Kobo Patcher

[DNSB]

Originally Posted by GeorgK

Well...
What I did
1. downloaded FW 3.0.1
2. upgraded Aura HD to 3.0.1
3. downloaded Patcher V.10
4. extracted libnickel from FW 3.0.1. package
5. replaced the libnickel-file within the patcher folder (3.0.0) by the extracted 3.0.1 version
6. edited the .conf as follows:
<Patch>
patch_name = `Custom reading footer style`
patch_enable = `yes`
# Decrease footer example
replace_xor_CC = 13EA170, ...
</Patch>
7. run kgp.exe

And got the error message: Data for libnickel and kpg.conf does not match.

What's wrong with my attempt?

I would suspect that replace_xor_CC does not exist. I invented that item in line with the other replace_xor_XX routines used by Metazoa. My suggestion was that Metazoa could use a routine similar to GeoffR and set the XOR value before calling the routine to generalize it. The addresses used by the other routines would not match either so hopefully, they were all commented out.

One other item of interest is that the modified libnickel.so.1.0.0 from 3.0.0 will work with 3.0.1. At least in limited testing after reading a comment here that suggested someone had done so, it seemed to work.

Regards,
David

[jackie_w]

Originally Posted by DNSB

I would suspect that replace_xor_CC does not exist. I invented that item in line with the other replace_xor_XX routines used by Metazoa. My suggestion was that Metazoa could use a routine similar to GeoffR and set the XOR value before calling the routine to generalize it. The addresses used by the other routines would not match either so hopefully, they were all commented out.

This is funny I believed you when I thought you suggested
replace_xor_CC = 13EA170, `ReadingFooter etc... for fw 3.0.1 should work. Now it seems you didn't believe yourself Have faith, sir!

Anyway that setting for the Footer patch does indeed run through the v10 Patcher kpg.exe without error, see log below

Spoiler:

Kobo Patch Generator ver. 1.5 (c) 2013 MZ

Loading config file ...
OK

Checking config syntax ...
OK

Loading file for patching ...
OK

Applying patches ...
	Applying patch [Custom reading footer style] ...
	OK
	Applying patch [My line spacing values] ...
	OK
	Applying patch [Custom fonts sizes] ...
	OK
	Applying patch [Custom left & right margins] ...
	OK
	Applying patch [Fix page breaks bug] ...
	OK
OK

Writing result file ...
OK

Creating Kobo update ...
OK

but... I have to confess I haven't actually updated to fw3.0.1 to try it for real because I'm still hoping the 3.0.1 settings for `Freedom to advanced fonts control` will turn up.

[DNSB]

Originally Posted by jackie_w

This is funny I believed you when I thought you suggested
replace_xor_CC = 13EA170, `ReadingFooter etc... for fw 3.0.1 should work. Now it seems you didn't believe yourself Have faith, sir!

Anyway that setting for the Footer patch does indeed run through the v10 Patcher kpg.exe without error, see log below

Actually, I was suggesting what I thought the change would look like. Since I haven't extracted the python code from the executable as has been suggested was possible, I hadn't realized that the routine was already generalized so I had not tried the change. Live and learn.

Regards,
David

[jackie_w]

I choose to believe in your prophetic powers

[tshering]

Originally Posted by DNSB

Live and learn.

Maybe listening to people who try to answer your questions might help, cf, this and that.

[DNSB]

Originally Posted by tshering

Maybe listening to people who try to answer your questions might help, cf, this and that.

I did realize from your comments that it was a general purpose routine. Not having looked at the source code, I did not realize that name used to reference the routine defined the XOR value to be used.

Regards,
David

[tshering]

DNSB,

sorry, I should have expressed it more precisely.

[GeoffR]

Recent firmware has 15 positions on the line spacing slider. 5 positions are set from immediate values in code which the original line spacing patch bypasses, and the other 10 positions are set from a table of values which can be changed with the patch. The result is that after applying the original patch there are only 10 positions on the line spacing slider.

I have extended the patch to allow 4 of the code-injected values to be modified rather than bypassed. This gives a slider with 14 positions, although to simplify the patch the 4 code injected values must be chosen from a limited table of values.

This is the extended line spacing patch patch for firmware 3.0.1 (644485e229) (tested on my Glo).

