docx->epub weirdness, but only for Apple readers

[mwlucas]

Hi,

I'm using Calibre 7.16.0 to convert my docx manuscripts to epub. The resulting epub works wonderfully in calibre, Kindle, and Adobe Digital Editions, but the spacing keeps changing in Apple's epub reader. I *think* I've attached everything, but just in case there's screenshots at:

https://cdn.mwl.io/junk/apple/bad1.png
https://cdn.mwl.io/junk/apple/bad2.png

A docx that reproduces the problem:

https://cdn.mwl.io/junk/apple/ryoms-appledebug.docx

A conversion of this sample:
https://cdn.mwl.io/junk/apple/ryoms-appledebug.epub

The conversion log:
https://cdn.mwl.io/junk/apple/convertlog.txt .

Is there some Apple-specific setting I should be using? Something in my docx style sheet that Apple's choking on?

Realistically: as conversion problems go, it's not horrid. It's definitely readable, just annoying for my readers. Any suggestions would be appreciated!

[DNSB]

I was noticing a couple of items in your stylesheet. You are using a mix of absolute (px, pt, etc.) and relative (%, em, etc.) measurements. I would suggest sticking to relative other than for thin borders.

Also:

Several <span class="????"></span> scattered through the text.

You might try changing the vertical-align: super; to vertical-align: top;

There is a mix of line heights for text. 1.2 seems to be the commonest. I would suggest removing the line-height: 1.2; lines and letting the reader set their preferred line height with the line-height 1.5 being relative to that user set value.

Several mixes of text-align: center and non-zero text indents in the stylesheet.

Courier New is used as the monospace font in two places and Lucida Sans Typewriter in a third.

Also note that unless you embed any specified fonts, the reader will use replacement fonts for ones that are not available.

Someone who more enthused about Word and styles will likely chime in about the issues with the .docx file.

[JSWolf]

Redo the .docx file and remove all formatting and use styles. Highlight the text and apply nothing but a style. Otherwise you end up with a mess as you can see.

[BetterRed]

Originally Posted by JSWolf

Redo the .docx file and remove all formatting and use styles. Highlight the text and apply nothing but a style. Otherwise you end up with a mess as you can see.

Did you actually look at the DOCX !?

There is no inline formatting, everything is done by styles, most of which I suspect were created by @mwlucas.

BR

[DNSB]

Originally Posted by BetterRed

There is no inline formatting, everything is done by styles, most of which I suspect were created by @mwlucas.

That was the reason I suggested that any issues with .docx and styles should be commented on by someone else.

[BetterRed]

@mwlucas - curious, what did you use to create the DOCX - Word, LO Writer, Wordperfect… etc.

I looked at your sample DOCX with Word, the one thing I noticed was that the paragraphs with 'tight' line spacing have styling within them — e.g.filename and Users and Hosts… which converts to something like <span class=text4>blah blah</span>.

Is that true for all the paragraphs that are badly formed on Apple.

If you use the same styling (template) in multiple manuscripts you might want to consider using the Mammoth DOCX to EPUB converter rather than calibre… there's a Sigil plugin called DOCX Input which is a shim to Mammoth. It provides a mechanism to map DOCX styles directly to CSS entries… takes a bit of effort to set up, but providing it is reusable its worth doing.

BR

[mwlucas]

thanks, everyone.

Yes, I created the styles. I'm all styles all the time. It's the only way to keep tech books sane.

I'll take a look at Mammoth, thanks! Calibre is the best I've found so far for these horribly format-heavy tech books, but I'm open to ideas.

Thank you all!

[BetterRed]

@mwlucas - I changed the 'filename' styling in the "You must know how to manage your operating system" paragraph to 'normal' with inline italics. It's attached, try converting it to EPUB and throwing that at Apple.

BR