Instructions for patching firmware 4.25.15875

[JSWolf]

Originally Posted by genoasalami

Has anyone been able to get multiple dictionaries working on the latest (November 2020) firmware? I have never been able to get this working and if anyone has step-by-step instructions, I'd appreciate it. I am pretty tech literate and comfortable with coding, but for whatever reason this has never worked for me.

Is there a maximum file size?

I suggest you ask this question for the new 4.26.x firmware.

[ps67]

Originally Posted by genoasalami

Has anyone been able to get multiple dictionaries working on the latest (November 2020) firmware? I have never been able to get this working and if anyone has step-by-step instructions, I'd appreciate it. I am pretty tech literate and comfortable with coding, but for whatever reason this has never worked for me.

Is there a maximum file size?

Since firmware 4.24.15672 no more patch is needed, You have to create a directory called "custom-dict" in the .kobo directory of the device.
and then follow the instruction by @geek1011 (AKA pgaskin) here:

https://github.com/pgaskin/dictutil/issues/14

I don't know if there is a maximum file size, I installed two custom dictionaries and they works well.

[genoasalami]

Originally Posted by ps67

Since firmware 4.24.15672 no more patch is needed, You have to create a directory called "custom-dict" in the .kobo directory of the device.
and then follow the instruction by @geek1011 (AKA pgaskin) here:

https://github.com/pgaskin/dictutil/issues/14

I don't know if there is a maximum file size, I installed two custom dictionaries and they works well.

Thanks. Does dictutil handle converting stardict to dicthtml, or do I still need to use penelope for that? Reading the documentation on dictutil it looks like it's just meant to copy the file to the kobo (I can do that in explorer, don't need a program for that) and to create new dictionaries from scratch (but I already have a dictionary I want to use).

I might be misunderstanding though, so I'd appreciate any clarification.

[geek1011]

Originally Posted by genoasalami

Thanks. Does dictutil handle converting stardict to dicthtml, or do I still need to use penelope for that? Reading the documentation on dictutil it looks like it's just meant to copy the file to the kobo (I can do that in explorer, don't need a program for that) and to create new dictionaries from scratch (but I already have a dictionary I want to use).

I might be misunderstanding though, so I'd appreciate any clarification.

It does everything except the stardict conversion. You can use PyGlossary to convert directly to a Kobo dictionary (it's based on dictutil), or you can convert it to dictutil's custom human-readable dictfile format to make manual edits and/or merge it with other dictionaries first.

Dictutil is a set of tools which can compile, decompile, unpack, pack, install, and do other things with Kobo dictionaries.

The main advantage of using dictutil to install dictionaries is that it'll handle differences between firmware versions for you.

[genoasalami]

Originally Posted by geek1011

It does everything except the stardict conversion. You can use PyGlossary to convert directly to a Kobo dictionary (it's based on dictutil), or you can convert it to dictutil's custom human-readable dictfile format to make manual edits and/or merge it with other dictionaries first.

Dictutil is a set of tools which can compile, decompile, unpack, pack, install, and do other things with Kobo dictionaries.

The main advantage of using dictutil to install dictionaries is that it'll handle differences between firmware versions for you.

Thanks for the quick reply - you're on it! Does this mean I can skip the dicthtml step and go stardict>kobo dictionary format (.df i guess)?

[geek1011]

Originally Posted by genoasalami

Thanks for the quick reply - you're on it! Does this mean I can skip the dicthtml step and go stardict>kobo dictionary format (.df i guess)?

It's the other way around . .df is my format.

[genoasalami]

Originally Posted by geek1011

It's the other way around . .df is my format.

This was a huge pain in the ass. Thank you for trying to help me. It turns out that dicthtml.zip had become corrupted, likely because of special characters which cannot be compressed (or with which some zip programs have trouble compressing).

Providing the below for people in case it is helpful to them. I know that Penelope has been abandoned, so if there's a better way to do this using your software, it would be great if you could share. Cheers!

----------------------

I fixed this by recreating dicthtml.zip using Penelope. However, Penelope will not install under windows because it requires the marisa-trie dependency which for some reason won't install either. Welcome to Python. Ultimately, I used my *nix system to get Penelope working, and as I think about this, that's what I had to do last time as well. If you're able to get python working under Windows, more power to you, but this was a mess and using *nix was faster for me, although that doesn't really help people who aren't tech savvy enough to have such a system handy.

The large file size doesn't seem to be a problem, but it does cause the Forma to chug for half a tick when you go to define a word. Not that big of a deal in my opinion, but it's worth noting.

Re the dictionary itself: the latest OED edition is the second. The third is underway according to the wiki; so, when you see a 1989 publication date, don't freak out. If you're here because you're looking for the OED, something something copyright and try google-fu.

Instructions:

1. get Penelope
2. get stardict file
3. depending upon where you happen across your stardict file, you will need to put it into a regular zip; so extract it and then zip it (literally .zip, not .7z or .tar etc); Penelope is expecting a .zip
3.

penelope -i oed1filename,oed2filename -j stardict -o oed.zip -p kobo -f en -t en

this: runs penelope, feeds it your two-part OED files, tells them to expect stardict format, tells it to output oed.zip (which penelope ignores; it will output dicthtml.zip), tells it to output in kobo format, and tells it that the to and from languages are both english
4. rename dicthtml.zip to dicthtml-oed.zip
this is desirable because without renaming the file the current Kobo FW will not remember that you've selected a different dictionary causing you to keep getting the default when you highlight a word. By renaming the file, it will persist for each lookup
5. make sure you're on the current firmware
6. connect your Kobo to your PC
7. drag dicthtml-oed.zip onto your kobo, placing it in .kobo\custom-dict
8. eject and unplug your kobo
9. go lookup a word
10. click the english dropdown and then select dicthtml-oed
11. Enjoy.

[geek1011]

It turns out that dicthtml.zip had become corrupted, likely because of special characters which cannot be compressed (or with which some zip programs have trouble compressing).

That shouldn't be possible if you're using dictutil. It shouldn't be possible with PyGlossary either, but I'm not 100% certain of that. Can you send the original file and the corrupted zip via PM so I can take a look?

fixed this by recreating dicthtml.zip using Penelope.

What happens if you use dictzip-decompile on that zip?