https://bugs.documentfoundation.org/show_bug.cgi?id=137737
Mike Kaganski <mikekagan...@hotmail.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
Keywords| |needsUXEval
CC| |er...@redhat.com
See Also| |https://bugs.documentfounda
| |tion.org/show_bug.cgi?id=70
| |142
--- Comment #22 from Mike Kaganski <mikekagan...@hotmail.com> ---
(In reply to Justin L from comment #21)
> This is difficult because we have two very real situations here. The most
> common one would be OPs - where LO implementation specifics are not cared
> about, and the regex is only concerned about the "text". So I think this is
> the one that needs the most focus, even though regex is a "pro user" feature.
>
> The second would be someone who is actually searching for footnote markers
> etc. But that "developer class" level of user will be very rare.
Thank you!
Let me mention tdf#70142 for some context here. People *in general* expect that
what they *see* on screen is searchable (which includes fields, numbering,
etc.); and in this regard, footnote anchor at lease should *participate* in
search (even if its character is not (yet) searchable).
OTOH, making dynamic text inside special objects searchable has some drawback:
e.g., replacing would be impossible when find-and-replace searches inside
fields (how would it replace inside or across boundaries of a field?).
So conceptually, there might be two modes for find: "normal" search, and
"search inside fields" (the latter one disabling replace).
But anyway, recognizing special objects as parts of text (even if just as
"control characters" in normal mode) would change workflow, but not disallow
things (possibly just require somewhat more complex expressions - consider
comment 19). I suppose that separating "normal power users" from "advanced
power users" when talking about regexes is too much, and IMO just mentioning in
the release notes (and related help page) that now the selected special objects
(footnote anchors) match [\p{Control}] would be enough - this adds flexibility.
OTOH, things like floating object anchors or reference or index entry are
completely "metadata". We need to modify the code to remove those objects from
the search string completely (which should be possible IMO, just needs more
work; and doing that would enable tdf#70142, because that way, we would be able
to replace the special characters *universally* including in the middle of the
string).
--
You are receiving this mail because:
You are the assignee for the bug.
