KPV: a PDF reader for Kindle, based on muPDF, GPLv3

[asdas]

Originally Posted by hawhill

:-) No, eventually, we will get there. But the Touch is yet another beast, requiring a careful implementation of touch gestures. This will be certainly interesting, but of course more developers are very welcome to go on and try.

Thanx for comment.
This is important progress for Kindle Touch.
Most probably yours pdf viewer project will revolution for KT.
i and my KT user friends continue to wait.
More power to you and other developer friends

Originally Posted by knc1

My apologies.
I did not realize I was writing to Google Translate.

Don't mention it, please.
You are right dear friend. No problem

[geekmaster]

A "quick and dirty" (script compatible) way to read touch gestures is by tailing the Xorg log file. It works surprisingly well. Purists can look elsewhere. The problem is that it stops logging when you pause the framework, and if you leave the framework active but capture the touchscreen device in your native mode app the gesture engine gets locked out. So you end up doing your own native mode gesture recognition (which is not all that hard).

[thomass]

Originally Posted by hawhill

:-) No, eventually, we will get there.

I wish I could help you

[NuPogodi]

ATM, I'm trying to add the bmp-header to raw-screenshots. In order to solve this task for all supported Kindle models, I need real screenshots made in kindlepdfviewer by pressing Shift-P (Alternatively, you may just run a command "dd if=/dev/fb0 of=/mnt/us/anyname.raw"). I've already successfully converted raw-shots to bmps on K2 and K3, so any help with KDX(G) (and K4, K5???) will be appreciated.
1. Open kindlepdfviewer (filemanager or any reader), press Shift-P
2. Copy /mnt/us/kindlepdfviewer/screenshots/YYYYMMDDHHMMSS.raw to your PC, pack it with any suitable archiver, upload it somewhere (for examlpe, here) & send me a link.
Thanks a lot.

[EDIT] I've already obtained the requested raw-screenshot made with KDX. So, I do not need another one. The fb2bmp-conversion works properly and consumes reasonable CPU-time (~0.4s).

[hawhill]

NuPogodi, good idea, however, I think it could save screenshots as PNG if I added an interface to an already existing function in muPDF (I could also add an interface to libpng, which is compiled in, too, but libpng's API is somewhat over-abstracted). Or you might.

Anyways, the standalone tool you're heading at (a general framebuffer dump to recognized image format converter) is usable in more context, so I encourage you to go on!

[NuPogodi]

Well, i've selected bmp since it requires nearly no processing time (one has just to add a header and to inverse the raws order) and, hence, relatively fast (~0.15-0.17s @ K2 and K3). The bmp-size is certainly too large, but one may compress it by gzip- or bzip2-compressors available in any Linux.
My first attempts to convert 4bpp to common 8bpp for 600x800 pixels took muuuch more CPU-time (~2.5s @ K3). I suspect that png-compression would continue even longer (especially for 824x1200 KDX).
But png-output test is nevertheless "versuchswert"

[magarcan]

I've installed Kite and followed this installation steps:https://github.com/hwhw/kindlepdfvie...allation-Guide

This is how my kindle looks:






The problem is that when I press over KPDFViewer File Manager nothing happends, any idea?

Thanks!!!

[NuPogodi]

Originally Posted by magarcan

The problem is that when I press over KPDFViewer File Manager nothing happends, any idea?
Thanks!!!

AFAIK, kite-script commands refer to kpdf.sh which is (by default) stored in launchpad-folder. Did you copy this folder from the last kpdfviewer-release to your Kindle? If you're not going to use launchpad, you actually need to copy the only file: /launchpad/kpdf.sh
@dev: If my guess was correct, it is probably better in future to move kpdf.sh to /mnt/us/kindlepdfviewer/ and respectively edit the path in scripts - both for kite and for launchpad

[hawhill]

Yep, I agree. Or get rid of the shell script all together. Feel free to file a bug.

[tigran]

Thank you for this MOST EXCELLENT software!

I have noticed only one bug so far: the gamma correction for DjVu files doesn't work (but works fine for PDF files). When you fix this I will be able to delete all those massive PDF files from my Kindle and replace them with equivalent DjVu files (about 10x smaller in size, pixel-by-pixel identical to PDF).

The test files showing the bug are here:

http://www.bibles.org.uk/testfile.zip

The pdf file works fine, but the djvu file ignores the setting of gamma. Actually, there seems to be no page redraw after pressing +/- keys for djvu (but there is for pdf).

