EPUB output

[kovidgoyal]

Originally Posted by Hadrien

That's the first step but it's not enough.

calibre will split on page breaks first and then on <div> and <p> tags if that is not enough. That should take care of most cases.

@JeffElkins: I just tested metadata on the reader. Works for me. What does epub-meta on the epub file give you?

[Hadrien]

I honestly can't think of a good way to split automatically > 300k flows though. Sure, using those <div> and <p> tags is necessary, but you might still break the flow in the middle of a dialog.

[kovidgoyal]

Yeah but having a couple of misplaced page breaks is much preferable to having a file that doesn't work at all. Oh and I should correct myself. The split order is actually going to be:

page-break
h1-h6
div
p

[Hadrien]

Originally Posted by kovidgoyal

Yeah but having a couple of misplaced page breaks is much preferable to having a file that doesn't work at all. Oh and I should correct myself. The split order is actually going to be:

page-break
h1-h6
div
p

On the Sony Reader. It'll work on DE desktop, FBReader and Stanza. I hate the fact that we have to potentially add page breaks that have nothing to do in the text for a single device...

[JeffElkins]

Originally Posted by kovidgoyal

calibre will split on page breaks first and then on <div> and <p> tags if that is not enough. That should take care of most cases.

@JeffElkins: I just tested metadata on the reader. Works for me. What does epub-meta on the epub file give you?

Title : Table of Contents
Author : Unknown
Language : Unknown

When viewed in calibre all metadata is correct.

[JeffElkins]

Originally Posted by Hadrien

On the Sony Reader. It'll work on DE desktop, FBReader and Stanza. I hate the fact that we have to potentially add page breaks that have nothing to do in the text for a single device...

Are you saying that epub on the Sony is broken?

[Hadrien]

Originally Posted by JeffElkins

Are you saying that epub on the Sony is broken?

I wouldn't say broken. I'd rather say that ePub on the Sony Reader has a different set of rules than ePub based on the official specs, that makes things much more complicated and require to arbitrarily add page breaks where there should be none.

[kovidgoyal]

Originally Posted by Hadrien

On the Sony Reader. It'll work on DE desktop, FBReader and Stanza. I hate the fact that we have to potentially add page breaks that have nothing to do in the text for a single device...

FBReader and stanza dont have support for CSS. The situation is that if you want to view epub files on current generation devices with full support for XHTML1.1+CSS2 your only hope is adobe DE. And since all devices are memory limited, there is going to be some form of constraint on all of them. Which means that you can only view unsplit EPUB files on a desktop. To me that's completely irrelevant, since I have no desire to read ebooks on a desktop.

[Hadrien]

Originally Posted by kovidgoyal

FBReader and stanza dont have support for CSS. The situation is that if you want to view epub files on current generation devices with full support for XHTML1.1+CSS2 your only hope is adobe DE. And since all devices are memory limited, there is going to be some form of constraint on all of them. Which means that you can only view unsplit EPUB files on a desktop. To me that's completely irrelevant, since I have no desire to read ebooks on a desktop.

But as you've stated yourself, DE mobile could at least provide a fallback with no or minimal CSS support in this case, rather than a cryptic "Paging Error" message.

There's also a whole difference between unsplit ePub and an ePub file where one flow is 310k.

[JeffElkins]

My desire was to maintain one library, usable in both my Nokia tablets and Sony Readers (Two 505s, an N800 and 770). Looks like that's a non-starter for a while

[kovidgoyal]

True, I agree the Adobe DE could have been designed a lot better, but the fact remains that the only way users are going to be able to use epub in all its glory on their devices is by some sort of splitting.

@JeffElkins: Once I implement splitting support you will be able to move wholly to using epub on all your devices.

EDIT: And I should add that I doubt splitting is going to change the meaning of the content in EPUB files significantly, in a vast majority of cases.

[JeffElkins]

Originally Posted by kovidgoyal

True, I agree the Adobe DE could have been designed a lot better, but the fact remains that the only way users are going to be able to use epub in all its glory on their devices is by some sort of splitting.

@JeffElkins: Once I implement splitting support you will be able to move wholly to using epub on all your devices.

