Plato: A document reader for PDF, ePUB, DjVu, etc.

[The_Immortal]

Originally Posted by CH23

Is there a particular reason you want these formats instead of another format?

Yup. I have a lot of doc/docx files and they are hardly convertable to other formats.

[DezReads]

Originally Posted by baskerville

One last question: what's the initial rotation value?

The image does match my device and description.
The initial Rotate Value when i started Plato was 90, so the ui was upside down.

[jcn363]

Originally Posted by The_Immortal

Guys, could you tell me please:

1. Does the recent Plato support MS Office Word formats .rtf/.doc/.docx) and F2B? Will be happy to see their support!

2. Is there instuction how to build the recent version of Plato under Windows? I found the instruction for Ubuntu only.

Thank you!

1. Read
2. VM

Good luck.

[The_Immortal]

Could someone please provide the ready binaries of the latest release?

Thank you!

[jcn363]

Originally Posted by The_Immortal

Could someone please provide the ready binaries of the latest release?

Here and github.
Good luck.

[baskerville]

I've released 0.6.1.

The screen rotation should work fine now, on the Aura H₂O and the Aura H₂O Edition 2 Version 2. Please check that:

• The initial rotation is correct.
• You can successfully switch from 90 to 270, or 0 to 180.
• The rotation gesture rotates the screen in the correct direction.

I stumbled upon a segmentation fault while using 0.6.0 but was unable to reproduce ever since. As a consequence, I have enabled core dumps on both my devices. If you want to help track down the bug, add the following line in plato.sh, before the line that starts Plato:

ulimit -c unlimited

And use the release binary with debug symbols and the corresponding version of libmupdf.so enclosed in plato-0.6.1-debug.zip. If Plato crashes, you will find a file named core alongside the binary: post it here zipped.

List of other changes:

• Add default styles for sub and sup.
• Parse BISAC Subject Headings present in dc:subject.
• Update thirdparty dependencies: MuPDF (master branch at cf039bea), libpng 1.6.36, jbig2dec 0.15, freetype 2.9.1, HarfBuzz 2.1.0.

[mobama]

Originally Posted by Ken Maltby

There is a plato.sh that you would have after moving the plato folder into .adds. You could try running that from KSM.

How do you navigate to it in KSM and run it?

[tshering]

Originally Posted by mobama

How do you navigate to it in KSM and run it?

Install this patch of KSM.

[mobama]

Originally Posted by tshering

Install this patch of KSM.

I chose Install Partial Update. (No idea which one I should have chosen, because there was no mention of it.)

The question remains: Where is Plato in KSM? Where do I see it? How do I open it?

Edit: I'm passionately burning to try this thing, but I find the instructions disjointed and incomplete, practically illegible. It's still a long way until an average person can try this app.

[tshering]

Originally Posted by mobama

I chose Install Partial Update. (No idea which one I should have chosen, because there was no mention of it.)

The question remains: Where is Plato in KSM? Where do I see it? How do I open it?

Edit: I'm passionately burning to try this thing, but I find the instructions disjointed and incomplete, practically illegible. It's still a long way until an average person can try this app.

You need KSM 9 with all updates and the above mentioned patch installed. You also have to copy the folder "plato" (which you can find in the package provided by Baskerville) into the folder ".adds". After restarting the device you should then see a new entry "start plato" on the first page of KSM.

Edit: More recent version of plato do not contain the folder plato inside the archive. You have to create the folder yourself under .adds, and then copy the content of the plato archive into it.

[NiLuJe]

@baskerville: Still a no go on the inverted switches . As I mentioned earlier, the ioctl do fire and return *successfully*, but to no effect.

My strace build doesn't decode fbio ioctls content, so I can't vouch for the accuracy of the ioctl content. I can either whip up a quick C test-case to double-check, of if you can add logging around the ioctl calls, that'd be helpful too .

[baskerville]

Originally Posted by NiLuJe

@baskerville: Still a no go on the inverted switches . As I mentioned earlier, the ioctl do fire and return *successfully*, but to no effect.

Thanks, well, I'm just resending the ioctl call until the read value changes (and stop after 3 failures).

I will try to send an intermediate value instead.

Originally Posted by NiLuJe

]If you can add logging around the ioctl calls, that'd be helpful too .

I'm logging the written and read value, so you should see messages of the form Framebuffer rotation: X -> Y in info.log where X is the written value and Y the read value.

[NiLuJe]

@baskerville: Oops, missed the logging, my bad .

I was under the impression only the ioctl return value was effectively checked, since rotation() only pulls the "cached" rotate value, it doesn't do a new GET. But I only skimmed through that, so I may have missed it .

I'll do another round of tests .

[baskerville]

Originally Posted by NiLuJe

I was under the impression only the ioctl return value was effectively checked, since rotation() only pulls the "cached" rotate value, it doesn't do a new GET.

The ioctl call mutates self.var_info.

I've attached a test binary that takes a different approach: if writing n fails, then I write (n+1)%4 and finally n.

[NiLuJe]

@baskerville: We (roughly) came to the same conclusion .

Turns out it really, really doesn't like it when swapping between inverted layouts in the *same* orientation layout.

The ioctls happily go through, and return successfully, but the rotation doesn't actually happen.

Attached a C testcase .

TL;DR: if the value to be set is equal to the current rotate value, go through an intermediary rotation to do an initial portrait/landscape swap (i.e., (n+1)%4).

