mailing-list for TeXmacs Users

[Dennis Nezic <address@hidden>]

On Mon, 4 Nov 2019 14:03:09 -0500, Dennis Nezic wrote:
> On Sun, 3 Nov 2019 04:11:21 +0100, TeXmacs wrote:
> > Hi Dennis,
> > 
> > On Thu, Oct 31, 2019 at 12:16:29PM -0400, Dennis Nezic wrote:
> > > I was finally able to compile 1.99.11 without Qt, and it was
> > > significantly faster, I think basically the same as the 1.0.7
> > > lagginess that I was used to. For example, with the Qt5
> > > compilation, simply pressing Home and End on a line of basic text
> > > took about a second at 100% cpu, whereas with the X11 port it took
> > > about half a second and about 50% cpu. I wish both could be a lot
> > > faster, I'm not sure why this would be cpu intensive, but with Qt5
> > > it was unbearable.
> > 
> > I would be interested trying to reproduce the slowness that you
> > are experiencing.  Can you please send me a simple test document
> > that demonstrates the problem and tell me where to put the cursor
> > and press Home and End?  What kind of computer are you working on?
> > 
> > The algorithm for Home and End may be somewhat suboptimal,
> > but 0.5 or 1 second is too long for ordinary text.
> 
> Using a couple of QDateTime::currentMSecsSinceEpoch() printf's in
> src/Edit/Interface/edit_keyboard.cpp in handle_keypress, I recorded
> the following times (in milliseconds) between the top and bottom of
> the function:
> 
> 1094
> 1086
> 1097
> 1115
> 1089
> 1091
> 1094
> 1090
> 1089
> 1091
> 1083
> 1093
> 1083
> 1241
> 1084
> 
> Just by alternatively pressing Home and End on a line of text such as:
> "abcdefghijklmnopqrstuvwxyzabcdefghijklmnopqrstuvwxyz"
> 
> Using a shorter string of text like "abcdef" produced the following
> much shorter delays:
> 
> 35
> 27
> 29
> 33
> 28
> 38
> 
> (I'm working on a 64bit amd-cpu 1.6Ghz laptop)

The delay is in the call too "keyboard-press", which calls "key-press"
in TeXmacs/progs/kernel/gui/kbd-handlers.scm. So it looks like some
kind of issue with scheme/guile?