![]() |
#1 |
Groupie
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() Posts: 155
Karma: 6672188
Join Date: Apr 2019
Device: Ipad Mini, Kindle PW5, Libra 2
|
AZW3 > Kepub conversion causes Kobo to crash
Issue: Converting an AZW3 to Kepub, causes a Kobo Libra 2 to hang for nearly a minute, then it crashes (no displayed error) and restarts.
Background: I've converted many other AZW3/MOBI to Kepub, haven't seen this behavior. This book is the "Cambridge Dictionary of Philosophy", ed. Robert Audi, file size about 3.6 MB. It renders cleanly in an alphabetic dictionary format in Kindle mobile apps, Kindle e-ink, and in AZW3/Epub formats in Calibre. However when I send the AZW3 or Epub to the Kobo from Calibre (or if I convert first to Kepub in Calibre, then send), it hangs for a while then crashes. |
![]() |
![]() |
![]() |
#2 |
Grand Sorcerer
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() Posts: 24,905
Karma: 47303824
Join Date: Jul 2011
Location: Sydney, Australia
Device: Kobo:Touch,Glo, AuraH2O, GloHD,AuraONE, ClaraHD, Libra H2O; tolinoepos
|
That sounds like a sickel triggered reboot. Sickel is a process watcher on the Kobo devices and reboots the device if something hangs. This can happen when opening larger complicated books. And I'd be pretty sure that book would fall into that category. Because this is related to what the kepub renderer does when opening a book, it is possible that it might open on the second, third or later time you try. And after that it will open with no issues.
Other than that, without seeing the book and the code in it, I can't say for sure. |
![]() |
![]() |
Advert | |
|
![]() |
#3 |
Groupie
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() Posts: 155
Karma: 6672188
Join Date: Apr 2019
Device: Ipad Mini, Kindle PW5, Libra 2
|
Thx @davidfor, that makes sense. I recall having a similar problem here a few years ago with koreader, a 'complex' kepub file that caused it to hang/crash. Contacted dev, sent a copy of the file, eventually figured out the file was ok but just had many XHTML 'nodes' due to the complex structure of the content.
This being a dictionary, it could be similar issue, but I'm surprised that it renders pretty easily in Kindle e-ink/apps and in Calibre all formats. I'll check the source of the kepub file, see if anything jumps out and PM you with details. |
![]() |
![]() |
![]() |
#4 |
Grand Sorcerer
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() Posts: 24,905
Karma: 47303824
Join Date: Jul 2011
Location: Sydney, Australia
Device: Kobo:Touch,Glo, AuraH2O, GloHD,AuraONE, ClaraHD, Libra H2O; tolinoepos
|
I have had a look, and there is a very simple solution: get a Kobo Sage.
When I downloaded the book, I only had the Sage near me so I put it on that first. It opened with no issues. When I fetched my Libra H2O, it rebooted and was definitely a sickel reboot because it hung. The issue is basically the size of the ToC. There are over 100005000 entries. When a kepub is first opened, the word count for each internal file is calculated. This is used for progress and the chapter bar charts. It looks like the CPU in the Sage can process the book a lot quicker than the Libra H2O. As a test, I did reduce the ToC to just the alphabet. That loaded with no problems on the Libra H2O. At 10000+ entries, I don't think the ToC is useful. Maybe if the ToC had a search function, it would be useful, but Kobo does not have that. Edit: I counted how many times "navPoint" was in the NCX and forgot to halve. Last edited by davidfor; 11-09-2021 at 05:57 AM. |
![]() |
![]() |
![]() |
#5 |
Groupie
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() Posts: 155
Karma: 6672188
Join Date: Apr 2019
Device: Ipad Mini, Kindle PW5, Libra 2
|
Thanks David, makes sense about the CPU. In fact: I noted that if I waited long enough as you suggested in your first post (it was nearly 10 minutes, with my Libra 2), eventually I could get the converted kebup to open. It is essentially, a limited CPU issue.
I also noticed a second issue: unfortunately, this type of book as a large-ish dictionary is a challenge on a basic e-ink device due to its content structure. The TOC, which as you said has over 5,000 entries in the dictionary, is SUPPOSED to render in a nested fashion: "A" with child "a" entries, "B" with child "b" entries, and so on. And it does on Calibre, or on mobile devices. Unfortunately what happens with this book in both my PW5 and the Libra 2, is that the TOC renders as a giant flat list of entries--there's no nesting. This makes the content all but impossible to 'browse' or navigate. You can scroll, but it'll take a LONG long time. You can search--but only if you already know what you're searching for. This is quite similar with other large dictionaries, for example, the Oxford Dictionary of English included with the PW5, it's the same experience. |
![]() |
![]() |
Advert | |
|
![]() |
#6 |
Grand Sorcerer
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() Posts: 24,905
Karma: 47303824
Join Date: Jul 2011
Location: Sydney, Australia
Device: Kobo:Touch,Glo, AuraH2O, GloHD,AuraONE, ClaraHD, Libra H2O; tolinoepos
|
I have to agree. There is a size were the ToC gets to be pretty useless without a search. For something like this, the more traditional Index would probably work better. And I don't think the paper version of this would have that gigantic ToC. It might have the letters but would rely on the fact it is alphabetical with short entries for finding things. And probably have a well constructed index.
But, the ToC on the Kobo device is nested. The entries under each letter are indented. But, they don't collapse and expand. If they did, it would help a little, but, honestly, I don't think by much. |
![]() |
![]() |
![]() |
#7 |
Still reading
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() Posts: 13,992
Karma: 105092227
Join Date: Jun 2017
Location: Ireland
Device: All 4 Kinds: epub eink, Kindle, android eink, NxtPaper
|
Create a page of links for each letter and only have outer level (i.e. A to Z) in the TOC.
That will work, but you'd want some automated way to edit the existing Toc and create the pages from each nested level. So you'd have 26 extra pages at the start. |
![]() |
![]() |
![]() |
|
![]() |
||||
Thread | Thread Starter | Forum | Replies | Last Post |
Image Quality Degraded (AZW3-->KEPUB) | twowheels | Conversion | 7 | 04-21-2025 09:36 PM |
Kobo Epub to AZW3/PDF conversion problem | hanhansen | Conversion | 3 | 12-19-2019 03:44 PM |
Kepub Conversion | Gibbo | Conversion | 2 | 09-24-2019 12:26 PM |
AZW3 Manga to KEPUB | RZetlin | Conversion | 16 | 08-05-2017 07:58 AM |
6.13 Crash during conversion | jquest | Calibre | 1 | 09-21-2009 05:02 PM |