EPUB output

[ekaser]

I have two ePubs (amongst others), one from each of two different publishers. They both started as .LIT files, and both .LIT files look fine in MS Reader. I converted both of them to ePub using Calibre, and they both look fine when viewed in Stanza. One of them looks fine in Calibre's viewer, but the other doesn't wrap. Each 'paragraph' is a single very long line in the Calibre viewer. I've tried unZIPing both ePubs and comparing some of the html, but I'm not great html wizard and haven't spotted anything obvious to me.

Any idea what might cause this behavior in one of the files (and/or is there anything in the Calibre conversion settings I can change that might fix it)?

I don't have an e-reader yet (it's on order), so can't test the files in actual e-reader hardware.

Everett

[kovidgoyal]

Look for a <pre> tag in the html and convert it to <p> tag.Don't forget to convert the closing tag as well.

[ekaser]

Originally Posted by kovidgoyal

Look for a <pre> tag in the html and convert it to <p> tag.Don't forget to convert the closing tag as well.

Thanks! That turned out not to be the problem, but it got me looking in the right place. It turned out (I think) to be a problem with the <DIV> blocks (after comparing the two files further). I ended up removing all of the <DIV> and </DIV> markers, except for a few around headings, and that fixed the problem.

Now here's the (possibly) interesting thing. The mis-performing .LIT file is one that I bought over 6 years ago to use on a Pocket PC. At the time, for whatever reason, I'd used the CLIT program to split the LIT file apart into it's HTML components. When recently starting conversions of favorite books to epubs (in preparation for receiving my first eInk reader), I downloaded Calibre, started figuring out how to use it, and used the original .LIT files as the source to convert from, resulting in the EPUB file that didn't wrap. After unZIPing the epub and figuring out that the DIVs appeared to be the problem, I remember/realized that I also had the HTML files that had been generated by CLIT. So out of curiousity, I used Calibre to convert that HTML file to an epub, and it worked fine! So, I think (not 100% sure, but pretty sure) that the problem occurred in Calibre's conversion of the LIT file to its internal HTML format.

Everett

[ekaser]

I think I may have figured out what the errant statement is in the LIT->EPUB output that's causing the "no wrap" problem, it's <div class="calibre5"/>. What is in the source LIT file that's causing this problem, I don't know. I've added a note to the ticket regarding this.

Everett

[ekaser]

I've found another case, where converting a single HTML file to an EPUB causes bogus formatting, this time the wrong font and other stuff. The problem again appears to be the insertion of a flakey <div class="calibre10"/> block (only the first one has the '/', all subsequent ones omit it). I've added this (and the 'offending' HTML source file) to the same ticket as the above EPUB problem.

Everett

[kovidgoyal]

Thanks, I'm a tad busy at the moment, so it will be a few days before I can look at it.

[JSWolf]

Originally Posted by ekaser

I've found another case, where converting a single HTML file to an EPUB causes bogus formatting, this time the wrong font and other stuff. The problem again appears to be the insertion of a flakey <div class="calibre10"/> block (only the first one has the '/', all subsequent ones omit it). I've added this (and the 'offending' HTML source file) to the same ticket as the above EPUB problem.

Everett

What you need to do is look at the CSS for calibre10 and see what it is doing and fix what needs fixing.

[Spit]

Hi,

When I use calibre to create an epub file from a pdf to read on a sony 505 can i change the font size?

I tried putting 8,10,12 etc in the font size box but the smallest font is always at default it seems. I just wanted more text on screen.

Any tips for this?

[ekaser]

Originally Posted by JSWolf

What you need to do is look at the CSS for calibre10 and see what it is doing and fix what needs fixing.

If I knew more about HTML and was more familiar with Calibre itselt, I might. But I'm still a newbie with Calibre, so I have to settle (for now) for just being a "bug reporter."

[JSWolf]

Originally Posted by ekaser

If I knew more about HTML and was more familiar with Calibre itselt, I might. But I'm still a newbie with Calibre, so I have to settle (for now) for just being a "bug reporter."

If you attach the CSS and tell us which entry is the problem I can have a look and see what might be the problem.

[ekaser]

Originally Posted by JSWolf

If you attach the CSS and tell us which entry is the problem I can have a look and see what might be the problem.

I had opened a ticket on the Calibre web site, and attached the two example files to that ticket. I can't predict what code in the input HTML is causing the problems with the EPUB HTML output, so trying to post snippets here would probably be less than fruitful.

[mtravellerh]

Originally Posted by Spit

Hi,

When I use calibre to create an epub file from a pdf to read on a sony 505 can i change the font size?

I tried putting 8,10,12 etc in the font size box but the smallest font is always at default it seems. I just wanted more text on screen.

Any tips for this?

In a CSS sheet, try this

body {font-size: 120%}

That should work!

Try a few different percentages to see what size you prefer. I don't know if this works on the Sony, but it does work on ADE so I guess it will work on a Sony, as well!

Please be advised, though, that converting directly from PDF still is not really a good idea (convert your PDF to txt or html first before converting it to ePub)

[JSWolf]

Yes, font size by % will work on the 505/700.

[frabjous]

I mentioned this in the Viewing software forum, but I thought I'd bring it up here too since it involves calibre.

I recently used calibre (0.6.5 linux) to convert an HTML document to an epub, commandline:

ebook-convert imp-c.html imp.epub --output-profile="sony" --chapter="//h:h2" --cover="CoverDesign.jpg" --authors="Bertrand Russell" --title="Introduction to Mathematical Philosophy" --author-sort="Russell, Bertrand" --extract-to=epub_extract

In the HTML source used to generate it, the chapter headings looked like this:

<h2 class="chaptertitle" id="chapter13">CHAPTER 
XIII: THE AXIOM OF INFINITY AND LOGICAL TYPES</h2>

...with a (typically harmless) linefeed after "CHAPTER".

Calibre's chapter detection put items like the following into the toc.ncx zipped inside the epub:

    <navPoint id="cdd7d91a-1d3f-48d7-8546-19fc1a8ae123" playOrder="16">
      <navLabel>
        <text>CHAPTER 
XIII: THE AXIOM OF INFINITY AND LOGICAL TYPES</text>
      </navLabel>
      <content src="imp-c_split_16.html#chapter13"/>
    </navPoint>

Again with a linefeed after "CHAPTER".

On my Sony, the TOC was fine, but if viewed in ADE, the linebreak in the TOC file seems to result in parts of chapter name displayed on top of themselves:



If I changed the toc.ncx file inside the ePub to avoid the linefeed, as in:

    <navPoint id="cdd7d91a-1d3f-48d7-8546-19fc1a8ae123" playOrder="16">
      <navLabel>
        <text>CHAPTER XIII: THE AXIOM OF INFINITY AND LOGICAL TYPES</text>
      </navLabel>
      <content src="imp-c_split_16.html#chapter13"/>
    </navPoint>

(Or, equivalently, changed the source HTML and reconverted...) ...then things worked fine in ADE.

Insofar as this could be considered a "bug", it seems like it's a bug with ADE, since I don't think linefeeds in XML files are ever supposed to do anything (and this looks like an XML file). But you still might consider altering how calibre generates these files to avoid this issue.

(Along similar lines, I've noticed that calibre's viewer itself often responds inappropriately to linebreaks in source documents...)

[kovidgoyal]

Now fixed.