From: james <james@mansionfamily.plus.com>
<= span style=3D"font-weight: bold;">To: Dimitar Dobrev <dpldobr= ev@yahoo.com>; KDE bindings for other programming languages <kde-bind= ings@kde.org>
Sent: Friday, July 12, 2013 9:32 PM
S= ubject: Re: [Kde-bindings] Qt & SWIG

=0A =0A=0A =0A= =0A
=0A
What is Qt-J= ambi doing, or are they in=0A difficulty too?
=0A
=0A = Qt seems to shed bindings like they are going out of fashion.
=0A =
=0A I can imagine that's no problem for Phil with PySide, bit perh= aps=0A a little less NIH elsewhere would work.
=0A
=0A = I mean - maybe the Qt-Jambi system could be extended? Even if it=0A= means going through IKVM?
=0A
=0A Or maybe the PySid= e stuff could be reused - even if the glue is=0A IronPython?
=0A =
=0A James
=0A
=0A
=0A On 12/07/2013 1= 3:41, Dimitar Dobrev wrote:
=0A
=0A
=0A
=0A=
=0A

=0A =
=0A
   =0A Hello, Ruth,
=0A =

=0A
= =0A
    Glad= =0A to hear from you again. I am sorry Qyoto didn't work out= =0A for you but I do agree SWIG would be a better solution, with better performance indeed. I'll= =0A try to find free time to work on it as soon as possible.= =0A However, I cannot make any promise yet.
=0A =

=0A =0A
    Bes= t=0A regards,
=0A
   =0A Dimitar
=0A
=0A
=0A
=0A
=0A
=0A =
=0A
From: Ruth= =0A Ivimey-Cook <ruth@ivimey.org>
=0A = To:=0A = kde-bindings@kde.org
=0A Sent:=0A Friday= , July 12, 2013 2:41 PM
=0A Subject:=0A Re: [Kde-bindings] Qt &am= p; SWIG
=0A
=0A

=0A
=0A
=0A
Dimitar=0A and others,
=0A =
=0A I evaluated using Qyoto for= a project I'm working=0A on and - with regret for all= the hard work Dimitar=0A had obviously put in - decid= ed against it, both=0A because it was buggy and becaus= e it was very slow.=0A When I investigated the speed i= ssue SMOKE relies=0A on textual matching of function s= ignatures to call=0A functions. Admittedly the hashes,= once matched,=0A are hashed and cached, but still...&= nbsp; It also=0A involved at least 2 transitions= from managed to=0A unmanaged code, which IIRC are exp= ensive in=0A themselves.
=0A =
=0A People I know have used SWIG to good effect bu= t I=0A have no personal experience.
=0A =
=0A A statically linked shim file (not a= DLL, if=0A possible) - i.e. a set of functions coded = in=0A managed code that each make an unmanaged call to= =0A the "real" function would be the way I would=0A = expect. Coding this by hand would be painful, so a=0A = program to generate them is the obvious response.=0A = I believe this is the SWIG way.
=0A =
=0A If it were possible I would look into using= =0A something like an XML file containing the=0A = interface definition, maybe itself initially=0A = generated from the .h files, and which can then be=0A = adapted and enhanced to improve it. If the Qt=0A = and/or SWIG community were receptive this might=0A = become something people would support generally -=0A = e.g. including PyQt et al. - which would benefit=0A e= veryone.
=0A
=0A HTH
= =0A Ruth
=0A
=0A =
=0A Dimitar Dobrev wrote:
=0A =
=0A
=0A =

=0A =     Dylan,
=0A =
=0A     Than= k you for your=0A suggestion. I know about = CXXI but it=0A wouldn't be my choice for the followi= ng reasons:
=0A =     1. It is both=0A incomplete and= abandoned;
=0A =     2. It=0A = relies on Reflection.Emit which is not=0A = supported on iOS; Qt will soon officially run=0A = on iOS and it'd be nice if the bindings worked on that=0A platform = too.
=0A
=0A =    =0A = Regards,
=0A    =0A Dimitar=0A    =0A

=0A
=0A =
=0A
=0A =
=0A <= hr size=3D"1"> =0A = From:=0A = "Moonfire, D." <d.moonfire@mfgames.c= om>
=0A To:=0A KDE bindings fo= r other programming=0A languages &= lt;kde-bindings@kde.org>=0A
=0A = Sent:=0A Thursday, July 11, 2013 9:06 PM<= br>=0A = Subject:=0A Re: [Kde-bindings] Qt= & SWIG
=0A
=0A =

=0A =
=0A =
On=0A = Thu, Jul 11, 2013 at 10:25 AM, Gour <gour@atmarama.net>=0A = wrote:
=0A On Thu, 11=0A = Jul 2013 06:31:43 -0700 (PDT)
=0A = Dimitar Dobrev <dpld= obrev@yahoo.com>=0A=0A wrote:
= =0A
=0A = > 3. I don't want to have=0A a= nything to do with SMOKE any more=0A so = if you'd
=0A > like to take this p= ath, you'd=0A be on your own.
=0A =
=0A = Isn't it pity that there is no more=0A i= nterest in having Qt bindings for
=0A = the .NET/Mono? :-(
=0A
= =0A
=0A =
=0A There was an announcement some year= s=0A back on one of the Mono blogs about an= =0A interop layer for C++ (https://github.c= om/mono/cxxi).=0A=0A It was never mentio= ned again and I think=0A a grue ate it, but = it looked like a=0A really nice set of libra= ries if someone=0A maintained it beyond the = initial proof=0A of concept. I also don't kn= ow how=0A tightly it is tied to Mono itself = and if=0A a cross-platform version could be = done=0A with it.
=0A =
=0A Looking at the code, they= were also=0A using Qt as the example in the= ir POC.
=0A
=0A = - Dylan
=0A
=0A =
=0A_______________________________________________=0A Kde-bindings mailing list
=0A = Kde-bindings@k= de.org
=0A htt= ps://mail.kde.org/mailman/listinfo/kde-bindings
=0A =
=0A
=0A =
=0A
=0A =
=0A
=0A
=0A =
=0A
=0A
_______= ________________________________________=0AKde-bindings mailing list=0AKde-bindings@kde.org=0Ahttps://mail.kde.org/mailman/listinfo/kde-bind= ings=0A
=0A
=0A =
=0A
=0A
-- =0ASoftware Manager & Engineer=0ATel: 01223 41= 4180=0ABlog: http://www.ivimey.= org/blog=0ALinkedIn: http://uk.linkedin.com/in/ruthivimeycook/
=0A =
=0A
=0A
=0A = _______________________________________________
=0A = Kde-bindings mailing list
=0A Kde-bindings@kde.org
=0A https://mail.kde.org/mailman/listinfo/kde-bindings
=0A =
=0A
=0A

=0A =