Feature request: CBR support

[BloodRagg]

I would like to request to CBR support, I could not find any topic on this, but then again that could be my search skills also

My current workaround: (depends on 7zip and unrar)

This is a recursive version: **EVERYTHING ending on .cbr will be converted to cbz**

sudo apt install p7zip-full unrar

Conversion:

find -type f -iname '*.cbr' -exec sh -c 't="`mktemp -d`";f="{}";unrar x -y "$f" "$t" && 7z a -r -sdel -tzip -mx=9 "${f%r}z" "$t/*";rmdir "$t"' \;

Inplace conversion: (removes original cbr)

find -type f -iname '*.cbr' -exec sh -c 't="`mktemp -d`";f="{}";unrar x -y "$f" "$t" && 7z a -r -sdel -tzip -mx=9 "${f%r}z" "$t/*" && rm "$f";rmdir "$t"' \;

[Frenzie]

There's an issue about it: https://github.com/koreader/koreader/issues/1517

[BloodRagg]

Could a workaround be added in the file browser? run you custom script against the current file.
It must still show the cbr files , now they are hidden.

[NiLuJe]

Nothing ships unrar, IIRC.

And it'd be fairly IO/CPU intensive, with minimal UI callback possible (think OTA updates), which pretty much punts that off to the "gross hack" category .

That's a task much better left to the other side of the equation (i.e., do it before/during sending the file to the device instead).

[BloodRagg]

Originally Posted by NiLuJe

Nothing ships unrar, IIRC.

And it'd be fairly IO/CPU intensive, with minimal UI callback possible (think OTA updates), which pretty much punts that off to the "gross hack" category .

That's a task much better left to the other side of the equation (i.e., do it before/during sending the file to the device instead).

Kobo firmware ships unrar , but you probably meant anything based on CR3 ?
Source code of unrar is open, but im guessing this uses a library for direct handling of
zip files ? So you would need something similar for rar files ? Is some kind of a stub possible
that uses the unrar executable ?

[NiLuJe]

Huh. I wasn't expecting that .

I'd go with libarchive, it has a great API, and it just recently picked up RAR5 support.

[Frenzie]

Originally Posted by BloodRagg

Kobo firmware ships unrar , but you probably meant anything based on CR3 ?
Source code of unrar is open, but im guessing this uses a library for direct handling of
zip files ? So you would need something similar for rar files ? Is some kind of a stub possible
that uses the unrar executable ?

You can compile unrar to a library (unrar.dll, unrar.so). This is directly supported by crengine for opening rarred HTML filed, but we don't build it.

However, cbz is handled by MuPDF, not crengine.

[rkomar]

Originally Posted by NiLuJe

Huh. I wasn't expecting that .

I'd go with libarchive, it has a great API, and it just recently picked up RAR5 support.

I'm going from memory, but I thought libarchive was for unpacking entire archive files. The nice thing about using the unrar library is that you can open the files one at a time as needed. It saves on diskspace while reading, and makes startup much faster (you don't have to wait for the entire archive to be unpacked).

The pbimageviewer code I worked on for PocketBook devices unpacks single files. The source code for that is available if you want an idea of how much work it would be to implement.

[NiLuJe]

I'm not familiar with the unrar API (but I do know it's there ), but you can certainly skip through the archive's content without unpacking anything with libarchive .