Kobo Patcher

[Anak]

`Freedom to advanced fonts control`

This patch works for both .ttf and .otf fonts files when reading kepubs or saved articles to Pocket (the Access Renderer). But when reading regular epubs (Adobe) TypeGenius only seem to work with .ttf font files and not with .otf.

Can someone confirm this?

BTW, can someone ask on the Russian the-ebook.org forum if Metazoa has searched or looked for possible hacks in the libadobe.so file?

[icallaci]

Originally Posted by Danger

Thought I would post my settings for the page # display. Been playing around with this for a few days now and finally have it where I want it for my Glo:

{\n\tmin-height: 30px;\n\tmax-height: 30px;\n\tqproperty-footerMargin: 30;\n\theight: 30px;\n\tpadding-bottom: 4px;\n\tfont-size: 20px;\n}

Danger, is this for FW 3.1.0? And is padding-bottom the space below the text or the space below the page number? I can't get this to work on my Aura HD with FW 3.1.0. Yet.

[forster01]

Change the lines as in the this post and working just fine, just pached

Originally Posted by jackie_w

Try changing the end part from

`* {\n\theight: 30px;\n\tfont-size: 25px;\n}\n\n`

to

`ReadingFooter {\n\tmin-height: 50px;\n\tmax-height: 50px;\n\tqproperty-footerMargin: 20;\n}\n\n* {\n\tfont-size: 26px;\n}\n\n[localeName="ja_JP"] {\n\tfont-family: A-OTF Gothic MB101 Pr6N;\n}`

The Footer height can be increased/decreased by changing the two 50px values. The size of the 'Page x of y text' can be inc/dec by changing the font-size: 26px value.


That was probably me. It is relevant when using the ReadingFooter suggestion above.

[met67]

Would it be possible to use this to re-enable reading stats for side loaded books?
I would try it myself, but the last time I did some reversing & patching was at least 15 years ago! :-)

[PeterT]

Originally Posted by met67

Would it be possible to use this to re-enable reading stats for side loaded books?
I would try it myself, but the last time I did some reversing & patching was at least 15 years ago! :-)

The criteria for displaying stats is that the book is in the .kEpubs directory and (I think) has a name that is an uuid.

Yes it is conceivable that the location could be found where these comparisons are made, but some one needs to work on it.

Our experience so far has been that even if a book is named correctly and stored correctly that the next sync with Kobo via desktop or WiFi will delete the book.

[Danger]

Originally Posted by icallaci

Danger, is this for FW 3.1.0? And is padding-bottom the space below the text or the space below the page number? I can't get this to work on my Aura HD with FW 3.1.0. Yet.

Yes this is for 3.1.0, sorry should have said that. The padding-bottom is for the space below the Page # of #.

How exactly is it not working? I found whenever I made any changes to the footer I had to tap the Aa to open & close the fonts settings and then page forward/back. After that the footer would display with the proper settings. I found the best way to judge what the footer space looked like was to have a book open that had the cover displayed full screen.

[icallaci]

Originally Posted by Danger

Yes this is for 3.1.0, sorry should have said that. The padding-bottom is for the space below the Page # of #.

How exactly is it not working? I found whenever I made any changes to the footer I had to tap the Aa to open & close the fonts settings and then page forward/back. After that the footer would display with the proper settings. I found the best way to judge what the footer space looked like was to have a book open that had the cover displayed full screen.

When I use your settings (which are for the Glo) on my Aura HD, the Page # of # is off the bottom of the screen. When I try to insert only the padding-bottom statement into the patch (no matter what numbers I use for the other settings), the KoboRoot.tgz file is created with no error messages, but the Aura HD just ignores the patch and reverts to its default spacing.

Below is the relevant part of what I am currently using, which works ok. I would prefer slightly more space under the Page # of #, but I can live with this. I'll probably keep playing with it to see if I can get padding-bottom to work.

{\n\tmin-height: 40px;\n\tmax-height: 70px;\n\tfont-size: 26px;\n}

Edit: I did try your suggestion of opening the Aa menu and turning a page, but it didn't change anything.

[GeoffR]

Originally Posted by icallaci

When I use your settings (which are for the Glo) on my Aura HD, the Page # of # is off the bottom of the screen. When I try to insert only the padding-bottom statement into the patch (no matter what numbers I use for the other settings), the KoboRoot.tgz file is created with no error messages, but the Aura HD just ignores the patch and reverts to its default spacing.

I've noticed that kpg.exe leaves the replacement string unterminated when using replace_xor. This means that if you replace say `Original` with `New` then the patched file will contain the string `Newginal` instead of `New`.

I think this might be causing some of the problems people have with getting the footer patch to work, as the end of the old string might override the new string depending on what the new combined string looks like.

