Duplicate IDs

[gregf]

Hi,

I just upgraded from 0.8.0 to 0.9.5 and now I get duplicated id from epubcheck and flight crew using the exact same ebook-convert command and input files. If I reinstall 0.8.0 it passes epubcheck.

Here is a example of a duplicate id of what 0.9.5 produces:

<a name="z01001008" id="z01001008"><span class="bcv">8</span>God called the dome “sky.” Evening came, and morning followed—the second day.
</a></p><a name="z01001008" id="z01001008">
</a><p class="calibre4"><a name="z01001008" id="z01001008">

Same content from 0.8.0:
<a name="z01001008" class="calibre15" id="z01001008"></a><span class="bcv">8</span>God called the dome “sky.” Evening came, and morning followed—the second day.
</p>
<p class="calibre5">

HTML Source:
<a name="z01001008" /><span class="bcv">8</span>God called the dome “sky.” Evening came, and morning followed—the second day.
</p>
<p>

Here is the ebook-convert command that I use for 0.8.0 and 0.9.5

ebook-convert index.html zzz.epub --flow-size 99999999 --page-breaks-before "//h:h1" --preserve-cover-aspect-ratio --chapter "/" --chapter-mark none --output-profile ipad --breadth-first --no-chapters-in-toc --max-toc-links 84 --cover zzz.jpg --author-sort "xxx" --authors "yyy" --publisher zzz --title "xyz" --isbn 111-1-1111-1111-1

thanks for the help,
Greg

[kovidgoyal]

You need to post a complete html file demonstrating the issue rather than a snippet.

[gregf]

Attached is the one of the input html files and related output file giving the duplicate ids. An example duplicate id=z01001008

thanks.

[kovidgoyal]

calibre has switched to an html 5 based parsing algorithm.As per that algorithm, that is the way that the particular markup you use is supposed to be parsed, each <a name="whatever" /> tag becomes two tags after parsing, leading to the duplicates. Open your html file ina modern browser and you will see the same thing if you use the "Inspect element" functionality, the <a> tag will have become two <a> tags.

If you change the self closing tag to an explicitly closed tag the problem with probably go away.

Change <a name="" /> to <a name=""></a>

[gregf]

Thank You.