Instructions for patching firmware 4.15.12920

[geek1011]

This patch thread is for firmware 4.15.12920 (b231bfea368f4d641c17a337e80e9ee4219d55be).

Instructions for patching on Windows, Linux, or macOS with kobopatch:

1. Download the firmware from here.
2. Download and unzip kobopatch_4.15.12920.zip from here.
3. Copy the original firmware you just downloaded into the kobopatch src folder.
4. Install the firmware on your kobo BEFORE proceeding with the next steps.
5. Enable patches in the files in the src folder (or use the overrides in kobopatch.yaml to keep your options separate).
6. Run kobopatch.bat on Windows, or kobopatch.sh on Linux.
7. If the patching succeeded, a file named KoboRoot.tgz will be created in the out folder. Copy it to the .kobo folder of your device.
8. Safely remove your device.

Download

See the kobopatch - An improved patching system thread for more details.

[geek1011]

Changes to the kobopatch zip:
v51 (2019-07-17): Updated kobopatch to v0.14.0 (geek1011).
v50 (2019-06-20): New "Set font scale factor" patch (GeoffR), improved automatic build scripts (geek1011).
v49 (2019-06-09): New "ePub uniform font scale" patch (GeoffR).
v48 (2019-05-29): Updated missing 24 line spacing values fixed/adjustable margins patches (GeoffR).
v47 (2019-05-28): Initial release for 4.15.12920, 2 new TOC indentation patches (geek1011).

Missing patches:
- Disable wishlist (causes really unusual side effects)

Patches for other files:
- None yet

Notes about this firmware version:
- Parental controls is now an official feature
- New indented TOC level feature (but it's buggy and hard to see, see the new patches to fix this).

New patches in this version:
- Increase TOC level indentation (@geek1011)
- Increase TOC level indentation and fix extra indentation bug (@geek1011)
- ePub uniform font scale (@GeoffR)

[jackie_w]

Originally Posted by geek1011

New patches in this version:
- Increase TOC level indentation and fix extra indentation bug (@geek1011)

Hi, geek1011,

Thanks for your speedy work. I've done a test drive on the above new patch and it seems to work fine. There's always a BUT ... I would prefer an even bigger indent if that is possible - maybe multiply by 8 rather than 4? - if the indents are going to be space-characters. I'm assuming this is the patch line which would need to be customised but I don't have a clue what to change it to. Or is it not as simple as that?

  - ReplaceBytes: {Offset: 2, FindH: 00 20, ReplaceH: 82 00}  # shift level by 2 (multiply by 4) into r2

[geek1011]

Originally Posted by jackie_w

I would prefer an even bigger indent if that is possible - maybe multiply by 8 rather than 4? - if the indents are going to be space-characters. I'm assuming this is the patch line which would need to be customised but I don't have a clue what to change it to. Or is it not as simple as that?

You can multiply it by any power of 2. You'd want to replace that replacement with a LSL instruction like: LSL r2, r0, #x , where x is the power of 2 (2^x) to multiply it by. To make it multiply by 8, you'd want 3 to be the power, which gives you C2 00. You can compile the instruction online using http://armconverter.com/, and taking the Thumb Hex output.

[Semwize]

Originally Posted by geek1011

C2 00

It's too much, IMHO. It is better to leave your original version.

[jackie_w]

Originally Posted by Semwize

It's too much, IMHO. It is better to leave your original version.

Just to be clear, I'm not suggesting geek1011 changes the patch. I just wanted to know whether a personal customisation was possible.

@geek1011,
I didn't understand all the technicalities of your reply, but I got the gist, so thanks very much. I'll give it a try when I have a bit more time later today.

[Semwize]

Originally Posted by jackie_w

Just to be clear, I'm not suggesting geek1011 changes the patch. I just wanted to know whether a personal customisation was possible.

I understood it, of course.

Just tried both options and concluded that C2 00 too much, ugly. But this is of course individual.

[jackie_w]

For anyone trying to decide whether to enable the new TOC patch 'Increase TOC level indentation and fix extra indentation bug'. Here are some screencaps for comparison purposes.

• left: patch disabled (indent = 2, top level indented)
• middle: patch enabled (indent = 2-squared, top level not indented)
• right: customised patch enabled (indent = 2-cubed, top level not indented)

ETA: I should have mentioned that these screencaps are from a Forma. Bigger indents on a 6" screen might not be ideal.

[Michal Jancik]

Thank you for PATCH ! Excellent work!

[GeoffR]

Originally Posted by jackie_w

I would prefer an even bigger indent if that is possible - maybe multiply by 8 rather than 4? - if the indents are going to be space-characters. I'm assuming this is the patch line which would need to be customised but I don't have a clue what to change it to. Or is it not as simple as that?

Maybe this would work? The replacement value can be any number 0-255.

Increase TOC level indentation:
  - Enabled: yes
  - PatchGroup: TOC Indentation
  - Description: |
      The space indented Table of Contents levels are quite hard to see, as the
      difference is almost too small to easily notice.
  - ReplaceBytes: {Offset: 0x8B2FD6, FindH: 42 00 4F F0 20 01, ReplaceH: 02 22 42 43 20 21}
  # Replacement value:
  - ReplaceInt: {Offset: 0x8B2FD6, Find: 2, Replace: 6}

[geek1011]

Originally Posted by GeoffR

Maybe this would work? The replacement value can be any number 0-255.