I would say this is the second most-important program ever written for Kindle (the first one is CoolReader and the third one is Kindle Chess, of course).

-------
Ok, the above "gamma doesn't work" is a red herring. The real bug lies elsewhere and I think we have tracked it down, thanks to NuPogodi's excellent analysis of the real cause (on the "Russian forum", the-ebook.org)!

Namely, the problem is probably due to the hardcoded rendering mode of DDJVU_RENDER_COLOR in djvu.c:handle():

ddjvu_page_render(page->page_ref, DDJVU_RENDER_COLOR, &pagerect, &renderrect, pixelformat, bb->w, imagebuffer);

In a DjVu viewer for Hanlin V3 (Lbook V3) called libdjvu I set the rendering mode depending on whether the djvu file is bitonal or not:

if (page_type == DDJVU_PAGETYPE_BITONAL)
djvu_render_mode = DDJVU_RENDER_MASKONLY;
else
djvu_render_mode = DDJVU_RENDER_COLOR;

Please consider this. I don't have the environment (arm gcc, etc.) to test this at the moment. But if there are clear instructions on building kindlepdfviewer (I'll check the README) then perhaps I will try.

Anyway, if you (hawhill, the Author) have a spare minute to try and compile the above fix (page_type is returned by ddjvu_page_get_type(djvu_page)) and see if it fixes the problem (the testfile.zip contains everything necessary to test) it would be great.

[tigran]

I left a comment in github as there is an existing issue about this:

https://github.com/hwhw/kindlepdfviewer/issues/217

[aliabbasjp]

so whens the new release?

[tigran]

Here are the improvements to the djvu support (thanks to NuPogodi for the idea of contrast change and the lua-based implementation which I simply converted to djvu code):

http://www.bibles.org.uk/kpdfviewer-...ncements.patch

There are three main changes (affecting only djvu files):

1. Added code to wait for djvulibre acknowledgement of completion of page rendering before manipulating the buffer. (i.e. safer handling of djvu files)

2. Added code to increase contrast for djvu files, linked to the same Volume +/- key as before, it is just for the majority of djvu files pressing Vol +/- would make no difference without this enhancement.

3. Allow the user to toggle the rendering mode between colour and b&w by pressing 'R' key. (currently requires re-opening the file)

[Dr.Seltsam]

I hope somebody can help me with installation, as I am little bit confused...

"You need to jailbreak and install launchpad first" sounded difficult, as I have no idea what jailbreak means and what launchpad is. So I would like to choose the other method with kite which seems to be easier.

Unfortunatley I found different instructions. The first message in this thread says

unzip the contents of the installation ZIP file in the root directory of the volume the Kindle

The current zip file contains two folders, so I need to unzip both folders with their contents (and not the contents only) in the root directory?

But on 12-08-2011 hawhill writes:

You are supposed to copy the contents into a folder "test" (sorry, still didn't manage to come up with a clever storage point...) in your document partition. Except for the contents of the "launchpad" subfolder, which you can copy into the launchpad config folder on your document partition if you want to use this integration.

I also looked into installation guide in the wiki and there it is again different:

Create a folder named "customupdates" on the Kindle and copy the zip file into it.
Firstly download the latest release from the Download page.
Unzip it into you Kindle 3, DX or 2.
If you want you can delete the "launchpad" folder.

So what is currently the right way to install kindlepdfviewer with kite?
Should there be a kindlepdfviewer folder in the root directory? Or in /documents/test? Or in /customupdates?
Or should I put the 17 files (which are in the kindlepdfviewer folder of the zip file) directly in any of the mentioned folders?

[hawhill]

@aliabbasjp: sorry, the developer team is small and I've been literally buried in work the last weeks (and even this weekend). I know that others from the team are also a bit on/off the job as we all do this in our spare time, which might sometimes be a bit scarce. That said: Hopefully soon!

@tigran: thank you, will review ASAP.

@Dr.Seltsam: that 2011 stuff is quite outdated. You can ignore it. Then there are different possible ways to install it. In all cases, the folders from the ZIP end up in the document partition of the kindle (in the root of the exported volume, _not_ the "documents" folder beneath it). I'm not so sure about that guide in the wiki. The "customupdates" approach only works with launchpad. Installing with kite is unfortunately still a bit broken, see https://www.mobileread.com/forums/sho...&postcount=577