Crazy style sheet

[jiembe]

I just stumble on an ePub with 16915 coding line in the style sheet for 12240 lines of text (after erasing a lot of font-face{, empty lines and commentary notes). Ever top that?

That give text like :

<p class="Style1141"><span class="FontStyle1"><span class="calibre21"><span class="title">—</span></span></span><span class="FontStyle1"><span class="calibre18">

and the Style number change every line or so!

I don't know if I have the courage to finish correcting that ePub.

[roger64]

Hi

There are two kinds of books, the ones with a simple structure (which may have a complicated code) and the others.

If your book belongs to the first category, for terminal cases like this one, I use a Sigil plugin named HTMLgen which saves the book to html format and, doing so, drastically simplifies the styles. Calibre has probably the equivalent but I do not have the practice.

I import the end result in LibreOffice (odt) and, after formatting publishes as an ePub. It's much quicker than it seems and provides you at the end with a clean code.

[retiredbiker]

I've seen books like that, with a gazillion "fontsylexxx" spans and/or a separate style name for every single paragraph. I think about 15,000 css lines was the biggest I've personally seen. I'd love to know what tools make these horrors.

I have saved searches to clean these up. I look for spans that do italics, and replace them all with simple <i>...</i> pairs. Then I blow away all spans and all styles and start over, basically.

[JSWolf]

One thing to try is to convert the ePub > ePub in Calibre and see how the code ends up. You might be able to easily edit the code from there.

Another thing to try is to remove unused CSS and use the new feature "Merge CSS rules with identical properties" in case there are a lot of CSS styles the same with different names. But you need to be using the current Calibre 5 for this to work.

And finally, if the format of the eBook is not complex, just drop the styles and make your own.

[jiembe]

Originally Posted by retiredbiker

I've seen books like that, with a gazillion "fontsylexxx" spans and/or a separate style name for every single paragraph. I think about 15,000 css lines was the biggest I've personally seen. I'd love to know what tools make these horrors.

Well in this case :
<meta name="Generator" content="Microsoft Word 14"/>
meta name="Originator" content="Microsoft Word 14"/>

I correct all and finish with 9064 lines in html files and 158 lines in the stylesheet.
____________________
(Adding)

I got a new one, this time the fellow use tag names where the the longest is 111 characters! Well, at least it is descriptive!

[JSWolf]

Originally Posted by jiembe

Well in this case :
<meta name="Generator" content="Microsoft Word 14"/>
meta name="Originator" content="Microsoft Word 14"/>

I correct all and finish with 9064 lines in html files and 158 lines in the stylesheet.
____________________
(Adding)

I got a new one, this time the fellow use tag names where the the longest is 111 characters! Well, at least it is descriptive!

I want to know how this person did this in Word? I've never seen anything like this before.

Would you please post the original CSS in a code block? Thanks.

[phossler]

Originally Posted by JSWolf

I want to know how this person did this in Word? I've never seen anything like this before.

I was thinking that maybe the original author's text was in Word, but loaded into an ePub maker that created that nightmare CSS, and then it was loaded into Calibre

Or maybe Word to Calibre to some ePub maker???

[jiembe]

Originally Posted by JSWolf

I want to know how this person did this in Word? I've never seen anything like this before.

Would you please post the original CSS in a code block? Thanks.

Here it is.

Other topic :

I see more and more using <div> instead of <p> or even <span>. Why ?

[Brett Merkey]

I've never seen anything like this before.

I was once assigned by a long-suffering employer to evaluate a browser-based application from SAP, the German software giant. Altho the app was essentially only one screen, the CSS file was over 500k. (!) Given that I was the visual designer and code lord and master of a suite of 6 multi-screen apps, all using a common stylesheet less than 6k, I was skeptical the damn thing would even run on a typical client machine.

It did. The code was parsed, somehow, without smoke coming out of the CPU. I was very disappointed.

That is the wonder of modern computers—even shit code can run fine.

[roger64]

Originally Posted by jiembe

I just stumble on an ePub with 16915 coding line in the style sheet for 12240 lines of text (after erasing a lot of font-face{, empty lines and commentary notes). Ever top that?.

About HTMLgen

If possible, can you send me the ePub in a PM? I'll return back the html file, the odt file, and the new ePub with some comments. It's a few minutes process, mostly automatic.

As the CSS file is a little verbose (...), I would begin with pruning unused styles using either Sigil or the Calibre editor.

[JSWolf]

@jiembe can you please install the ScrambeEbook plugin for Calibre and use it to scramble this eBook? Once scrambles you can attach it and I can take a look at how much of a hassle it would be to fix it. Thanks.