On Apr 10, 2007, at 10:40 AM, Abdelrazak Younes wrote:
Bennett Helm wrote:
On Apr 10, 2007, at 9:18 AM, Abdelrazak Younes wrote:
Bennett Helm wrote:
On Apr 10, 2007, at 8:37 AM, Abdelrazak Younes wrote:
Testing if typing lags behind within either a math or a caption
inset. If the delay is acceptable on a PPC platform we can
ditch the wide() hack.
I just tested on 800 MHz PPC Mac with an unpatched, post-beta1
version of LyX (again with Qt-4.2.3). There was a noticeable lag
when trying to enter text as fast as I could, even when not in
an inset. It could just keep up with me when typing normally. I
did not notice that insets were any slower than normal text.
Wait... do you mean that typing within a math inset is not slower
than within normal top level text? If it is not slower then we
the trouble we're having with the wide() hack is not worth it.
I'm not following you here. As I understood it, it's not slower
precisely because of the wide() hack; prior to that hack, typing
in insets was *much* slower.
Right, this is precisely the reason why I've asked you to test
within mathed or the caption inset where this hack is not used. If
typing within caption for example is the same as within normal text
(or within Foot note which uses the hack) then we don't need the
hack anymore.
OK, this is making more sense. I've tried the caption inset on my
fast Intel Mac, and I can -- just barely -- type faster than LyX can
keep up. (LyX uses 100% of one of the CPUs in this case.) I'm sure it
would be much worse on PPC. That argues all the more strongly for
keeping the wide() hack.
On the other hand, this is on Qt-4.2.3. If you really want, I can try
Qt-4.3, but not today!
Bennett
