data conversion strangeness

[A LeDonne]

On 6/17/06, Alex Dean <alex at crackpot.org> wrote:
>
> On Jun 16, 2006, at 9:08 AM, Craig White wrote:
>
> > On Fri, 2006-06-16 at 12:03 -0400, A LeDonne wrote:
> >> Another option (though I understand that you're past this): I was
> >> able
> >> to use $ by itself to match line endings in OOo's Replace with RegEx.
> > ----
> > duh - that was an important piece of info there...I guess that's why I
> > asked/responded as I did because I was frustrate at not being able to
> > 'find' the line ending but able to 'replace' with them.
>
> ^, $, etc. are position assertions rather than characters.  They seem
> similar to things like \n or \r, but they're not.  Various editing
> tools seem to handle attempts to find/replace the positions in non-
> intuitive ways, at least as I've seen.
>
> I always thought '^$' would find all blank lines, but in some
> programs it finds nothing at all, because blank lines have a newline
> character, so you have to use '^\n$' as the regular expression.
>
> I'm pretty sure you'll get more predictable results if you always
> include some real characters.  An expression built only out of
> positional assertions will be a frustrating little beast.  (That's
> how it was for me, anyway.)
>
> alex

(some references below for archive completeness re: matching newlines
in OpenOffice)

Absolutely agreed - you start getting very much regex-engine-dependent
behavior when using regexs comprised solely of assertions. I was just
describing what happens to work in that one particular engine. In
short, the OO.o regex engine is known to be flaky about newline... it
treats $ as a positional assertion UNLESS it has nothing to the left,
with the possible exception of the ^ assertion, in which case it
matches the newline.

http://www.oooforum.org/forum/viewtopic.phtml?t=27643
http://www.oooforum.org/forum/viewtopic.phtml?t=32963

There's an LGPL'ed open source workaround, though: IannzFindReplace.sxw from
http://homepages.paradise.net.nz/hillview/OOo/

-A