Why two separate page break xpaths in 0.6.x?

[ldolse]

This is with regard to bug #2718, which was about page breaks being inserted.

I understand now why the bug is invalid, I wasn't aware that an additional xpath had been intentionally added to insert more page breaks in final release. I'm just wondering why? It seems like a fair number of books could have a chapter headings that look like the following:

<h1>CHAPTER X</h1>
<h2>chapter title</h2>
<p>beginning of chapter text</p>

Which all displays on a single page at the beginning of the chapter. This additional xpath will now cause <h1>Chapter X</h1> to appear on a page all by itself.

I can see the benefit of having an additional xpath available to detect page breaks in some other fashion, but I'm not sure that it makes sense to have both the chapter detection xpath AND this one enabled by default....

[kovidgoyal]

The reasoning is that what's really important for chapters is inserting page breaks before them. Since the chapter detection expression requires the word chapter/part etc it wont match a lot of cases the page break expression will.

And having a few extra page breaks is not too bad. What could be done is make the page break insertion a little more intelligent so that it does not insert page breaks for elements "close" together if the expression is the default one

[ldolse]

Agree that inserting page breaks before chapters is important, the chapter detection xpath is the one accomplishing this if I understand correctly.

Adding a bit of smarts to when the second xpath triggers works for me, will just disable it for now when it causes trouble.

[kovidgoyal]

Opena ticket for it and I'll get around to it