Eink Framebuffer Update Trigger

[jmseight]

Hi,

Does anyone have the man or help pages for eips? What commands are there?

Thanks,
James

[geekmaster]

eips is a custom amazon command that we do not have source for. You can do just "eips" and then try all the options, or better yet, look at all the ways it is used in the startup scripts in the kindles.

It can display text (eips x y "text"), and images (eips -g file.png), and can even scroll the text display (eips -z 2 38). It can also display a pretty pattern (eips -p), and clear the screen (eips -c). Adding -f makes it do a full flash update (eips -f -c, eips -f -g, etc.)...

To just refresh changes to the framebuffer you can do eips '' (as seen in the "algorithmic art script" thread).

Combining -f with other options is really limited to K4 and Touch. On K3 you are better not combining it with other options.

[jmseight]

Thanks GM. I tried all the commands you listed and works.

Best Regards,
James

[geekmaster]

Check out examples showing how I abused eips beyond its usefulness in a script in this new "tinysh - tiny limited onscreen shell" thread:
https://www.mobileread.com/forums/sho...d.php?t=173657

[pirl8]

Originally Posted by geekmaster

I did notice that "dd" seems to be more reliable for copying from the framebuffer than cat (I do not know why)

It could be because by default cat uses buffering, while dd doesn't.

[knc1]

Check if the copy of dd you are using supports the seek= and skip= options.

Those would give you access to the last (or other selected) line.
skip= for skipping the un-wanted lines on a fb read.
seek= to seek past the un-changed lines on a fb write.

[geekmaster]

Originally Posted by knc1

Check if the copy of dd you are using supports the seek= and skip= options.

Those would give you access to the last (or other selected) line.
skip= for skipping the un-wanted lines on a fb read.
seek= to seek past the un-changed lines on a fb write.

Huh?

If you look at almost any script I have published that uses the framebuffer, I have been using dd skip and seek parameters, so apparently they *do* work. You can check if they are supported by typing "dd" with no parameters, to see usage. I have never yet seen any dd anywhere (even limited busybox versions) without both of those parameters. They are the main critical parameters (along with bs, count, if, and of) that distinguish "dd" from "cat".

And if what I said was that unclear, let me restate it: "dd" works great, but binary files copied with "cat" do not make good copies in all cases. I was just warning people to use dd instead. The busybox version of "cat" in some kindles may have a bug that makes it process embedded special characters even when its output is redirected. It seems that any time I try to reduce my text output word count, I get these confusing replies.

[knc1]

Originally Posted by geekmaster

Huh?

Perhaps a bit of context would help you out:

Originally Posted by novo

EDIT: GOT IT!
every line is count=20, so to only delete first line is:
dd if=/dev/fb0 bs=608 count=20 > /mnt/us/fb0.raw # save framebuffer

Originally Posted by knc1

Check if the copy of dd you are using supports the seek= and skip= options.

Those would give you access to the last (or other selected) line.
skip= for skipping the un-wanted lines on a fb read.
seek= to seek past the un-changed lines on a fb write.

See: You aren't the only poster on this forum I reply too.

[geekmaster]

Originally Posted by knc1

Perhaps a bit of context would help you out:
...
See: You aren't the only poster on this forum I reply too.

Yes, that missing context would have saved my poor little brain from all that confusion. Your post followed mine, so with no embedded quote snippets to which you were repying, it looked like you were replying to me.

[varnie]

May I ask you a question - what does the Home button use to clear the screen? is it 'eips -c' actually? I just discovered that 'eips -c' remains many artefacts when we're running some third-party applications on the Kindle Touch such as CoolReader3.
After 6 page refreshes I invoke 'eips -c' and redraw the full screen, but numerous artefacts remain.

[geekmaster]

Originally Posted by varnie

... what does the Home button use to clear the screen? ... I invoke 'eips -c' and redraw the full screen, but numerous artifacts remain.

As I mentioned in post#17 above, "adding -f makes it do a full flash update (eips -f -c, eips -f -g, etc.)..." Last time I used this command, I noticed that the "-c and -f" option order is important, and must be swapped like this:

You can use "eips -c -f" to do a "full flash clear" (black -> white), which clears off more eink display artifacts. (I fixed post #17, but the original is shown in blue above.)

You can look in the GPL source code to find the "definitive" answer to your Home button question, but my "guess" is that the Home button (indirectly) does an ioctl system call to do a "full flash clear". The "eips" command shown above probably does the same ioctl system call.

[varnie]

Thank you, it works for my task.
Now all I need to figure out is how to tell/instruct eips not to clear the first 10-20 pixels from the top. The problem I have is that after I perform 'eips -c -f', then redraw a page completely, the header and footer areas (the two narrow strips, ~10 and ~20 pixels in height, respectively) are untouched.

[geekmaster]

Originally Posted by varnie

Thank you, it works for my task.
Now all I need to figure out is how to tell/instruct eips not to clear the first 10-20 pixels from the top. The problem I have is that after I perform 'eips -c -f', then redraw a page completely, the header and footer areas (the two narrow strips, ~10 and ~20 pixels in height, respectively) are untouched.

Instead of "eips -c -f", you could use "eips -f -g 0 1 white.png" to "full flash" display an "all white" image starting at the 2nd text row, where the image is short enough that it does not over-write the footer area.

When using eips to display images with a starting position, the column position needs to be 0, or it just horizontally clips the image instead of moving it sideways. But the row position (0-39) works fine.

Another way (which I have actually used, unlike the method above) is to define a shell variable that contains 50 ascii space chars, and then just use eips to "print" that to lines 1 to 38 (with the -f option in your case):

PHP Code:

B="          ";B=$B$B$B$B$B # 50 spaces
for i in $(seq 1 1 38);do eips -f 0 $i "$B"; done 


P.S. I personally do not like automatic full flash updates, preferring to provide a button or key to do a manual full flash update when the USER decides that the eink screen is "too dirty". In my experience, eink using fast updates behaves a lot like an "old-school" chalkboard. I taught computer programming at a junior college, and the kindle screen (even when "dirty", is MUCH better than a chalkboard that has been erased too many times since it was last washed). Here is a great example of what I mean by a display (or chalkboard) getting "too dirty" (and an EXCELLENT lecture, by the way): http://video.google.com/videoplay?do...83735068377717

[handyguy]

Newby here looking for assistance with eips;
I am using eips to display small png files with single numeric digits for use as a clock. However, the display of the file is cropped and squashed. The files are 268px x 362px. Any ideas how to resolve this?
I figure I can display 4 clock digits by adjusting the x offset for the 2-4 digits.
Thanks!

[NiLuJe]

Shameless self-promotion: All Hail FBInk!.

As for an actual answer: eips is very particular as to how it wants a PNG to be encoded in order not to mangle it (IIRC, it expects what is roughly colloquially called PNG8. See the ScreenSavers' hack cover mode for all the nitty gritty details). That's probably what's happening here.

(FBInk is much more lenient on that front).