[Qt5-feedback] SWF, IPC, & QtService...

BRM bm_witness at yahoo.com
Tue Jun 14 21:52:41 CEST 2011


Starting a new thread so as not to hijack the other one...


----- Original Message ----

> From: Peter Kümmel <syntheticpp at gmx.net>
> On 14.06.2011 21:09, Thiago Macieira wrote:
> > Em Tuesday, 14 de June de  2011, às 20:55:08, Peter Kümmel escreveu:
> >> Reading "majority of use  cases" I wonder why it is not used by QtCreator? 
>Is
> >> this only due to  the release of SFW after  starting QtCeator
> >> development, are  are there also technical reasons to not use SFW in
> >>  QtCreator?
> >
> > Why does Qt Creator need SFW for?
> 
> In an  initial blog of SFW,
> 
>       http://labs.qt.nokia.com/2009/05/26/qt-service-framework/
> 
> I read "This  initial release focuses on in-process".
> And QtCreator also has a  infrastructure to load and unload services,
> they call it  "ExtensionSystem":
> 
>      http://doc.qt.nokia.com/qtcreator-extending/extensionsystem.html
> 
> It's  only in-process but at a first glace it looks that SFW
> could also be used for  the job.
> 

The blog notes that out-of-process IPC will eventually be supported.
Has this been done? Or is it planned for Qt5?

If the QtService replacement is to use SFW to communicate between controller and 
daemon - which from some of what I read on SFW may make it a very nice fit for 
doing so - then the out-of-process IPC would certainly be necessary.

Also, how does Windows COM come into play on this? If a module like the 
QtService replacement were to use SFW on Windows and COM were the back-end to 
SFW then would two programs be able to use different versions of the 
QtService/SFW backend? Or would they conflict with each other? (E.g. different 
compilers)

Just trying to understand how we might be able to use SFW for the back-end comms 
in the new QtService replacement - why reinvent the wheel if we don't have to?

Thanks,

Ben



More information about the Qt5-feedback mailing list