Jürgen Spitzmüller ha scritto:
Pavel Sanda wrote:
i wouldn't do this until we have something really usable instead...
my thinking went rather on Find & Replace (Experimental )... ;)
I know you are joking, but if it still deserves the attribute "Experimental"
at release time, I'll vote for disabling it.
Just to mention that I'm trying to get rid of all the instabilities, in
order to get rid of the "Experimental" attribute (and on my laptop,
which comprises also some patches that have not been (yet) incorporated,
I don't get any crash using it), however I still need some help for
identifying common use cases and possible issues or deviations from the
expected behaviour.
Considering that it searches also inside maths, it will be anyway slower
than the standard search, which completely ignores math insets, however,
consider that:
-) searching for a non-existent word within the "UserGuide.lyx" or
"Extended.lyx" takes currently ~1sec before you get the question "search
from beginning ?"
-) searching for "$\hat{H}_{kl}$" in "Math.lyx", which appears almost at
end of file, takes currently ~1.5 sec before the match is highlighted
(and there's no way to find it with the standard text-only search facility)
-) the above (qualitative) benchmarks refer to the development version
compiled with "--enable-debug" and "--enable-stdlib-debug", running on a
"Intel(R) Core(TM)2 Duo CPU T7700 @ 2.40GHz" (where of course only
1 core is used for the search)
I think the performance of this feature is acceptable for a release.
Finally, you can still use the traditional simpler search facility for
common searches within text. If you don't use the new searching
facility, it won't hurt you anyway.
Concerning the names to use (menù, dialog and tooltips), I'm open to any
suggestion. As of now, the
-) "Find & Replace (plain text)"
-) "Find & Replace (formatted)"
seem to me the most "communicative" names.
Bye,
T.