epub to mobi using kindlegen - question

[codrutoctavian]

I have some books which I am trying to convert to mobi using kindlegen. I have unpacked the epub and modified the content.opf and toc.ncx files and the conversion goes nicely but kindlegen is raising a lot of warnings of the type:

Warning(inputpreprocessor):W29001: inserting implicit <p>
in file: xxxxxxx line: 0000017
Warning(inputpreprocessor):W29001: inserting implicit <p>

but it converts the file eventually with no fatal errors:

nfo(prcgen):I1009: Resolving guide items
Info(prcgen):I1017: Building PRC file, record count: 0000192
Info(prcgen):I1039: Final stats - text compressed to (in % of original size): 041.65%
Info(prcgen):I1041: The file format version is V8
Info(prcgen):I1036: Mobi file built successfully

However in Kindle Previewer the book is formatted nicely as in the original epub only when simulating Kindle Fire. When simulating a generic Kindle the images are not displayed as they should be and unnecessary spaces are introduced. I have checked the html code and it seems that it is using <svg> tags for images. For example:

<p id="d1114115821">
<svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" height="100%" version="1.1" viewBox="0 0 516 561" width="100%">
<rect fill="#000007" height="561" width="516" x="0" y="0"/>
<image height="561" transform="translate(0 0)" width="516" xlink:href="images/xxxx.jpg"/>
</svg>
</p>

In the Amazon Kindle Guidelines it is mentioned:

3.9.7 HTML Guideline #7: Avoid Using Scripting with SVG Images
Scripting is not supported. All scripts are stripped from the source during conversion. SVG with animation is not supported.

So, my questions are:

1. What about the errors:

Warning(inputpreprocessor):W29001: inserting implicit <p>
in file: xxxxxxx line: 0000017
Warning(inputpreprocessor):W29001: inserting implicit <p>

what do they mean?

2. What to do about the <SVG> tags. Should I rewrite the code? To what?

Thanks

[Whackatagin]

Try Calibre instead of Kindlegen, seems to handle conversion a little more smoothly. i.e. it is far more tolerant about what you input.

Kindle uses <img tags instead of <SVG. Another common conversion bug with Kindlegen is that it ignores blank lines tagged <br>, do a search and replace with Sigil, use &nbsp; Putting it after the closing </p> will give you a standard font size blank line. Putting it inside <p> tags on a line of its own will give you a nice chapter break sized space. ie <p>&nbsp;</P>

I would try running it through Calibre before you start digging into the code. If it is a well formed epub it should convert without much bother.

Good luck.

[codrutoctavian]

Thanks, I will try that. However I don't like Calibre... but I will use it for this task.