Kindle VNC viewer (native app, GPLv2)

[knc1]

Originally Posted by MagnusW

- - - - -
PS: I'm posting on a couple of threads like this, sorry if some of you might see this message other places as well.

Ref: https://www.mobileread.com/forums/faq...ing_guidelines
Section 3.

[edddh]

Hi Magnus,

I am really glad to hear about such project.

This idea would strongly interest a minority of people who suffer dramatically from the blue peak of white leds



This peak causes several indispositions, such as difficulty of concentration and insomnia.

Although as i said this problem strongly affects only a minority of people, it is likely that most other people have effects such as bad sleeping and disrupted daily activities without recognizing or even accepting it.

In fact unfortunately there is strong resistance to accept that white leds are bad for health because the majority of people apparently sleep well after using a laptop and do not even believe the remaining people when they assert it. However a massive number of scientific studies now are emerging fast.

Using Kindle as a secondary monitor works, however it is an unsatisfactory and incomplete solution. The main issue is NOT the whole screen refresh time as many people superficially assert. This because most users desperately require a secondary monitor of course not for playing video games, but for using much more basic productive software like word, excel, and eventually for browsing the Internet.

In my opinion the main issue is then JUST the cursor pointer reactivity.

I believe the solution to this issue is not really technological, but constructive and implementative:

1) having a refresh of only the portion involved by the cursor pointer

2) having a touch screen that is properly connected with the OS in order to allow most interactive operations like
- changing the position of the cursor pointer
- sliding and moving windows and lists

3) having a simple color schema in the OS (without useless frills like transparencies, shadows and so on) to allow easy usage of productive applications

4) good adaptation of the OS to the exact screen resolution of the e-ink

5) easy operations for connecting the secondary monitor and having the color LCD switched off immediately, without having to work with it for even few minutes before switching to the e-ink monitor.

PS: I would be pleased if you get connected with me also through Linkedin (https://www.linkedin.com/in/novation).

[MagnusW]

Hi edddh,

Wow that was a wonderful and insightful reply to my question, thanks a lot!

I'd love to see the references for the research you are mentioning, it seems really interesting and relevant.

I would also ask if you could share how you would use such a monitor? For work? Any thoughts on screen set-up? (Next to main screen, number of screens, etc.) As to understand your solution need in more detail..

Your comments on implementation trade-offs and workarounds are duly noted and I agree that many of the "technological" downsides can be offset.

I hope it's OK we keep it in this thread as the topic still is e-ink monitor related.


Thanks,
- Magnus

[Graphite]

I am glad to see that an e-ink screen monitor is being developed and I would be very interested in being among the first to buy it, or even in using a prototype for testing/feedback.

I use the computer a lot with an LCD screen and several months ago I was looking for a newer monitor that would be easier on the eyes. After much research I selected a good LED screen that does not have the rapid flickering (PWM) and can be dimmed right down (I find many monitors too bright even on the lowest brightness setting). However I found that I was significantly more fatigued with the new screen. As edddh explained it is very likely to be the blue light issue. I've had to return the LED screen and go back to the old LCD, choosing "the lesser of two evils".

Aside from gaming and watching video, a significant part of my computer use is e-mail, web browsing, reading and writing. I would love to have an e-ink display that I can use for these purposes and have an easy way of switching to LCD for gaming/video. Perhaps a keyboard shortcut that can disable one screen and enable another. Or potentially having a displayport IN and OUT on the e-ink display with a switch button at the front. Pressing this button could power off the e-ink and redirect the signal to the OUT port, effectively changing the monitor on the same cable and forcing the OS to quickly reconfigure from one display to another.

I imagine that I would use both screens fairly equally, perhaps the e-ink even a bit more than LCD. I do not have severe symptoms with the LCD screen but I am very keen to have something that is easier on the eyes. For users like me, fast and easy switching between two screens would be an important feature.

Another suggestion is to ensure that this monitor does not generate electro-pollution to also make it the ultimate choice for electro-hypersensitive people. There are four main types of electro-pollution:
- Electromagnetic Fields (EMF)
- Electric Fields (ELF)
- Radiofrequency Radiation (RFR)
- "Dirty Electricity"
The main concern in this case would be the power supply used in the monitor. I'd be happy to explain how to test for all of these things.

