ePub Output Bug, Caused by MSWord

[Daddy Warpig]

There is an annoying bug in the Calibre ePub conversion module, linked to a "feature" of MSWord.

This original text:

to Unseelie Court on King Street and tease

is converted to the following text:

to
Unseelie Court
on

King Street
and tease

Cause:

MS Word Generated HTML/XHTML includes "smart tags." When such an HTML file is converted to ePub, these tags are translated, but errant <p> tags are inserted into the new html.

Original HTML code:

to <st1:Street w:st="on"><st1:address
 w:st="on">Unseelie Court</st1:address></st1:Street> on <st1:Street w:st="on"><st1:address
 w:st="on">King Street</st1:address></st1:Street> and tease

Translated HTML code:

to</p>
<address class="calibre8"><span>Unseelie</span> Court</address>
<p>on</p>
<address class="calibre8">King Street</address>
<p>and tease

Some solutions for end users:

Either erase the MSWord smart tags before converting, or fix the <p> tags by hand after converting (unzip ePub, edit .html or .xhtml files, rezip).

This has been reported as ticket #5671 in the Calibre Bug Tracking system.

[theducks]

Originally Posted by Daddy Warpig

There is an annoying bug in the Calibre ePub conversion module, linked to a "feature" of MSWord.

This original text:

to Unseelie Court on King Street and tease

is converted to the following text:

to
Unseelie Court
on

King Street
and tease

Cause:

MS Word Generated HTML/XHTML includes "smart tags." When such an HTML file is converted to ePub, these tags are translated, but errant <p> tags are inserted into the new html.

Original HTML code:

to <st1:Street w:st="on"><st1:address
 w:st="on">Unseelie Court</st1:address></st1:Street> on <st1:Street w:st="on"><st1:address
 w:st="on">King Street</st1:address></st1:Street> and tease

Translated HTML code:

to</p>
<address class="calibre8"><span>Unseelie</span> Court</address>
<p>on</p>
<address class="calibre8">King Street</address>
<p>and tease

Some solutions for end users:

Either erase the MSWord smart tags before converting, or fix the <p> tags by hand after converting (unzip ePub, edit .html or .xhtml files, rezip).

This has been reported as ticket #5671 in the Calibre Bug Tracking system.

Yup!
I noticed that street names seem to get broken up instead of just Italicized. Figured Kovid liked it that way

[jackie_w]

Hi Daddy Warpigs,

If you generate your HTML using MSWord, you should use the SaveAs Webpage-Filtered option rather than SaveAs Webpage. The "smart tags" should then not be created in your generated HTML and there is no need for manual editing.

[Dopedangel]

I would also recommend passing the html file through html tidy.
That cleans up many of the crap word add to the file.
I have seen files go down from 1 mb to about 500kb sometimes