[Libreoffice-bugs] [Bug 71481] Clone formatting appends copied formatting rather than overwriting applied formatting

[bugzilla-daemon at bugs.documentfoundation.org]

https://bugs.documentfoundation.org/show_bug.cgi?id=71481

--- Comment #18 from Mike Kaganski <mikekaganski at hotmail.com> ---
(In reply to Aron Budea from comment #17)
> How is it possible to see if something is paragraph or character formatting?
> For example I could set a 3-word paragraph to bold, and unbold the 2nd word,
> but instead I could set the 1st and 3rd words to bold, and not touch the
> 2nd. In one case bold is considered paragraph formatting, in other case it
> isn't.

I agree that it could be improved, because all this is not user-visible. But my
point was that currently-visible behavior is not a bug or regression, but a
consciously made decision.

To improve here, we must first decide how to continue with our combination of
character styles, paragraph styles that contain character settings, direct
formatting of paragraphs and characters. Because it's not easy to devise a good
format copy tool that would work "intuitively" in any circumstances, and in the
same time would not ruin style-based approach.

Simply applying all character-related properties to characters as direct
formatting, where some of them come from paragraph style, would tend to
multiply direct formatting and lead to surprises where user will be "unable to
apply another paragraph style (nothing changes)" where the reality will be that
new paragraph style's settings are overridden by direct character formatting.
Also consider, e.g., when a user copies character formatting (and makes many
paragraph-defined formatting the manual one), then decides to copy also
paragraph settings (including style), and has both paragraph style *and*
same-looking manual formatting applied.

Skipping paragraph-style-related character formatting leads to surprise when
user sees that some of character properties are not copied.

What is the best approach when manual character formatting (e.g., italics) does
not conflict with paragraph-style-defined character properties (font and its
size, e.g.)? Do we need to overwrite all the settings from source, and make
destination text formatted exactly like source, despite some properties of
origin and destination differ only because of para styles?

I suppose that dealing with this like with a "regression" is wrong, and
UX-advise is required, with thoughtful analysis and some design specs before
"fixing" it.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/libreoffice-bugs/attachments/20170806/49b4d0f7/attachment.html>