[hawhill]

While I understand that you really love to discuss this matter and while I consider this matter important, I like to ask you to stop hijacking this thread and open a new discussion instead. Also, I don't see how this would be Kindle specific, so this whole subforum might not be the place. You can then edit your posts to include a link to where you continue the discussion. Thank you!

[MagnusW]

hawhill, you're right, thank you for a fair feedback. I will start a new thread specifically on this in the coming days. Will post a link so people can continue the discussion there.

Sorry for the noise on your thread, and best of luck going forward.

- Magnus

[dfect]

Great app!

Is it possible to let the kindle vmc client to show mouse pointer (if this is sent from server, I use Vine Server for mac)?

Best Regards
dfect

[hawhill]

It doesn't do a software mouse pointer. If the server sends the pointer (in addition to its position), it is shown. Unfortunately, I don't know how to configure Vine for such behaviour. The other way around is not implemented in kindlevncviewer.

[tclazier]

Hello,

I was wondering if there is a way to convert my kindle paperwhite to a secondary monitor for Android. Thanks

[PieMan597]

You could root your android device, install a vnc server, and then set the kindle and the vnc server to connect over WiFi.

[botm]

Hello,
I'm new here. I wasn't able to read the whole thread so sorry if this was already answered.
I started using kindlevncviewer today and it works great. I'm going to use it for programming so that I don't have to look at a backlit screen. The problem is that when I type I have to wait long until the characters appear. However when I'm moving a window or something like that it updates much faster. Does it wait until a large enough number of pixels has changed or something like that? I noticed that it will update if i press a button on the Kindle keyboard. But clicking a key 2 times per seckond is not a good idea. Can it be configured to automatically update faster?
Thanks in adavance.

Also, this post was written using the kindle as a screen. Kind of difficult to write nwithout seeing the letters appear.

Edited to add:
I looked at the source. I noticed that there are commandline options. Buk kvncviewer seems to ignore them. I cloned the repository so that I can make changes to make it print information so that i know what's going on.
I wasn't able to build it. I figured out that I have to download luajit, libvncserver and zlib.
The "-m32" option is not recognized. I removed it to see what happens.
I still get errors.

b:~/kvnc/kindlevncviewer> make ARCH=arm-linux-gnueabihf
fatal: No names found, cannot describe anything.
fatal: No names found, cannot describe anything.
make -C luajit-2.0 HOST_CC="gcc" CROSS=arm-linux-gnueabihf-
make[1]: Entering directory `/home/b/kvnc/kindlevncviewer/luajit-2.0'
==== Building LuaJIT 2.0.2 ====
make -C src
make[2]: Entering directory `/home/b/kvnc/kindlevncviewer/luajit-2.0/src'
AR        libluajit.a
make[2]: arm-linux-gnueabihf-ar: Command not found
make[2]: *** [libluajit.a] Error 127
make[2]: Leaving directory `/home/b/kvnc/kindlevncviewer/luajit-2.0/src'
make[1]: *** [default] Error 2
make[1]: Leaving directory `/home/b/kvnc/kindlevncviewer/luajit-2.0'
make: *** [luajit-2.0/src/luajit] Error 2

I have no idea what to do now.

Also, when a part of the screen is updated the rightmost pixel is not updated. An off-by-one error?

ETA: Ok, I found out that I have to upgrade binutils. I have successfully built luajit, jibjpeg-turbo and zlib. I'm stuck on libvncserver now.

