pandoc 3.6.2: unwanted blockquote tags, while bold and italic frequently vanish

[UnixRonin]

I just got pointed at pandoc as a possible less dysfunctional way to generate EPUB from ODT. I've found two issues that I'm unsure whether they're just "option I don't know about" or actual incorrect behavior.

First problem is that in the actual text, every p object gets wrapped in blockquote lines. I don't want that. I've written a filter to strip them out, but it would be better to not have them in the first place. Is there a way to turn that off? I don't see one.

The second, more serious problem is one of vanishing tags. I SUSPECT it is actually a symptom of LibreOffice's dysfunctional formatting behavior, but what I'm seeing is that, mostly at the beginnings of lines, italic or bold text formats just fall off the face of the earth and are never seen again.

Has anyone else encountered this problem?

[jgm]

If you can upload an ODT that illustrates these issues, we could take a look. The ODT reader may be using a heuristic for block quotes (such as indentation) that is inadvertently triggered in your document.

[UnixRonin]

Let me see if I can cut down a manuscript sample that illustrates BOTH problems then.

[UnixRonin]

Here you go ... a 1.5 page excerpt that illustrates both issues.
Bearing Gifts EXCERPT.odt

[UnixRonin]

Forgot to add — commandline used:

pandoc -s --split-level=2 -t epub3 Bearing\ Gifts\ EXCERPT.odt -o BGE.epub

[jgm]

IIRC, the ODT reader uses some heuristics to identify block quotes, and it's probably the fact that your paragraphs are indented that is triggering these. Probably the heuristics can be modified so this won't happen.

AS for the other issue: what is going on here is that a (temporary) paragraph style is being used that sets text properties to italic, and then a separate text style is used to reset things to normal. The ODT reader isn't picking up on this. It does okay with the more normal situation in which a text style sets the text to italics.

I'm recording these observations here, but I haven't had a chance to look at the ODT reader yet.

[UnixRonin]

Ah, so as I suspected, part of what's going on here is that LibreOffice is trying to be too clever for my own good. 😁

For additional information, the fundamental problem I'm trying to solve here is that LibreOffice generates technically readable, but internally godawful EPUBs that are almost completely infeasible to edit after generation (to correct errors, for example). And I'm trying to find a way of generating a readably (and editably) clean, EPUB3.3 compliant eBook from ODT source, with less manual labor than my current expedient of wiping ALL of LibreOffice's dysfunctional formatting out of the EPUB, writing new sane CSS, and re-applying the formatting by hand in a way that doesn't look as though an XML generator barfed in it. I'm hoping that pandoc can be a part of that solution ... IF I can get it to preserve all of the formatting parts that I actually need to keep, because re-inserting formatting is where MOST of my post-production errors creep in. Then I'd mostly just need to write new CSS and clean up the metadata.

[jgm]

See #9505 for the block quote issue.

[jgm]

I fixed the blockquote issue with two recent commits. See #2623 for the other issue.