Writer: character styling in TOC items

ajlittoz · January 17, 2021, 9:40am

Has anybody experienced the following “accident”?

I add a new chapter in an existing document as a Heading 2 heading followed by Text Body paragraphs. The heading is formatted by its paragraph style without any additional character style nor direct formatting.

When the TOC is updated, one of the middle words (not the first nor the last one, always the same) in the TOC record shows up in italics! The TOC is protected against manual changes (as usual). I disabled the protection to apply Ctrl+M which reverted the word to its Contents 2 formatting. But whenever I update the TOC, italics is back.

I went to this unique Heading 2 (all others behave as expected) to make sure no extra formatting was there. I applied Default Character Style and Ctrl+M to make sure. But this didn’t fix the TOC issue.

As a last resort, I styled my heading Text Body then Heading 2 back. This time, the TOC is correct.

Has anybody an explanation for this?

In case there is a trick to forward a character style to the TOC, I am interested in it.

LO version: 7.0.4.2
Fedora Linux 33 with KDE Plasma desktop

anon87010807 · January 17, 2021, 12:41pm

Sample document? Can’t reproduce on 7.0.3 LO from TDF on Ubuntu 20.*.

ajlittoz · January 17, 2021, 12:48pm

Unfortunately not. As soon as I fixed it with the style switch and back, I couldn’t reproduce it. The file is no longer is its prior state, so it is useless to attach it now.

anon87010807 · January 17, 2021, 1:08pm

So it might be a fluke in the xml?

ajlittoz · January 17, 2021, 1:24pm

I’m rather confident in the deterministic aspect of computers (at least as long as an application is deeply tested and debugged) and I don’t understand why, all of a sudden, something as trivial as adding a new heading without added formatting would trigger a change of formatting somewhere else (the heading showed “normal” and a single word of it was italicised in the TOC).

It is now very difficult to analyse the issue as the file has changed. This is why I asked. On my side, the problem is fixed by a stance which worked in many other instances (though I also son’t understand why it should work). So the question is rather academic.