Re: empty document at Graphos startup

[Riccardo Mottola]

Hi,

thanks for looking into the issue.
> > I guarded updatePrintInfo so that returns without setting the page rect if the 
> > information is clearly invalid (nil or<= 0). However, this doesn't fix the 
> > problem (commited though, it doesn't harm surely).
> >
> > the pageRectSize is set in the designated initializer.
> >    pageRect = NSMakeRect(0, 0, 695, 942);
> > line 47
> >
> > I notice two interesting thigns:
> > 1) updatePrintInfo is not called on startup witha  new blank document for me.
> In that case you should get a log message "printInfo nil!". It means that the 
> document's -printInfo method returns nil (maybe because you have no printer configured?). 
> In that case the size of the GRDocView created in -init is not changed and that view is 
> created with (0,0,0,0) so it is certainly not visible!
Well, initWithFrame of GRDocView essentially ignores the parameter and 
sets some default stuff.
I don't get printInfo nil.

My startup looks like:
grid$ Graphos.app/Graphos
2011-12-20 09:25:31.023 Graphos[5151] No theme named 'Neos' found
Xlib:  extension "MIT-SHM" missing on display "dev11:0.0".
2011-12-20 09:25:31.418 Graphos[5151] initing doc view with win
2011-12-20 09:25:31.438 Graphos[5151] initing doc view with win
2011-12-20 09:25:31.518 Graphos[5151] Scanner initialised with nil string
2011-12-20 09:25:31.536 Graphos[5151] initing GRToolsWindow
2011-12-20 09:25:31.543 Graphos[5151] initing GRToolsView
2011-12-20 09:25:31.549 Graphos[5151] inited GRToolsWindow

> > 2) the init method is called twice (I get "initing doc view with win" on the 
> > console)
> Of course it gets called twice. Once for the GRDocView that you create in 
> GRDocument's -init method and once for the GRDocView that is unarchived from 
> the .gorm file. Since you don't create a window in the -init method, I think it 
> is pointless to create a GRDocView there. You really should use the document 
> view that is present in the gorm file (by linking it to the docView attribute 
> of the document in Gorm).
>
> BTW, I've noticed this bogus code in the -init method:
>      [super init];
>      if (self)
> This should read
>      self = [super init];
>      if (self)
Yes, that must be a leftover of the old code I refactored. My usual 
style is the verbose line you show here. I fixed that. It did not help 
though, but thanks for the catch.

I'll check in the double initialization too. It is probably just useless 
and not harmful.

Riccardo

> > 3) I also get "Scanner initialised with nil string", I haven't debugged yet 
> > where it is coming from.
Do you get this one? it might be a clue...