Spoiler:

<Patch>
# For firmware 3.0.1 (644485e229)
patch_name = `My 14 line spacing values` 
patch_enable = `yes`
# Bypass code-injected values 1.0, 1.2, 1.5, 2.0, 3.0
#replace_bytes = 7493AE, 26 F6 AE EA, 00 46 00 46
#replace_bytes = 7493FC, 26 F6 86 EA, 00 46 00 46
#replace_bytes = 749422, 26 F6 74 EA, 00 46 00 46
replace_bytes = 749458, 26 F6 58 EA, 00 46 00 46
#replace_bytes = 7494B4, 26 F6 2A EA, 00 46 00 46

# Table A
# E0 = 0.50,   E1 = 0.53125,  E2 = 0.5625,  E3 = 0.59375
# E4 = 0.625,  E5 = 0.65625,  E6 = 0.6875,  E7 = 0.71875
# E8 = 0.75,   E9 = 0.78125,  EA = 0.8125,  EB = 0.84375
# EC = 0.875,  ED = 0.90625,  EE = 0.9375,  EF = 0.96875
# F0 = 1.0,    F1 = 1.0625,   F2 = 1.125,   F3 = 1.1875
# F4 = 1.25,   F5 = 1.3125,   F6 = 1.375,   F7 = 1.4375
# F8 = 1.5     ...
#
# Table B                      ...            E7=0.725
# E8 = 0.75625,  E9 = 0.7875,  EA = 0.81875,  EB = 0.85
# EC = 0.88125,  ED = 0.9125,  EE = 0.94375,  EF = 0.975
# F0 = 1.0125,   F1 = 1.075,   F2 = 1.1375,   F3 = 1.2
# F4 = 1.2625    F5 = 1.325    F6 = 1.3875,   F7 = 1.45
# F8 = 1.5125    ...

# First code-injected value 1.0
# Replace byte with a value from table A:
replace_bytes = 7493A2, F0, E8

# Second code-injected value 1.2
# Replace byte with a value from table B:
replace_bytes = 7493F2, F3, EB

# Third code-injected value 1.5
# Replace byte with a value from table A:
replace_bytes = 749418, F8, ED

# Last code-injected value 3.0
# Select Table A:
replace_bytes = 7494A8, C4 F2, C3 F6
replace_bytes = 7494AB, 03, 73
# Replace byte with a value from table A.
replace_bytes = 7494AA, 08, F4

# Other values for replace
#                       1.0   0.75
replace_float = 7494F8, 1.05, 0.765
replace_float = 749500, 1.07, 0.8
replace_float = 749508, 1.1,  0.83
#                       1.2   0.85
replace_float = 749510, 1.35, 0.865
#                       1.5   0.90625
replace_float = 749518, 1.7,  0.925
replace_float = 749520, 1.8,  0.95
#                       2.0   ----
replace_float = 749528, 2.2,  1.0
replace_float = 749530, 2.4,  1.05
replace_float = 749538, 2.6,  1.1
replace_float = 749540, 2.8,  1.175
#                       3.0   1.25
</Patch>

This is the extended line spacing patch patch for firmware 3.0.0 (0cbc3626b3) (not tested).

Spoiler:

<Patch>
# For firmware 3.0.0 (0cbc3626b3)
patch_name = `My 14 line spacing values` 
patch_enable = `yes`
# Bypass code-injected values 1.0, 1.2, 1.5, 2.0, 3.0
#replace_bytes = 817CEE, 57 F5 16 EE, 00 46 00 46
#replace_bytes = 817D3C, 57 F5 EE ED, 00 46 00 46
#replace_bytes = 817D62, 57 F5 DC ED, 00 46 00 46
replace_bytes = 817D98, 57 F5 C0 ED, 00 46 00 46
#replace_bytes = 817DF4, 57 F5 92 ED, 00 46 00 46

# Table A
# E0 = 0.50,   E1 = 0.53125,  E2 = 0.5625,  E3 = 0.59375
# E4 = 0.625,  E5 = 0.65625,  E6 = 0.6875,  E7 = 0.71875
# E8 = 0.75,   E9 = 0.78125,  EA = 0.8125,  EB = 0.84375
# EC = 0.875,  ED = 0.90625,  EE = 0.9375,  EF = 0.96875
# F0 = 1.0,    F1 = 1.0625,   F2 = 1.125,   F3 = 1.1875
# F4 = 1.25,   F5 = 1.3125,   F6 = 1.375,   F7 = 1.4375
# F8 = 1.5     ...
#
# Table B                      ...            E7=0.725
# E8 = 0.75625,  E9 = 0.7875,  EA = 0.81875,  EB = 0.85
# EC = 0.88125,  ED = 0.9125,  EE = 0.94375,  EF = 0.975
# F0 = 1.0125,   F1 = 1.075,   F2 = 1.1375,   F3 = 1.2
# F4 = 1.2625    F5 = 1.325    F6 = 1.3875,   F7 = 1.45
# F8 = 1.5125    ...