EDIT: And here's what it spouts on my H2O:

Variable fb info: 1080x1440, 32bpp @ rotation: 1 (Clockwise, 90°)
Fixed fb info: ID is "mxc_epdc_fb", length of fb mem: 6684672 bytes & line length: 4352 bytes

FB_ROTATE_UR to FB_ROTATE_CCW, +1 increments
Setting rotate to 0 (Upright, 0°)
Rotate is now 2 (Upside Down, 180°)
Variable fb info: 1440x1080, 32bpp @ rotation: 2 (Upside Down, 180°)
Fixed fb info: ID is "mxc_epdc_fb", length of fb mem: 6684672 bytes & line length: 5760 bytes
Setting rotate to 1 (Clockwise, 90°)
Rotate is now 3 (Counter Clockwise, 270°)
Variable fb info: 1080x1440, 32bpp @ rotation: 3 (Counter Clockwise, 270°)
Fixed fb info: ID is "mxc_epdc_fb", length of fb mem: 6684672 bytes & line length: 4352 bytes
Setting rotate to 2 (Upside Down, 180°)
Rotate is now 0 (Upright, 0°)
Variable fb info: 1440x1080, 32bpp @ rotation: 0 (Upright, 0°)
Fixed fb info: ID is "mxc_epdc_fb", length of fb mem: 6684672 bytes & line length: 5760 bytes
Setting rotate to 3 (Counter Clockwise, 270°)
Rotate is now 1 (Clockwise, 90°)
Variable fb info: 1080x1440, 32bpp @ rotation: 1 (Clockwise, 90°)
Fixed fb info: ID is "mxc_epdc_fb", length of fb mem: 6684672 bytes & line length: 4352 bytes

FB_ROTATE_UR to FB_ROTATE_CCW, +2 increments
Setting rotate to 0 (Upright, 0°)
Rotate is now 2 (Upside Down, 180°)
Variable fb info: 1440x1080, 32bpp @ rotation: 2 (Upside Down, 180°)
Fixed fb info: ID is "mxc_epdc_fb", length of fb mem: 6684672 bytes & line length: 5760 bytes
Setting rotate to 2 (Upside Down, 180°)
Rotate is now 2 (Upside Down, 180°)
Variable fb info: 1440x1080, 32bpp @ rotation: 2 (Upside Down, 180°)
Fixed fb info: ID is "mxc_epdc_fb", length of fb mem: 6684672 bytes & line length: 5760 bytes

FB_ROTATE_CW to FB_ROTATE_CCW, +2 increments
Setting rotate to 1 (Clockwise, 90°)
Rotate is now 3 (Counter Clockwise, 270°)
Variable fb info: 1080x1440, 32bpp @ rotation: 3 (Counter Clockwise, 270°)
Fixed fb info: ID is "mxc_epdc_fb", length of fb mem: 6684672 bytes & line length: 4352 bytes
Setting rotate to 3 (Counter Clockwise, 270°)
Rotate is now 3 (Counter Clockwise, 270°)
Variable fb info: 1080x1440, 32bpp @ rotation: 3 (Counter Clockwise, 270°)
Fixed fb info: ID is "mxc_epdc_fb", length of fb mem: 6684672 bytes & line length: 4352 bytes

FB_ROTATE_UR to FB_ROTATE_CCW, +2 increments, intermerdiary rota if ==
Setting rotate to 0 (Upright, 0°)
Rotate is now 2 (Upside Down, 180°)
Variable fb info: 1440x1080, 32bpp @ rotation: 2 (Upside Down, 180°)
Fixed fb info: ID is "mxc_epdc_fb", length of fb mem: 6684672 bytes & line length: 5760 bytes
Intermerdiary rotation...
Setting rotate to 3 (Counter Clockwise, 270°)
Rotate is now 1 (Clockwise, 90°)
Variable fb info: 1080x1440, 32bpp @ rotation: 1 (Clockwise, 90°)
Fixed fb info: ID is "mxc_epdc_fb", length of fb mem: 6684672 bytes & line length: 4352 bytes
Requested rotation
Setting rotate to 2 (Upside Down, 180°)
Rotate is now 0 (Upright, 0°)
Variable fb info: 1440x1080, 32bpp @ rotation: 0 (Upright, 0°)
Fixed fb info: ID is "mxc_epdc_fb", length of fb mem: 6684672 bytes & line length: 5760 bytes

FB_ROTATE_CW to FB_ROTATE_CCW, +2 increments, intermerdiary rota if ==
Setting rotate to 1 (Clockwise, 90°)
Rotate is now 3 (Counter Clockwise, 270°)
Variable fb info: 1080x1440, 32bpp @ rotation: 3 (Counter Clockwise, 270°)
Fixed fb info: ID is "mxc_epdc_fb", length of fb mem: 6684672 bytes & line length: 4352 bytes
Intermerdiary rotation...
Setting rotate to 0 (Upright, 0°)
Rotate is now 2 (Upside Down, 180°)
Variable fb info: 1440x1080, 32bpp @ rotation: 2 (Upside Down, 180°)
Fixed fb info: ID is "mxc_epdc_fb", length of fb mem: 6684672 bytes & line length: 5760 bytes
Requested rotation
Setting rotate to 3 (Counter Clockwise, 270°)
Rotate is now 1 (Clockwise, 90°)
Variable fb info: 1080x1440, 32bpp @ rotation: 1 (Clockwise, 90°)
Fixed fb info: ID is "mxc_epdc_fb", length of fb mem: 6684672 bytes & line length: 4352 bytes

EDIT²: I wasn't convinced we could trust PUT_V not to mangle rotate without an extra GET_V to confirm, but, yeah, you're right, that works .