Comparing epubs

[DrChiper]

I just upgraded to Sigil 7.2 and fetched the accompanying new Sigil manual. Then I wondered: what are the differences between manual version 7.1 and 7.2?
Which leads to my question: Is there actually a clean method to compare 2 epubs with each other to spot differences in text? Layout differences would be a nice-2-have too, but seems to me rather difficult to realize and of limited use (but I might be wrong, as I have demonstrated before )

[Toxaris]

No, that feature does not exist. You can compare two XHTML files of course, but not automatic.
It can actually be quite complicated to do.

[DrChiper]

Would be a nice feature though. For the moment I can try comparing by using BeyondCompare (highly recommended), but unfortunately file by file ...

Perhaps new user manual versions could include change marks to indicate the changed/added/removed segments?

[Toxaris]

You can also check the release notes of Sigil. That should tell you what topics were changed.

[Arios]

If I do understand correctly, I wonder: wouldn't it be possible to merge all files in the epub and then open it with a good text editor, save the file as html (or whatever), redo it for the next one and compare? It is not a comparaison from ebub to epub, but not to far away (and styles classes are also there).

[OddCosine]

Originally Posted by DrChiper

Would be a nice feature though. For the moment I can try comparing by using BeyondCompare (highly recommended), but unfortunately file by file ...

You can tell BeyondCompare (yes, great tool) to treat epubs as zip archives, and it will give you essentially a folder compare of the contents of two epubs. Often useful, but maybe not so much for what you're trying to do.

[Doitsu]

ComponentSoftware's free CSDiff file comparison tool will also compare epubs, if you change the file extension to .zip.
You'll need to run CSDiff once and change the Analyze option to Folder. You'll also need to enable the two check boxes. (Once you're done, click OK and close the app.)
After you've changed these options, all you have to do is select two renamed epubs and drag & drop them on the CSDiff shortcut.

[DrChiper]

@OddCosine: You do not even have to "fake" the file extension for BeyondCompare, just add "*.epub" to Tools->Settings->General (Archive Associations) and BC will know from then on how to handle epubs.

[OddCosine]

Originally Posted by DrChiper

@OddCosine: You do not even have to "fake" the file extension for BeyondCompare, just add "*.epub" to Tools->Settings->General (Archive Associations) and BC will know from then on how to handle epubs.

That's actually what I meant, but I see you already know how to do it. Mine is set a little differently though - it looks like we have different versions of BC (I have 3.3.7).

[Man Eating Duck]

Originally Posted by DrChiper

Is there actually a clean method to compare 2 epubs with each other to spot differences in text? Layout differences would be a nice-2-have too, but seems to me rather difficult to realize and of limited use (but I might be wrong, as I have demonstrated before )

I use the free (and excellent) tool WinMerge. It will not compare your epubs as-is, but if you unpack them with winrar it will compare all folderse recursively. You could also convert both epubs to text with calibre, and compare those. Text files are a bit easier to work with. Both of these options can be automated and integrated with Explorer or Nautilus using a suitable shell script, this enables you to work with epubs outside of calibre by using the command line ebook-convert. I use this to check over corrections I've made in ebooks I've read, and it works fine. For best results, open and save the epub in Sigil before making a copy for editing.

None of these methods will work well if you try to compare epubs with significant differences, such as two different editions of the same book. A comparison of the Feedbook and Gutenberg versions of the same book will show *a lot* of differences, some because of different titlepages and layout, but also in the body text due to "irrelevant things" like differing dashes, quotation marks and so on.

Good luck, and if you find a brilliant solution to this problem, please post it here

[varlog]

Quick solution for single .(x)html file would be to use Sigil feature "Open With". You have to open two Sigils with books to compare, right click in both of them the section you want to compare and open with something like this script:

#!/bin/bash
SAVEIFS=$IFS
IFS=$(echo -en "\n\b")

if [ -e "$HOME/.CmpFiles" ] 
then
  echo -e "${1}\n">>$HOME/.CmpFiles
  files=`cat $HOME/.CmpFiles`
  xxdiff $files
  rm $HOME/.CmpFiles
 else
 echo -e "${1}\n">$HOME/.CmpFiles
fi

IFS=$SAVEIFS

Instead of xxdiff you could use program of your choice.
This is for Linux of course but must be doable in Windows too.