# First code-injected value 1.0
# Replace byte with a value from table A:
replace_bytes = 817CE2, F0, E8

# Second code-injected value 1.2
# Replace byte with a value from table B:
replace_bytes = 817D32, F3, EB

# Third code-injected value 1.5
# Replace byte with a value from table A:
replace_bytes = 817D58, F8, ED

# Last code-injected value 3.0
# Select Table A:
replace_bytes = 817DE8, C4 F2, C3 F6
replace_bytes = 817DEB, 03, 73
# Replace byte with a value from table A.
replace_bytes = 817DEA, 08, F4

# Other values for replace
#                       1.0   0.75
replace_float = 817E38, 1.05, 0.765
replace_float = 817E40, 1.07, 0.8
replace_float = 817E48, 1.1,  0.83
#                       1.2   0.85
replace_float = 817E50, 1.35, 0.865
#                       1.5   0.90625
replace_float = 817E58, 1.7,  0.925
replace_float = 817E60, 1.8,  0.95
#                       2.0   ----
replace_float = 817E68, 2.2,  1.0
replace_float = 817E70, 2.4,  1.05
replace_float = 817E78, 2.6,  1.1
replace_float = 817E80, 2.8,  1.175
#                       3.0   1.25
</Patch>

To change the code injected values just choose a value from the appropriate table and replace the byte at the end of the line after the "# Replace byte with value from table ..." comments. If you don't modify the patch you will get the following 14 line spacing values: {0.75, 0.765, 0.8, 0.83, 0.85, 0.865, 0.90625, 0.925, 0.95, 1.0, 1.05. 1.1, 1.175. 1.25}

[GeoffR]

I've made a patch to fix a little bug that is annoying me in firmware 2.10.0/3.0.0/3.0.1. The bug occurs when I long-press a book in my Glo's library and select the "View Details" menu item. The page that is displayed has a book cover image, but instead of using the N3_LIBRARY_FULL image file that usually already exists on the device, a new one named AndroidBookLoadTablet_Aspect is generated or downloaded. This causes two problems:

1. If the book is a Kobo-bought kepub, then the new image has to be downloaded from the server. This means that if I have wifi turned off I don't see a cover image on the details page, just a box with the title. The new image is not saved after it is downloaded, so it will have to be downloaded again once it leaves the cache.

2. If the book is a sideloaded epub then an ugly new cover image is generated from the first page of the book, instead of using the nice image I sent from Calibre when I sideloaded the book. The new cover is saved though.

The patch works by changing the file name of the new image to match the old one, so when the device checks whether the new file needs to be generated/downloaded it finds the old one already exists and uses that.

Patch for 3.0.1 (644485e229) (tested on my Glo):

Spoiler:

<Patch>
patch_name = `Avoid generating/downloading redundant cover image`
patch_enable = `yes`
replace_string = 10F91A4, `AndroidBookLoadTablet_Aspect`, `N3_LIBRARY_FULL`
</Patch>

Patch for firmware 3.0.0 (0cbc3626b3) (not tested):

Spoiler:

<Patch>
patch_name = `Avoid generating/downloading redundant cover image`
patch_enable = `yes`
replace_string = 11065D8, `AndroidBookLoadTablet_Aspect`, `N3_LIBRARY_FULL`
</Patch>

[neil74]

So is this safe for the new aura yet?

[neil74]

Ok I took the plunge on 2.10 with the new 6" aura and it seems to have worked fine.

Not tested it properly yet but have opened a book and other than the font/margins adjustment feeling just a tad slower in refreshing it seems to be working ok and the lack of giant footers is very nice indeed

[GeorgK]

Something strange happend: I patched my AuraHD 3.0.1 experimentally not to show the footer at all. Worked fine.
Fell back to decrease footer. The (decreased) footer is back but the header is gone!

How's that?

Regards
Georg

[GeoffR]

I have updated my extended line spacing patch in this post, it now gives an extra 4 line spacing values, and there are patches for firmware 3.0.0 and 3.0.1.

[neil74]

Wondering if it's possible to get the page number in the footer back? i.e make the footer just larger enough to display it? Or is it all or nothing?

Loving that the huge footer has now gone though.

[jackie_w]

Originally Posted by neil74

Wondering if it's possible to get the page number in the footer back? i.e make the footer just larger enough to display it? Or is it all or nothing?

Loving that the huge footer has now gone though.

There are 2 options for the footer patch in the kpg.conf file. One disables it completely (no page x of y text). The other reduces the footer size and keeps the page x of y.

When you run the patch program you need to have one disabled (commented out with a '#' as the first char of the line) and the other enabled (no '#').

I think the default as per Metazoa's original kpg.conf has the 'Disabled' option selected rather than the 'Decreased' option. So if you didn't swap the options before proceeding it would explain why you can't see page x of y.