Increase TOC level indentation:
  - Enabled: yes
  - PatchGroup: TOC Indentation
  - Description: |
      The space indented Table of Contents levels are quite hard to see, as the
      difference is almost too small to easily notice.
  - ReplaceBytes: {Offset: 0x8B2FD6, FindH: 42 00 4F F0 20 01, ReplaceH: 02 22 42 43 20 21}
  # Replacement value:
  - ReplaceInt: {Offset: 0x8B2FD6, Find: 2, Replace: 6}

That should work, but then I wouldn't be able to fit the bugfix in there too.

[jackie_w]

Originally Posted by GeoffR

Maybe this would work? The replacement value can be any number 0-255.

Increase TOC level indentation:
  - Enabled: yes
  - PatchGroup: TOC Indentation
  - Description: |
      The space indented Table of Contents levels are quite hard to see, as the
      difference is almost too small to easily notice.
  - ReplaceBytes: {Offset: 0x8B2FD6, FindH: 42 00 4F F0 20 01, ReplaceH: 02 22 42 43 20 21}
  # Replacement value:
  - ReplaceInt: {Offset: 0x8B2FD6, Find: 2, Replace: 6}

I really like the customisability (is that a word?) of this, but I do like geek1011's fix to have top-level items not indented (which seems like a bug to me, but calling it that seems to have created a bit of tension earlier today).

Having both would be great, if it was possible.

ETA: I see geek1011 got in first.

[jackie_w]

Fat fingers, please ignore

[GeoffR]

Originally Posted by geek1011

That should work, but then I wouldn't be able to fit the bugfix in there too.

I think this would work, but have not tested. it uses the fact that r7 = sp+8 which might change in the future.

Spoiler:

Increase TOC level indentation and fix extra indentation bug:
  - Enabled: yes
  - PatchGroup: TOC Indentation
  - Description: |
      The space indented Table of Contents levels are quite hard to see, as the
      difference is almost too small to easily notice. In addition, this fixes
      the bug where all TOC entries are indented by 1 level extra.
  - BaseAddress: 0x8B2FD2
  # Clear out existing instructions with NOPs
  - ReplaceBytes: {Offset: 0, FindH: 07 F1 0C 0B, ReplaceH: 00 20 00 20} # put string address in fp
  - ReplaceBytes: {Offset: 4, FindH: 42 00, ReplaceH: 00 20}             # shift level by 1 (multiply by 2) into r2
  - ReplaceBytes: {Offset: 6, FindH: 4F F0 20 01, ReplaceH: 00 20 00 20} # set r1 to 0x20 (space char)
  - ReplaceBytes: {Offset: 10, FindH: 58 46, ReplaceH: 00 20}            # put fp in r0
  # Put new instructions
  - ReplaceBytes: {Offset: 0, FindH: 00 20, ReplaceH: 40 1E}             # decrease level by 1
  - ReplaceBytes: {Offset: 2, FindH: 00 20, ReplaceH: 02 22}             # load multiplier into r2
  - ReplaceBytes: {Offset: 4, FindH: 00 20, ReplaceH: 42 43}             # multiply level, result into r2
  - ReplaceBytes: {Offset: 6, FindH: 00 20, ReplaceH: 20 21}             # set r1 to 0x20 (space char) 
  - ReplaceBytes: {Offset: 8, FindH: 00 20, ReplaceH: 05 A8}             # put string address (sp+20==r7+12) into r0
  - ReplaceBytes: {Offset: 10, FindH: 00 20, ReplaceH: 83 46}            # save r0 into fp
  # Replacement value (multiplier):
  - ReplaceInt: {Offset: 2, Find: 2, Replace: 6}

[geek1011]

Originally Posted by GeoffR

I think this would work, but have not tested. it uses the fact that r7 = sp+8 which might change in the future.

Spoiler:

Increase TOC level indentation and fix extra indentation bug:
  - Enabled: yes
  - PatchGroup: TOC Indentation
  - Description: |
      The space indented Table of Contents levels are quite hard to see, as the
      difference is almost too small to easily notice. In addition, this fixes
      the bug where all TOC entries are indented by 1 level extra.
  - BaseAddress: 0x8B2FD2
  # Clear out existing instructions with NOPs
  - ReplaceBytes: {Offset: 0, FindH: 07 F1 0C 0B, ReplaceH: 00 20 00 20} # put string address in fp
  - ReplaceBytes: {Offset: 4, FindH: 42 00, ReplaceH: 00 20}             # shift level by 1 (multiply by 2) into r2
  - ReplaceBytes: {Offset: 6, FindH: 4F F0 20 01, ReplaceH: 00 20 00 20} # set r1 to 0x20 (space char)
  - ReplaceBytes: {Offset: 10, FindH: 58 46, ReplaceH: 00 20}            # put fp in r0
  # Put new instructions
  - ReplaceBytes: {Offset: 0, FindH: 00 20, ReplaceH: 40 1E}             # decrease level by 1
  - ReplaceBytes: {Offset: 2, FindH: 00 20, ReplaceH: 02 22}             # load multiplier into r2
  - ReplaceBytes: {Offset: 4, FindH: 00 20, ReplaceH: 42 43}             # multiply level, result into r2
  - ReplaceBytes: {Offset: 6, FindH: 00 20, ReplaceH: 20 21}             # set r1 to 0x20 (space char) 
  - ReplaceBytes: {Offset: 8, FindH: 00 20, ReplaceH: 05 A8}             # put string address (sp+20==r7+12) into r0
  - ReplaceBytes: {Offset: 10, FindH: 00 20, ReplaceH: 83 46}            # save r0 into fp
  # Replacement value (multiplier):
  - ReplaceInt: {Offset: 2, Find: 2, Replace: 6}

I think it should work. That's a really clever way of doing things. I didn't think to look for ways to make the offset calculation shorter, and I didn't know MOV could be made into a 16-bit instruction by using SP. I'll try it later tonight.