On Wed, Nov 26, 2003 at 01:55:48PM +0000, Angus Leeming wrote:
> > This pure Tcl/Tk solution might be a bit unfair to compare with, but I
> > can assure you that it takes me about ten minutes to add a new module
> > (includeing C++ code) to my long-standing 'toy' application which is a
> > Tcl/Tk Gui with a C++ core.
>
> Shrug. What can I say? Let's return to the old days of xforms in the core?
That's neither desirable nor necessary.
If I had to compute 1+2 in my Gui/Core framework I'd use:
#include "ui.h"
void sum(ui & D)
{
int i1, i2;
D.in >> i1 >> i2;
D.out << i1 + i2;
}
ui_register("compute", "sum", sum);
on the Core side and
puts [compute sum 1 2]
on the Gui side. That's all that's needed for Gui/Core interaction.
The interface can pass arbitrary data.
> Incidentally, I trigger exactly the same crash with your patch as you did and
> I too have no idea why. However, I also think that Insert->Vertical Space
> whould not insert an inset directly. Rather it should launch the VSpace
> dialog, allowing the user to set the variables.
I thought this was taken care of by the code in factory.C:
case LFUN_INSERT_VSPACE:
if (!cmd.argument.empty())
return new InsetVSpace(cmd.argument);
bv->owner()->getDialogs().show("vspace");
return 0;
> In turn, this needs a
> VSpaceMailer class to transfer data between the core and the frontend.
Ah. So I may have to copy some Mailer stuff, too.
> One final thing. Why does 'doInsertInset' not take a 'LyXText const &' as its
> first argument? Works for me...
Don't know. Change it if it works.
Andre'
--
Those who desire to give up Freedom in order to gain Security, will not have,
nor do they deserve, either one. (T. Jefferson or B. Franklin or both...)
