epub 2 pdf widowed <h2> solution sought

[rjwse@aol.com]

It is my understanding that GTK does not honor widows because of "columns". I do not understand the ins and outs of the technicalities. I find <h2>heading</h2> at the bottom of pdf pages. When I designate break-inside: avoid this only treats the <h2> tag itself and, of course, does not break, but it is still at page bottom. If I wrap the <h2> pair and the following <p> tag with a <div> I can tell the <div> not to break and that works okay. But it is laborious to go through and do this manually for every <h2>. To avoid adding divs throughout the book I tried to 'group' with a combinator in a goofy way: <style>(h2+*) {page-break-inside: avoid}</style>. That does not work. I cannot see using :before, :after, :first-pudknocker, or anything else. To say it cutesy-wootsie, I am looking for a way to affect 'the antecedent of an adjacent sibling selector of a combinator'. Would appreciate input. Best regards, Pop

[kovidgoyal]

Am confused, if you dont want the h2 at the bottom why not just put a page break before it?

h2 { page-break-before: always }

and page-break-inside requires a container to apply to the inside of it, so if you want to use it you have to wrap the stuff you want to use it on in a div

[Quoth]

Only Page Break Before a heading is reliable. I always use level 1 to 3 with Break Before. That also helps TOC generation.
I use styles that map to CSS that only look like headings if I don't want a page break or don't want it in the TOC.

[rjwse@aol.com]

Thanks for input. Regarding your question:"why not just put a page break before it?
h2 { page-break-before: always }" it is because the h2 segments are extremely short and look wrong if only one is on a page, but look good if bunched together as many as possible per page, but look bad with a widowed h2. The h2s show up good in the TOC.

I mentioned in the questions that wrapping h2s and following Ps in a DIV certainly works and you said so:"and page-break-inside requires a container to apply to the inside of it, so if you want to use it you have to wrap the stuff you want to use it on in a div"

I was trying to find an elegant solution of no widows of h2 without wrapping all h2+ps. Basically, I'm learning the ins and outs of how to solve programming css limitations. Thanks for the help. Best regards, Pop

[Tex2002ans]

Originally Posted by rjwse@aol.com

When I designate break-inside: avoid [...]

If you're aiming for EPUB2 + older ereaders, you'd use:

page-break-inside: avoid

CSS2 = page-break-inside
CSS3 = break-inside

break-inside is a relatively recent renaming, so a lot of older devices/readers won't recognize that.

Originally Posted by rjwse@aol.com

I find <h2>heading</h2> at the bottom of pdf pages. [...] this only treats the <h2> tag itself and, of course, does not break, but it is still at page bottom.

I think I came up with something. (See below.)

Related Side Note: There are a few bleeding edge things in CSS3 Paged Media that deal with floats (mostly for Images+Captions/Figures/Tables/Sidebars).

It would allow those chunks to move to the top/bottom of next page, while the rest of the text fills the previous page.

I explained some of that here:

"Kindle Previewer 3.40 - 64bit, SVG text, snap-block" (Post #3+)

Originally Posted by rjwse@aol.com

Regarding your question:"why not just put a page break before it?

h2 { page-break-before: always }" it is because the h2 segments are extremely short and look wrong if only one is on a page, but look good if bunched together as many as possible per page,

If that screenshot is an example of your book... then I'd tackle it like this.

1. Wrap every "little section" in its own <section>:

Before:

<h2>Sample Heading</h2>
<p>Short and sweet paragraph.</p>
<p>Another short paragraph.</p>

HTML5:

<section class="chunk">
	<h2>Sample Heading</h2>
	<p>Short and sweet paragraph.</p>
	<p>Another short paragraph.</p>
</section>

2. Now you can apply CSS to your <section>s:

section.chunk {
	page-break-inside: avoid;
}

That should accomplish what you want, but it would only work in readers that understand <section> (so HTML5 = EPUB3 + newer readers).

And I think this is one of those cases where you can do your best to code it, then it's up to the readers themselves to be using a standards-compliant reader.