[cairo] Drawing to multiple surfaces, "cairo_recorder", zoom

[Hakki Dogusan]

Hi,

Tuom Larsen wrote:
> 
> 
> On Dec 14, 2007 11:13 PM, Carl Worth <cworth at cworth.org 
> <mailto:cworth at cworth.org>> wrote:
> 
>     On Fri, 14 Dec 2007 22:47:40 +0100, "Tuom Larsen" wrote:
>      > Hello. Yesterday's answers cleared up a lot but they provoked
>     some more.
>      > Please:
> 
>     No problem, that's what we're here for. :-)
> 
>      > - Say an application draws to a window and optionally, if a user
>     would like
>      > to, it exports the drawing to PDF. How?
> 
>     First of all, an application, (at least one targeting X or GTK+), must
>     be prepared to redraw its contents, (or any portion of its contents),
>     at any time. So an application has to have code for that already.
> 
> 
> But the drawing would be a raster image embeded in PDF, no? Well unless 
> I switch to a vector surface target, as you suggest bellow?
>  
> 
> 
> 
>      > If I understood Carl's
>      > recommendation correctly -- to use "data structures [which] know
>     how to
>      > 'draw themselves' given a cairo context -- I probably could write
>     some dummy
>      > drawing functions (lineto, circle, stroke, ...) which would just
>     record [1]
>      > the drawing operations (and serialize to disk, e.g. in a text
>     file). Later,
>      > I could parse that file and execute the real drawing function of
>     a given
>      > surface. Does it make sense to you?
> 
>     No, that's not what I was describing at all. I'm just talking about
>     using the same code that made the stuff appear on the window in the
>     first place. Call that same code again with a cairo_t* targeting a PDF
>     surface instead of the cairo_t* targeting a surface for the window,
>     and you should be done.
> 
>     Am I making sense yet?
> 
> 
> Yes, I see, thanks. But consider this:
> - The drawing is the result of some heavy computation - now it would 
> have to run twice.
> - Or even worse, the result of some stochastic processes.
> - How to pass the targeting surface through network or pipe? (Simple, 
> you [I] don't. You [I] send the drawing source code.)
> Please, I certainly don't want to exaggerate.
>  
> 

Regarding "cairo_recorder", I did something similar for 
printing/previewing in my wxPrintCommandProcessor. As suggested before, 
it records wxDC's drawing commands in a list, (so, only result of heavy 
process is required to saved), then it replayed to either to screen or 
printer. If I would do it for Cairo, it would look something like :

struct command {
   virtual ~command(){}
   virtual void play(cairo_t *cr)=0;
};

command_line_to : public command {
   double m_x, double m_y;
   command_line_to(double x, double y):m_x(x), m_y(y){}
   virtual void play(cairo_t *cr){ cairo_line_to (cr, m_x, m_y); }
};

etc.

Hence, instead of calling Cairo API directly, generate commands, save 
them somewhere, than iterate & play them against a real cairo_t.


--
Regards,
Hakki Dogusan