[Plugin] CustomCleanerPlus

[DNSB]

Originally Posted by Turtle91

<h3>Chapter 1 <span>This is the title of the chapter</span></h3>

The CSS would look something like this:

h3 {font-weight:bold; text-align:center; font-size:1.2em}
h3 span {display:block; font-weight:normal; font-size:.8em; font-style:italic; font-family:serif}

This very specifically defines which <span> is being styled (any <span> within an <h3>). It also minimizes code bloat (no need to say <span class="chaptertitle"> on every chapter).

It is certainly not required to do it this way...you can add extensively defined spans to every letter if you want...but having an empty naked <span> does not automatically mean it is wrong.

What I was seeing in the ebook where removing the naked spans caused very noticeable changes was the author/publisher using a span following a p to define the dropcap class, 3 flavours of h1, h2 and h3 with a span defined in the stylesheet, etc., all without using names.

My personal preference is to give these classes names just to make it easier for me to follow later. <p><span>A</span> is a bit harder to follow than <p><span class="dropcap1">A</span> when you come back to it a few months later.

[slowsmile]

Originally posted by DNSB
"My personal preference is to give these classes names just to make it easier for me to follow later. <p><span>A</span> is a bit harder to follow than <p><span class="dropcap1">A</span> when you come back to it a few months later. "

I really don't understand why you don't properly format those spans correctly as you go instead of using those empty spans as markers that you can go back to later. Sorry but I don't regard that as a good enough reason to prevent automatic empty span removal in the plugin.

[KevinH]

CSS combinator selectors appears to be an "in" thing to do to prevent the need to create and use css classes in your html/xhtml even though class names can actually be chosen to mean something!

That said ... they are not naked spans and should not be removed.

[slowsmile]

KevinH and @DiapDealer...In that case -- if removing empty spans really does effect the formatting as you say -- I will add another option for empty span removal in the plugin. This also means that I will have to create another separate user options dialog, since the current user options dialog window's height is now excessive.

[DiapDealer]

Originally Posted by KevinH

That said ... they are not naked spans and should not be removed.

Certainly not silently and/or without the ability to opt out of any such potentially damaging deletion, in my opinion.

[slowsmile]

Originally posted by DiapDealer
"Certainly not silently and/or without the ability to opt out of any such potentially damaging deletion, in my opinion."

I'm not quite sure why removing empty spans i.e. span tags containing nothing -- would cause any problems. If it does cause major problems, as you say, then perhaps it's best that I should just remove it altogether from the plugin(with no selectable option in the user dialog). Thoughts?

[DiapDealer]

Originally Posted by slowsmile

This also means that I will have to create another separate user options dialog, since the current user options dialog window's height is now excessive.

For what it's worth... the current dialog can't be expanded wide enough to see all of the text on my highdpi monitors. But there's plenty of room for the height to be increased.

I'd recommend multiple columns or even tabs, but then it is tkinter. So everything is easier said than done. I ultimately had to get away from using it (Tk/TKinter). Made my head hurt something fierce!

[DiapDealer]

Originally Posted by slowsmile

I'm not quite sure why removing empty spans i.e. span tags containing nothing -- would cause any problems. If it does cause major problems, as you say, then perhaps it's best that I should just remove it altogether from the plugin(with no selectable option in the user dialog). Thoughts?

Sorry for the confusion. I have no opinion on removing spans that contain no text between the opening and closing span tags. I'm not certain why they'd be there in the first place, but removing them should change nothing. I've been referring to spans that contain text, but whose tags have no attributes. Removing those can change how things render (as others have already illustrated).

One option would be to stop removing naked (no attributes) span tags by default with the plugin. Then make it a manual preference to remove them. Those who know the risks (and know what they're doing) and want to remove them anyway, could always manually add something like a

"remove_naked_spans": true,

to the plugin's .json preferences file that the plugin could then honor.

[slowsmile]

Originally posted by DiapDealer
"For what it's worth... the current dialog can't be expanded wide enough to see all of the text on my high dpi monitors. But there's plenty of room for the height to be increased."

Thanks for that. I am intending to split the plugin's user options dialog into two separate windows in the very near future because the plugin dialog must also be able to accommodate low rez monitors -- which is where problems might occur. And when I do the next upgrade I will also widen the dialog window to try and accommodate Linux machines as well.

Originally posted by DiapDealer
"I'd recommend multiple columns or even tabs, but then it is tkinter. So everything is easier said than done. I ultimately had to get away from using it (Tk/TKinter). Made my head hurt something fierce! "

I think I'll take the easier route by just splitting the user options in half and using two tkinter window dialogs. I haven't really had a chance to get up to speed using Qt.

And thanks for the suggestion about using the json file for selection but unfortunately I already use the json settings to save the user dialog selection for next time you use the plugin. So I think my only option is to perhaps add it as another user option in the user options dialog window.

[DiapDealer]

Originally Posted by slowsmile

I think I'll take the easier route by just splitting the user options in half and using two tkinter window dialogs. I haven't really had a chance to get up to speed using Qt.

That's fair. Qt has it's own learning curve. But since I'm already used to doing C++ Qt stuff with Sigil itself, Qt for Python just makes more sense to me. I never did fully warm up to tkinter. Nothing about it ever became intuitive for me. I had to look up online examples for everything I wanted to do with it.

[slowsmile]

Just to clarify and recap, I will make the following changes to this plugin in the near future:

• Add a new selectable option to remove line-height from the CSS
• Add a new selectable option to remove empty spans from the html.
• Create a second user options dialog window.
• Widen the dialog windows to properly accommodate Linux HD displays.

[JSWolf]

How about checking the CSS and if there is no style for a naked span, then go ahead and remove them?

[DiapDealer]

Originally Posted by JSWolf

How about checking the CSS and if there is no style for a naked span, then go ahead and remove them?

That would be much, much, much easier said then done. Prohibitively difficult (not to mention nearly impossible to guarantee its accuracy).

[DNSB]

Originally Posted by slowsmile

I really don't understand why you don't properly format those spans correctly as you go instead of using those empty spans as markers that you can go back to later. Sorry but I don't regard that as a good enough reason to prevent automatic empty span removal in the plugin.

It's not something I am creating. I'm editing someone else's creation. I personally do not like using combinator selectors but others seem to feel they are the greatest thing since sliced bread.

[slowsmile]

Just to be absolutely clear, when I say "empty span tags" I mean span tags that have nothing inside them -- no classes, style or any other attributes. Example below:

<h1 class="heading"><span>CHAPTER 1</span></h1>

Whereas "a naked span tag" I think refers to a span tag containing perhaps some attributes as well as text on it's own line in the html that is not surrounded by any other html tags. That's my understanding of "naked spans" anyway. As well, "naked spans" surrounding text or whatever can also contain classes and other attributes.

Generally, as a rule, I always try and avoid using both "empty span tags" and "naked span tags" in my own epubs. In most instances, empty spans serve no earthly purpose and naked spans alone and on their own line(and not surrounded by <div> tags) will always give you an error whenever you run FlightCrew or Epubcheck.