[Device Plugin] Kobo Touch Extended Driver

[Lucas Malor]

Thank you David. I think I have to use the Old Good Elbow Grease.

PS: I think Kobo had a good idea about ids. It's surely more faster to select a DOM element by id instead of searching it.

[Mivo]

Thanks for the work you put in the plugin. It is a shame that you're dropping the support for it, but I certainly understand not wanting to work for free for a company that frustrates the stuff out of you.

I have walked away from a company that I had given several thousands to over the course of years when they refused to refund ten lousy bucks for something that had been their mistake. So, I definitely understand the irritation and the desire to pull out.

But do know that the community appreciated you and your efforts, even if Kobo seems to have taken your work for granted. It is one of those companies who are hellbent on sabotaging themselves. I do like their readers, though.

[davidfor]

Originally Posted by Lucas Malor

PS: I think Kobo had a good idea about ids. It's surely more faster to select a DOM element by id instead of searching it.

Hmm, don't know about faster. The ADE method simply counts tags. And I think that is roughly the same way that has been defined for epub3. The latter is what calibre does in the viewer.

The ids are an overhead. They have to be added and unique. And if the book code changes, they need to be maintained.And the way that Kobo have done it by wrapping a span around each sentence, makes editing the book a little harder.

[PeterT]

As posted over in the Kobo forums there seems to be an error in the driver involving the handling of certain, unusual markup.

The original post reads:

Originally Posted by cybmole

in one specific book, I am seeing a most strange effect. I appreciate that teh authro of the relevant plugin which tweaks epub iunto kepub is not longer actively involced here. but I am hoping shomeone else can shed some light

what i see on AuraHD is that each h tag chapter header is printed twice.
when I intcept the converted-to-kepub code, then sure enought it is in there - see below, but it is not in the epub source,
so something in how this header line is constructed must be causing the plug-in to glitch. I have not seen it do this with any other book

kepub code ( plug in configured to save a copy to PC, as it send to device )

<h1 class="h"><span class="koboSpan" id="kobo.6.1"><img alt="002" class="imgepub" src="../Images/stro_9781101081983_oeb_002_r1.gif" />BOMB SCARE</span><span class="koboSpan" id="kobo.6.1">BOMB SCARE</span></h1>

epub (my calibre library file ]

<h1 class="h"><img alt="002" class="imgepub" src="../Images/stro_9781101081983_oeb_002_r1.gif" />BOMB SCARE</h1>

is it's not going to happen to any other books then I can live with it, i would guess that the <img bit is what triggers the glitch, and that if I stripped that out of the source file, then the doubling effect would go away.
PS the gif is just a little circle thingie, it's not anything meaningful or important.

so I what I should see for that snippet is (centered) BOMB SCARE

what I actually see on Aura HD is (centered) BOMB SCAREBOMB SCARE

and the same throughout the book, for each similarly constructed header

Attached find a small ePub that demonstrates this.

[jgoguen]

I suspect there's some oddity around handling the tail text of the <img> tag. In code, the 'BOMB SCARE' immediately following the tag is actually the element's .tail attribute. I think something is happening that the <img> tag is being copied as if it were text; perhaps there's an oddity in handling tags that have no .text attribute but do have a .tail attribute.

Hopefully that's enough for someone enterprising enough to go fixing this. And hopefully I'm at least close to right

[davidfor]

Yes, it's something like that.

I was also wondering were the spans should go. Should there be one span around it all? Or one for the img and one for the tail? I grabbed the preview and it has:

<span class="koboSpan" id="kobo.2.1"><img src="stro_9781101081983_oeb_002_r1.gif" alt="002" class="imgepub" xmlns="http://www.w3.org/1999/xhtml" /></span><span class="koboSpan" id="kobo.3.1">BOMB SCARE</span>

The tag and the tail text are being treated as separate sentences. Which makes sense as the tail could be multiple sentences.

[jgoguen]

I never looked at a real KePub with text following an image, but my hunch is that the tag and text should have each their own span.

[davidfor]

It was what we thought it was. Attached is a version that fixes the problem. It does this by handling the child text of a node separately from the rest of the node. It wasn't to hard to do, and I actually spent most of the time trying to get the counts on the ids right. Then I put the original back in and found it had the same problem.

As well as the plugin, I have attached a modified version of PeterT's sample. This adds some other combinations for where the image could be.

[jgoguen]

Once I'm in front of a computer I'll update the file in the first post with this. Should another thread be made and the plugin index updated to point at that thread? Updates to this are probably infrequent.

[PeterT]

I think a new thread would be a good idea, apart from the fact we don't know who will pick up issues (and ideally they should be the one to create the thread).

Also, remember this same fix is needed in the kEpuib Output plugin.

[jgoguen]

Thanks for reminding me

The shell/PowerShell scripts I have in the git repository (not included in the plugin ZIP) rebuild both plugin ZIP files. Once I merge davidfor's changes, that'll give me both updates at once. And hopefully whoever takes this over can make use of them as well.

[pblasi]

Hi,

I'm quite new here with my Kobo Aura.

I was very happy to know about this project and now I'm sad to know it will be not continued... Anyway a great work! Congratulations and thank you!

I have an issue with hyphenation in kepubs. It simply does not work properly while in epubs it does. First I send an epub version of a book using calibre and hyphenation works perfect. Then I use your plugin to send a kepub of the same book, but now hyphenation is full of errors, badly splited words.

I've already loaded hyphen dicts from libreoffice, following your instructions, and tried all combinations (tick-untick "hyphenate" option in your plugin, trying "hyphenate this!" plugin...) No matter what I do, hyphenation is ok in epubs, but completely wrong in kepubs... It's making me crazy because I really want to use kepubs but this issue is keeping me away from doing it.

Any idea where the problem could be?

Thank you very much! I will really appreciate any suggestion.

[jgoguen]

No promises, but post your debug log and I'll see if anything jumps out at me. Also, what language is the book in, and what language is your reader using?

[pblasi]

Originally Posted by jgoguen

No promises, but post your debug log and I'll see if anything jumps out at me. Also, what language is the book in, and what language is your reader using?

Thanks for your answer. You will need to excuse me, I'm just an user, not a developer. Could you tell me how to make a debug log?

Both book and reader are in spanish.

Thank you!

[jgoguen]

Sorry, the first post has details but basically you click the arrow beside Preferences, choose to restart in debug mode, then after calibre restarts send the book to the reader again. When you close calibre, a debug log will be displayed. Paste that here as a QUOTE or CODE block.