Re: [Denemo-devel] A Transformation of how Denemo is perceived.

[Richard Shann]

On Wed, 2012-09-26 at 23:25 +0200, Andreas Schneider wrote:
> Hi,
> 
> in principle, a regular update of the preview is a nice thing. But for
> continuous live-compilation with lilypond at least my computers (a
> netbook bought this year and a desktop a few years old) are way too
> slow, judging from how long it takes to update the preview.

yes, this is all triggered by my purchase of a new (quiet) computer - it
seems they only make quiet computers with lots of muscle, my new
computer has 4 cpus and 8 times the memory of my old one.
Note, however that you can configure Denemo to typeset just a small
context about the cursor - even a single measure. And it may not affect
your use of Denemo, that is you may not find the response to your
keypresses getting sluggish depending on how the background process is
handled. So, the updates might be a bit slow in arriving, but better
late than never.

>  The question
> is, what hardware is the average user assumed to have? I would suggest a
> setting for the time interval in which the compilation is triggered.

I am not sure if having blank times when no attempt to start a LilyPond
process is made would help. On my quad-core box I notice that top
reports the cpu usage as 100% or sometimes 101% or 102% while the
background typesetting is turned on. It goes to these values even when
no typesetting is actually taking place (because the score is just
standing there unaltered). I can only guess that this is the idle call
going mad. But why that should be is a mystery to me as there are a
number of other idle calls (for the audio threads) which are also
active. So I may have some fault in what I am doing to check if a
re-typeset is needed.
I would be very interested to hear about performance on other systems -
I will try to do some testing myself.

> Also, on the insertion of certain elements (e.g. start of a beam or
> slur) necessarily an error will occur until the finishing element is
> set. One has to think of how to deal with that.
What happens is that the display is marked as "Possibly Invalid" - at
the moment I am having trouble detecting whether it is actually invalid
or has just been generated within 1 second of the last update...

> 
> Regarding the edition of lilypond code, I think this is one of the most
> useful things, as there are situations where one has to go beyond the
> direct features of Denemo.
Absolutely, nothing is being changed there.

Richard


> 
> Andreas
> 
> _______________________________________________
> Denemo-devel mailing list
> address@hidden
> https://lists.gnu.org/mailman/listinfo/denemo-devel