b:~/kvnc/kindlevncviewer> make ARCH=arm-linux-gnueabihf
fatal: No names found, cannot describe anything.
fatal: No names found, cannot describe anything.
arm-linux-gnueabihf-gcc -fPIC -shared -DLIBVNCSERVER_HAVE_LIBZ -DLIBVNCSERVER_HAVE_LIBJPEG -Ilibvncserver/libvncclient/.. -Ilibvncserver/libvncclient/../common/ -Ilibvncclient.so -Izlib/ -Ilibjpeg-turbo-1.3.0/ -Wl,-E -Wl,-rpath,'$ORIGIN' -o libvncclient.so libvncserver/libvncclient/cursor.c libvncserver/libvncclient/listen.c libvncserver/libvncclient/rfbproto.c libvncserver/libvncclient/sockets.c libvncserver/libvncclient/vncviewer.c libvncserver/libvncclient/tls_none.c libvncserver/libvncclient/../common/minilzo.c zlib/libz.so.1 libjpeg-turbo-1.3.0/.libs/libjpeg.so.62
In file included from libvncserver/libvncclient/../rfb/rfbclient.h:39:0,
                 from libvncserver/libvncclient/cursor.c:25:
libvncserver/libvncclient/../rfb/rfbproto.h:72:27: fatal error: rfb/rfbconfig.h: No such file or directory
compilation terminated.
In file included from libvncserver/libvncclient/../rfb/rfbclient.h:39:0,
                 from libvncserver/libvncclient/listen.c:39:
libvncserver/libvncclient/../rfb/rfbproto.h:72:27: fatal error: rfb/rfbconfig.h: No such file or directory
compilation terminated.
In file included from libvncserver/libvncclient/../rfb/rfbclient.h:39:0,
                 from libvncserver/libvncclient/rfbproto.c:37:
libvncserver/libvncclient/../rfb/rfbproto.h:72:27: fatal error: rfb/rfbconfig.h: No such file or directory
compilation terminated.
In file included from libvncserver/libvncclient/../rfb/rfbclient.h:39:0,
                 from libvncserver/libvncclient/sockets.c:32:
libvncserver/libvncclient/../rfb/rfbproto.h:72:27: fatal error: rfb/rfbconfig.h: No such file or directory
compilation terminated.
In file included from libvncserver/libvncclient/../rfb/rfbclient.h:39:0,
                 from libvncserver/libvncclient/vncviewer.c:32:
libvncserver/libvncclient/../rfb/rfbproto.h:72:27: fatal error: rfb/rfbconfig.h: No such file or directory
compilation terminated.
In file included from libvncserver/libvncclient/../rfb/rfbclient.h:39:0,
                 from libvncserver/libvncclient/tls_none.c:20:
libvncserver/libvncclient/../rfb/rfbproto.h:72:27: fatal error: rfb/rfbconfig.h: No such file or directory
compilation terminated.
make: *** [libvncclient.so] Error 1

Looks like rfbconfig.h should be autogenerated somehow.
ETA: I needed automake and autoconf to build the config script to generate rfbconfig.h and the makefiles. using automake or autoconf gives errors. Anlo, using the generated ./config gives errors

[hawhill]

Oh, right, rfbconfig.h is missing. Hm, that's an oversight by me. I must think about how to fix that. I thought I could play clever by just building the libvncclient sources, because the rest of the libvncserver distribution is such a hog in terms of building. And for whatever reason it isn't prepared to restrict its build to libvncclient in the upstream version.

Errrm - are there problems left which you're still facing? About everything is implemented in the Lua sources anyway, so you might get away without touching the build system at all. Also, arm-linux-gnueabihf is a hard-float ABI, and Kindles usually don't use this. Just sayin'. You might want to switch to a arm-linux-gnueabi toolchain. Or - as I said - not use a toolchain at all and just fiddle with the Lua files.

Nice that finally some other eyes are looking at the code.

You might be onto something with the off-by-one failure. However, at a first glance, it just looks a bit complicated, but basically right.

You're right with the delay for small updates. To be exact, that delay is 250 msecs, which are being waited for additional updates, so a bunch of small updates won't hog the device (e-ink updates are pretty slow). And it's wrongly documented as being 150 msecs :-) You can also change that value via a command line switch.

[knc1]

See the ARMhf thread for using a hard float abi (from user storage).
The archive includes LuaJit-2.0.3 also.

Hmm...
Probably not such a good idea on a DXG - that SoC doesn't have enough floating point hardware to be worth the effort.

Use a soft float toolchain (and compile options) - LuaJit will build with that, without warnings or errors.
If you do get warnings or errors - check your toolchain setup - it isn't LuaJit.

[dfect]

Is there some easy way to quit running this app (kill) from the "button-set" of the kindle dxg?

I've tried to make a command via launchpad, but it only works when not running the vnc client.

I used to kill it via usb terminal ps aux and kill but would be awesome to be able to do it direct on the kindle.

[hawhill]

/me scratches his head. I think the "home" key is intended to quit the app. (looks at code...) Yes, it is. Doesn't that work?