EDIT: And I should add that I doubt splitting is going to change the meaning of the content in EPUB files significantly, in a vast majority of cases.

Good to know, thank you.

[wallcraft]

Originally Posted by kovidgoyal

And I should add that I doubt splitting is going to change the meaning of the content in EPUB files significantly, in a vast majority of cases.

I agree, but in any case once Calibre has epub2epub with splitting it will be ok for DRM-free ePubs to be "unsplit" because Calibre can split them for devices that need it.

Also, if anyone does not like where Calibre splits an ePub they can always add their own page-break to the original and have Calibre re-split it.

[kovidgoyal]

version 0.4.85 is now available with comic2epub

[fireproof]

I'm using the command-line tools that are installed with the MacOS X version of Calibre (v 0.4.85), and html2epub seems to be using a cached html document (codyw1003010030-8.html) instead of the XML file specified at runtime (collinsw155155.xml):

dsl81:/tmp matthew$ /usr/bin/html2epub -o /tmp/collinsw155155.epub --chapter-mark=both --cover=/tmp/collinsw155155.jpg -t "The Moonstone" -a "Wilkie Collins" /tmp/collinsw155155.xml
Building file list...
	Parsing /tmp/codyw1003010030-8.html
	Parsing /tmp/codyw1003010030-8_toc.html
Output written to /tmp/collinsw155155.epub


dsl81:/tmp matthew$ /usr/bin/html2epub -o /tmp/collinsw155155.epub --chapter-mark=both --cover=/tmp/collinsw155155.jpg -t "The Moonstone" -a "Wilkie Collins" -vvv /tmp/collinsw155155.xml
Building file list...
	Found files...
		HTMLFile:0:a:/tmp/codyw1003010030-8.html
		HTMLFile:0:a:/tmp/codyw1003010030-8_toc.html
[INFO] __init__.pyo:297: 	Parsing /tmp/codyw1003010030-8.html
[DEBUG] __init__.pyo:297: Written processed HTML to /tmp/codyw1003010030-8.html-parsed.html
[DEBUG] __init__.pyo:297: Written processed HTML to /tmp/codyw1003010030-8.html-nocss.html
[INFO] __init__.pyo:297: 	Parsing /tmp/codyw1003010030-8_toc.html
[DEBUG] __init__.pyo:297: Written processed HTML to /tmp/codyw1003010030-8_toc.html-parsed.html
[DEBUG] __init__.pyo:297: Written processed HTML to /tmp/codyw1003010030-8_toc.html-nocss.html
Output written to /tmp/collinsw155155.epub

dsl81:/tmp matthew$ rm -f codyw1003010030-8.html

dsl81:/tmp matthew$ /usr/bin/html2epub -o /tmp/collinsw155155.epub --chapter-mark=both --cover=/tmp/collinsw155155.jpg -t "The Moonstone" -a "Wilkie Collins" -vvv /tmp/collinsw155155.xml
Building file list...
Traceback (most recent call last):
  File "/Applications/calibre.app/Contents/Resources/loaders/html2epub.py", line 9, in <module>
    main()
  File "/Applications/calibre.app/Contents/Resources/lib/python2.5/site-packages.zip/calibre/ebooks/epub/from_html.py", line 165, in main
  File "/Applications/calibre.app/Contents/Resources/lib/python2.5/site-packages.zip/calibre/ebooks/epub/from_html.py", line 91, in convert
  File "/Applications/calibre.app/Contents/Resources/lib/python2.5/site-packages.zip/calibre/ebooks/html.py", line 616, in get_filelist
  File "/Applications/calibre.app/Contents/Resources/lib/python2.5/site-packages.zip/calibre/ebooks/html.py", line 227, in opf_traverse
  File "/Applications/calibre.app/Contents/Resources/lib/python2.5/site-packages.zip/calibre/ebooks/html.py", line 112, in __init__
IOError: Could not read from file: /tmp/codyw1003010030-8.html with error: [Errno 2] No such file or directory: u'/tmp/codyw1003010030-8.html'

Is there something I need to clear out of the filesystem? This command worked for several hundred files before something somewhere gave out...

Lest it sound like I'm ungrateful, I'd like to say that this looks like just the epub creation tool I've been waiting for!