Issue with HTML tags - fail to copy book to device

[riedje]

Calibre version: 0.9.31-x64 (Windows)
Plugin version: 1.3.2
Device: Kobo Aura HD (firmware 2.5.2)

Tag combinations such as <p><b><br /></b></p> cause the driver to emit an error.
I've created a new EPUB with Sigil and pasted this particular tag combination to isolate the issue, this book is available to be sent if you so desire.

Please note that other combinations similar to this one also cause errors, if I encounter them again I'll be sure to mention them. They do seem to have in common that there's no text within the innermost tags and/or the different style of tag notation: <tag> ... </tag> vs. <tag />.

[jgoguen]

Please post a link to the file. I'll happily take a test book duplicating the issue, especially if you've got one book with a few different examples of problematic combinations.

[riedje]

You can download the test epub from here: https://dl.dropboxusercontent.com/u/31409399/testbook.epub
If (when) I encounter other epubs (unfortunately I already fixed the older ones manually) I'll send them over right away!

[jgoguen]

It looks like there's an issue with the HeuristicProcessor class in calibre. I'll try to narrow down what exactly is wrong and file a bug report against calibre.

[riedje]

Another snippet that generates an error:
<p class="also-by-title"><i>STORM FRONT<br /> FOOL MOON<br /> GRAVE PERIL<br /></i> Available from Roc Books</p>

After playing around a bit with the HTML it looks like the error occurs when the <br /> tag is followed immediately by the </i> tag. If text is placed between those two tags, the file conversion is successful.

[jgoguen]

Same issue. I've filed a bug report (https://bugs.launchpad.net/bugs/1181873) and I'll come up with a workaround in the meantime.