Kobalt: an alternative reader app for Kobo devices

[MrSilverstrike]

Originally Posted by DNSB

I installed Kobalt beta 5 on my Sage (fw 4.38.23852). That seemed to go fine but when I opened a book and tried to set the default font, it wasn't fine. The first time I selected my Charis SIL 7 modified font, my Sage restarted. After the restart, I was able to select the font and set it as the default. I then started reading and after 7 pages, my Sage restarted again. I re-opened the book and so far, I've read ~100 pages and no restart.

Thanks for the log, I'll investigate!

[Dal Bhat]

Originally Posted by MrSilverstrike

I like to think I've covered all bases.

This is such a really nice piece of work. Such comprehensive consideration of the users' preferences is just the best.

Bravo.

[Desertway]

Originally Posted by MrSilverstrike

Kobalt on Android has a library feature right now. Tagging and collections are planned. What version are you trying right now? Beta 5?

Oh, I wasn't clear. I'm on beta 5, and it does have a good library feature. I was specifically asking about "collections" in future releases, because they do help with large libraries. I'm glad to hear that they're planned. Such a fantastic reading app! Thank you!

[DNSB]

It's me again. Two more crashes. See attached logs. It seems to happen when I am attempting to select a font and make it default. The other issue is that despite selecting a font and making it the default, the next book I open reverts back to using Libron as the font to use.

[MrSilverstrike]

Originally Posted by DNSB

It's me again. Two more crashes. See attached logs. It seems to happen when I am attempting to select a font and make it default. The other issue is that despite selecting a font and making it the default, the next book I open reverts back to using Libron as the font to use.

Hmm... how many fonts do you have on your device and are you having this problem after waking the device? One crash is the watchdog that kills the device upon a hang and one is a memory allocation fail, which seems to indicate we run out of memory.

Any further logs you can send me that are Kobalt specific?

Specifically, if they exist:

.adds/kobalt/events.jsonl
.adds/kobalt/diagnostics/events.jsonl
.adds/kobalt/diagnostics/events.jsonl.previous
.adds/kobalt/crash-reporting/ (all files)

I've run into a crash issue with descriptor exhaustion in the past, and that can happen under certain circumstances when a sync is happening in the background and Kobalt is loading the fonts at the same time.

[DNSB]

Hmmm... There are 26 fonts in total with a likely candidate for a trouble maker being Arial Unicode MS (a rather large at ~23MB font file).

[howyoudoin]

Originally Posted by MrSilverstrike

Well, that and there's also a percentage option too, which is very accurate because the entire book is laid out when it loads. Unlike Nickel which estimates anything beyond the current chapter in 4.x.

There's also the progress bar which is visual. I like to think I've covered all bases.

Would it be possible to exclude all the notes/appendices/indexes at the end of the book from the calculations for pages/time/percentage?

[DNSB]

How is any renderer to know what is text and what is notes/appendices/whatever fluff is included.

[howyoudoin]

Originally Posted by DNSB

How is any renderer to know what is text and what is notes/appendices/whatever fluff is included.

Chapter name is one way

[Sirtel]

Originally Posted by howyoudoin

Chapter name is one way

There are hundreds of possible variants for those things. And chapters themselves aren't always named "chapter" either.

[MrSilverstrike]

Originally Posted by howyoudoin

Would it be possible to exclude all the notes/appendices/indexes at the end of the book from the calculations for pages/time/percentage?

Exactly as the posts above say: there's no standard for this, so no. Not possible, at least not on a low-power device such as this. It'd take training a neural net and stuff like that if you realistically wanted to do this.

Programmatically, it'd be guesswork at best, and doing this for the many languages that are out there is essentially not something I'm interested in.

[DNSB]

Originally Posted by howyoudoin

Chapter name is one way

What if I am reading a French language book? Or one of those many books where the chapter name is simply a number? There is no standard way of naming a chapter.

In theory, you could do this for an ePub3 book that made the correct use of epub:type (epub:type="appendix" for example) or ARIA roles but those books are currently not all that common though accessibility should make them more common.

See EPUB Type to ARIA Role Authoring Guide for more on the ARIA roles and epub:type to role conversion.

[fogice]

Originally Posted by rantanplan

Readest is certainly my preferred reader on the iPhone.

Thanks for the suggestion, Readest looks nice. But syncing isn't working right for me.

So far I've:

• Installed Readest on my iPhone
• Signed up for a kosync at kosync.eu
• Set kosync.eu in Readest and Kobalt
• In Readest, set "checksum method: file content" (as instructed by Readest help)
• In Kobalt, set "document matching: binary" (ibid)

However, my reading progress is being stored on kosync.eu in a separate file from each device, and they don't communicate.

Initially, I had a kepub in Kobalt (USB loaded through Calibre) and an epub in Readest (AirDrop from Calibre to iOS Files, then opened in Readest). I thought that might have been the error, so I directly placed the epub into the Kobo directory rather than using Calibre, and deleted the records in kosync.eu. The two new records at kosync.eu continue to update independently of each other.

Any suggestions as to how to properly connect the two reading progresses into one? I'm willing to erase all kepubs on my Kobo and reload them as epubs, but that is time consuming if that's not the answer.

[MrSilverstrike]

Originally Posted by fogice

Any suggestions as to how to properly connect the two reading progresses into one? I'm willing to erase all kepubs on my Kobo and reload them as epubs, but that is time consuming if that's not the answer.

You need to copy identical files on both devices (so either kepub or epub) if you want them to sync with binary document matching.

(You may want to wait until beta 6 to try again as there were some issues with sync positions with some providers, and kosync.eu was one of them...)

[fogice]

Originally Posted by MrSilverstrike

You need to copy identical files on both devices (so either kepub or epub) if you want them to sync with binary document matching.

I believe that I did that the second time. Epub directly copied into Kobo directory system. Epub airdropped to iPhone/Readest. Would either of those methods alter the file? Is there another transfer method that would work better? (If I get this step working, I plan to set up a Calibre web instance that I can access with Tailscale in order to move books ad hoc to iPhone/Readest.)

(edit) just saw the other edit. I'll wait.