You could try adding some space to the end of the replacement string to overwrite more of the original string. Or maybe adding some junk characters to the end of the new string will cause the CSS interpreter to stop when it can't parse the junk.

If you are running Linux you could try to apply the patch with my program in this post to see if there is a difference, as it does terminate the string when the replacement is shorter than the original.

[DNSB]

Originally Posted by GeoffR

I've noticed that kpg.exe leaves the replacement string unterminated when using replace_xor. This means that if you replace say `Original` with `New` then the patched file will contain the string `Newginal` instead of `New`.

I think this might be causing some of the problems people have with getting the footer patch to work, as the end of the old string might override the new string depending on what the new combined string looks like.

I checked the output libnickel.so.1.0.0 file from running kpg.exe v1.5 and the entry for the footer patch was what is in the kpg.conf file with the terminating } followed by the pair of LFs.

The bytes at 13FA550 (fw 3.1.0) were:

2A 20 7B 0A 09 68 65 69 67 68 74 3A 20 32 35 70 78 3B 0A 09 66 6F 6E 74 2D 73 69 7A 65 3A 20 32 30 70 78 3B 0A 7D 0A 0A 68 65 69 67 68 74 3A 20

* {\n\theight: 25px;\n\tfont-size: 20px;\n}\n\nheight:

The only time I would see the termination as being a problem is if someone did not maintain the structure of the the replacement block with the terminating }\n. Admittedly that was why I originally started converting the ReadingFooter block to a human readable form to be able to see and avoid some of those errors.

Regards,
David

[GeoffR]

Originally Posted by DNSB

Just to disagree, I checked the output libnickel.so.1.0.0 file and the entry for the footer patch was what is in the kpg.conf file with the terminating } followed by the pair of LFs.

The bytes at 13FA550 (fw 3.1.0) were:

2A 20 7B 0A 09 68 65 69 67 68 74 3A 20 32 35 70 78 3B 0A 09 66 6F 6E 74 2D 73 69 7A 65 3A 20 32 30 70 78 3B 0A 7D 0A 0A 68 65 69 67 68 74 3A 20

I mean the replacement string is not terminated by a null character. Would the two LF's be enough to tell the CSS interpreter to stop processing the rest of the string? I don't think so, as the original footer string is structured like

ReadingFooter { ... }\n\n ReadingFooter[deviceCodeName="dragon"] { ... }\n\n

which I think means that the first clause is applied to all devices, then the second is read but only applied for AuraHD devices. So if the replacement string was the right length to overwrite the first clause but leave the second intact, and there was no terminating null to end the string, then the second clause from the original string would still be applied to AuraHD devices, possibly overriding the first clause.

Edit: I added code to my own program to test, if I terminate the replacement string with a null then I get binary output that differs from what kpg.exe produces, but if I don't terminate the string then I get output identical to kpg.exe.

[GeoffR]

Ok I have a test case for fw 3.1.0 (d1b2f00933) that demonstrates the problem with the way kpg.exe is applying the footer patch:

Spoiler:

<Patch>
patch_name = `Custom reading footer style`
patch_enable = `yes`
#
replace_xor_58 = 13FA550, `ReadingFooter {\n\tmin-height: 70px;\n\tmax-height: 70px;\n\tqproperty-footerMargin: 40;\n}\n\nReadingFooter[deviceCodeName="kraken"], \nReadingFooter[deviceCodeName="phoenix"] {\n\tmin-height: 82px;\n\tmax-height: 82px;\n\tqproperty-footerMargin: 15;\n}\n\nReadingFooter[deviceCodeName="dragon"] {\n\tmin-height: 120px;\n\tmax-height: 120px;\n\tqproperty-footerMargin: 20;\n}\n\n* {\n\tfont-size: 16px;\n}\n\n[deviceCodeName="kraken"], \n[deviceCodeName="phoenix"] {\n\tfont-size: 20px;\n}\n\n[deviceCodeName="dragon"] {\n\tfont-size: 26px;\n}\n\n[localeName="ja_JP"] {\n\tfont-family: A-OTF Gothic MB101 Pr6N;\n}`, `ReadingFooter {\n\tmin-height: 30px;\n\tmax-height: 30px;\n\tqproperty-footerMargin: 15;\n}\n\n`
</Patch>

When I apply the above patch with kpg.exe, the new footer is ignored on my Glo. However if I apply the patch using my own program which terminates the replacement string with a null character, then the new footer works as expected on my Glo.

The reason the kpg.exe applied patch is being ignored is that the end of the original string is still being read after the replacement, and so the clause in the original with the [deviceCodeName="kraken"] overrides the new footer, and I just get the original footer.

Edit: I should note for anyone trying this out that my program uses a slightly different syntax in the patch: instead of "replace_xor_58 = ...", write "encode_xor8 = 58" followed by "replace_string = ...".

