https://bugs.documentfoundation.org/show_bug.cgi?id=153710
--- Comment #4 from sdc.bla...@youmail.dk ---
(In reply to Dieter from comment #3)
> So is it a bug or a docmentation issue?
(imo) both aspects are relevant -- and may be relevant for a few other tickets
(e.g., bug 93904). I will explain, so that you can decide on a way to proceed,
and also to seek your feedback on the documentation part (because it is a
little complicated).
The documentation part:
The current description cited in comment 2 was introduced recently (22 feb) as
a resolution of bug 105628.
But after that change (26 feb), I got a new/different understanding (summarized
in bug 153560 comment 10), which is what motivated comment 1.
Subsequently, it was decided (9 mar) to document actual behavior (see bug
147004, comment 20).
I just checked now and can see that I have not (yet) added the "latest"
formulation for "user-defined", but was starting to work on it now, so I will
take care of the documentation part of this ticket, following the (9 mar)
decision.
Now the complicated part.
I have just done a similar updating for "Show up to level" for caption
numbering (e.g., AutoCaption [1} (to appear soon in online help). And today a
version was made for Alphabetical Index [2], which will also be used for this
ticket ("User-defined"). I mention these two so that I can seek your opinion
about how to handle the documentation.
The issue is this: If a document uses headings in a "typical" way [3], then
the current online help (cited in comment 2) is pretty accurate. It is only
when "non-typical" structures are used that seeming bugs appear (such as this
ticket, as well as bug 147004 and bug 93904, which also involve fields.)
In AutoCaption [1], I tried to indicate a "rule" for typical heading
structures, and then put the actual rule in a "note". The "actual" rule seems
a little too complicated to "think with", especially when it seems that most of
us (as seen in BZ) are assuming that the "level" option is for specifying how
many levels to show. And as noted, with a "typical" heading structure, it DOES
work that way.
So I tried to find a way to give simple understandable help, but signal that
this is in typical situations, but then also give the "actual rule" that can be
used to explain/understand "atypical" cases.
After the new versions in [1] and [2] appear, then let me know what you think.
No doubt they can be improved. Perhaps a similar solution could work here? At
present, the version for Alphabetical Index [2] does not mention the actual
rule, but maybe it should? (there are also tooltips and extended tips)
The (maybe a) "bug" part.
Both this ticket and attachment 185594 from bug 153560 involve a situation
where the search rule does not find any match, according to the selected level.
In that case, the field displays the first heading in the document (i.e., not
following the rule anymore). In this one (corner) case, the expected behavior
should be: "nothing is displayed"
I consider the current behavior to be a bug. (the attachment gives a good/easy
demonstration).
[1} https://help.libreoffice.org/7.6/en-US/text/shared/optionen/01041100.html
[2] https://help.libreoffice.org/7.6/en-US/text/swriter/01/04120222.html
[3] start with outline level 1, continue with outline 1 or outline level 2.
With outline level 2, all subsequent headings have an outline level of -1, 0,
or +1. In other words, no jumps or gaps in outline levels, each heading is
preceded or followed by another heading that has the same outline level, or is
one greater or one lesser.
--
You are receiving this mail because:
You are the assignee for the bug.
