Hallo ROOTers,
sorry for the late answer.
The advantage of threads in comparison to some timer treatment is obvious,
as far as I'm concerned. This was the major reason for us to deal with the
TThread-class in ROOT for our GO4-project.
We finally found a solution which might allow to perform also a graphics
update from within the ROOT thread class. This is already in use by us
here and by Martin Purschke at BNL. It is not yet perfect, but it is in my
view very close to a usable system (still the problem of object creation
and destruction in threads remains, esp. in lists).
Thanks to Victor for his help on this topic and to Rene, who included our
changes as soon as they worked into his development version. As far as I
see it, you can try it and work with it (also for your graphics!).
Greetings,
Marc
Dr. Marc Hemberger
|\ _,,,---,, GSI, Abteilung DV&EE
ZZZzz /,`.-'`' -. ;-;;,_ Planckstr. 1
|,4- ) )-,_. ,\ ( `'-'
'---''(_/--' `-'\_) 64291 Darmstadt
M.Hemberger@gsi.de
** Disclaimer: My views/comments/beliefs, as strange as they are, are my own.**
On Thu, 21 Oct 1999, Onuchin Valeriy wrote:
> KOSU_FOKIN@garbo.lucas.lu.se wrote:
> >
> > Hello!
> >
> > After all these changes and work around performed by GSI groop ... Do we
> > have any chance to get ROOT graphical objects working in multithread env.?
> > This is the most important to handle because it is not much use of ROOT's
> > TThread if you can't put your Canvas update in a thread.
> >
> > Best regards,
> > Anton
>
> Hi Anton,
> one of the alternative to multiheading is using of
> QTTimer class. Look how it is done at
> http://emcal06.rhic.bnl.gov/~onuchin/Rt/examples.html
>
> hsimplem.C example.
>
> One of the advantage of this approach is possibility
> to set "dynamic priority" to thread by dynamic changing
> of the timer interval.
>
> Best regards. Valeriy
>
This archive was generated by hypermail 2b29 : Tue Jan 04 2000 - 00:43:41 MET