 39" than a 50" just to reduce neck twisting.

Others may prefer to have a much higher DPI with multiple small high resolution monitors, but for my particular dev work a high DPI does not really confer a benefit.

Yeah, I looked at a handful of 50 inch monitors online and they are really pricey!
This would be worth saving up for if you can.

Sometimes they go on sale for a price low enough to almost tempt me. Right now, hhgregg seems to have a 48" Seiki 4K on sale for under $300. I'm not familiar with reviews of this model, but if it's anything like the Seiki 39" and 50", it's only capable of 30hz refresh at full 4K resolution and color reproduction is mediocre. Which sucks for 4K TV or high end gaming, but it's just peachy for software development, DBA work, document processing, etc...

Anyway, getting back to the original topic - I find that generally XFCE4 will open up a new window on the screen where my mouse cursor is, if the software doesn't specify. That seems good enough for me. It doesn't break any intuitive expectations I have, and it's easy to control where a new application opens up when I do so.

Red Squirrel:

Is XFCE 4 working for you with multiple monitors? (reasonably)

Yeah a 4k has actually crossed my mind would solve all of this, I would hope. I have a KVM though so I can use my windows machine too, not sure how 4k would behave with one of those. USB/DVI ones are hard enough to find as is let alone trying to find one that supports 4k.


I don't recall for sure since I've since clean installed, but I think it was 1.7.6. I could not use the one built into the OS repo since it was too old and I had no choice but to go from source for the Raspberry PIs and that one was newer and they all have to match. Something like that, I forget the details.

I tried to go to 1.8.2 but it just would not compile because of some new dependency they added. Something to do with openssl, but it was not openssl itself. Or think it was wrong version. Dependencies are bad enough as is, it's when it's wrong version where things get REALLY ugly and think that was the error I was getting. Something to do with a lib version.

But even if I got it to work I still had the issue of the RPI browser constantly closing on it's own. My left monitor is mostly for monitoring stuff or doing quick research that I may want to be able to reference while I'm working on main, so I tend to just leave a browser open and require it to stay open. Suppose I could use a PC based solution instead of raspberry PIs, but then it means I'm looking into 100w+ territory. The RPIs only use like 5w vs a regular PC that uses 100w at least.

If the compilation failed it should of told you why and what dependency it was.
Installing that dependency first and than try to compile should work.

I understand, BTDT-

Was the nVidia link on post #51 on how to set up multiple monitors of any help?

I know how to set it up with nvida, this is how it's setup now, it "works". The problem as I mentioned is how each app seems to behave differently, some very terribly, in how they support multi monitor. I have the centre set as Primary, yet some apps still insist on opening on the side ones, or spawning dialogs or other elements on the wrong one.

As for synergy this is the error I get:

This is a Mint 18 system fully up to date, no idea what to do about it wanting a different lib version, no way I want to start messing around with 3rd party repositories. I always end up botching my system when I do that.

Sorry to hear the applications are not opening on the monitor that you want them to.
I wish I knew how to fix it. If I find out how, you'll be the first to know.:)

From reading this output until these dependencies are met the compilation will not be successful.
Install each dependency in the order it's listed.

https://help.ubuntu.com/community/SynergyHowto


Synergy is out of Beta right?