[GeoffR]

Here is a demonstration of the problem with the way kpg.exe applies the footer patch that doesn't require any other program to test. I have added the missing string terminator manually with a replace_bytes line:

Spoiler:

<Patch>
patch_name = `Custom reading footer style test case`
patch_enable = `yes`
#
replace_xor_58 = 13FA550, `ReadingFooter {\n\tmin-height: 70px;\n\tmax-height: 70px;\n\tqproperty-footerMargin: 40;\n}\n\nReadingFooter[deviceCodeName="kraken"], \nReadingFooter[deviceCodeName="phoenix"] {\n\tmin-height: 82px;\n\tmax-height: 82px;\n\tqproperty-footerMargin: 15;\n}\n\nReadingFooter[deviceCodeName="dragon"] {\n\tmin-height: 120px;\n\tmax-height: 120px;\n\tqproperty-footerMargin: 20;\n}\n\n* {\n\tfont-size: 16px;\n}\n\n[deviceCodeName="kraken"], \n[deviceCodeName="phoenix"] {\n\tfont-size: 20px;\n}\n\n[deviceCodeName="dragon"] {\n\tfont-size: 26px;\n}\n\n[localeName="ja_JP"] {\n\tfont-family: A-OTF Gothic MB101 Pr6N;\n}`, `ReadingFooter {\n\tmin-height: 30px;\n\tmax-height: 30px;\n\tqproperty-footerMargin: 15;\n}\n\n`
#
# add a terminating null (xor 0x58) to the above 86-character string: 
#replace_bytes = 13FA5A6, 0A, 58
</Patch>

To test: Apply with the replace_byte line commented out and the patch is ignored on my Glo. But uncomment the replace_bytes line and the patch works as expected. The replacement byte is an encoded null (zero XOR 0x58) applied at the location following the last '\n' character in the replacement string.

This test case is for firmware 3.1.0 (d1b2f00933) on Glo/Aura/AuraHD (I don't think this particular footer would have a problem on other devices).

[icallaci]

GeoffR, you are definitely on to something. Your code works on my Aura HD, but the Page Number font size is 16pt, which is way too small for me. But I have no idea which bytes to replace with a null because my string is longer than 86 bytes. Here is my string:

replace_xor_58 = 13FA550, `ReadingFooter {\n\tmin-height: 70px;\n\tmax-height: 70px;\n\tqproperty-footerMargin: 40;\n}\n\nReadingFooter[deviceCodeName="kraken"], \nReadingFooter[deviceCodeName="phoenix"] {\n\tmin-height: 82px;\n\tmax-height: 82px;\n\tqproperty-footerMargin: 15;\n}\n\nReadingFooter[deviceCodeName="dragon"] {\n\tmin-height: 120px;\n\tmax-height: 120px;\n\tqproperty-footerMargin: 20;\n}\n\n* {\n\tfont-size: 16px;\n}\n\n[deviceCodeName="kraken"], \n[deviceCodeName="phoenix"] {\n\tfont-size: 20px;\n}\n\n[deviceCodeName="dragon"] {\n\tfont-size: 26px;\n}\n\n[localeName="ja_JP"] {\n\tfont-family: A-OTF Gothic MB101 Pr6N;\n}`, `ReadingFooter {\n\tmin-height: 40px;\n\tmax-height: 70px;\n\tfont-size: 26px;\n\tqproperty-footerMargin: 15;\n}\n\n`

I assume qproperty-footerMargin is the margin below (or above AND below?) the Page Number footer?

Using the code I show above, what should I use instead of replace_bytes = 13FA5A6, 0A, 58? Thanks--I think you have figured out why some code works while other code doesn't.

Edit: the only change from your original code is the very last part, where I added an entry for font-size for my Aura HD: {\n\tmin-height: 40px;\n\tmax-height: 70px;\n\tfont-size: 26px;\n\tqproperty-footerMargin: 15;\n}

[GeoffR]

Originally Posted by icallaci

Using the code I show above, what should I use instead of replace_bytes = 13FA5A6, 0A, 58? Thanks--I think you have figured out why some code works while other code doesn't.

Edit: the only change from your original code is the very last part, where I added an entry for font-size for my Aura HD: {\n\tmin-height: 40px;\n\tmax-height: 70px;\n\tfont-size: 26px;\n\tqproperty-footerMargin: 15;\n}

I count your replacement string to be 104 characters long (counting '\n' and '\t' as one character each), so try adding the terminating null as

replace_bytes = 13FA5B8, 3B, 58

[tshering]

Originally Posted by GeoffR

I count your replacement string to be 104 characters long (counting '\n' and '\t' as one character each), so try adding the terminating null as

replace_bytes = 13FA5B8, 3B, 58

I guess another way to achieve this (without counting) is adding \x00 to the replacement string.