Re: zoom (was Re: commit (head): 7610)

From: <msevior_at_physics.unimelb.edu.au>
Date: Tue Nov 02 2004 - 10:24:34 CET

>
>
> Hi Dom,
>
>> The whole point of layout units is that you could use
>> *REAL* OS supported tranformation matrices/scaling if
>> they're supported. The important functions are all
>> virtual. Device Units and Layout units can be the same
>> thing (i.e. 1x scale).
>
> Having grepped through the code, I think this might be possible to do
> this on win32 without significant changes to the xp code, as it seems
> that outwith the graphics classes the tlu transformation is only
> applied to constants. It is possible that the only change required to
> the xp code is to add virtual GR_Graphic::_setZoomPercentage() to be
> called from the public function to allow for platform specific
> processing. I will investigate this further. If all the other changes
> will be restricted to win32, I will just go ahead and do it. If there
> are further issues, I will consult further.
>

Sounds good. I had a hunch that you could bury most of this in the win32
graphics class.

Regarding the desired behaviour my priorities would be:

1. Lines break on win32 at the point for all zooms
1. Lines break on win32 at the same point when printing
1. Pages break on the same line at all zooms
1. Pages break on the same line when printing
2. Words and widths look nice when printing.
3. Words and width look nice at 100% or 200 % only
4. Words and widths look nice at all zooms
5. Lines and pages pages break at the same place as other platforms.

Getting the 1's right are absolute blockers for 2.2
Getting down to 4 right is worthwhile if it can be done "quickly"

5 would be nice but I'd be very happy if you can get to 4 for 2.2

Cheers

Martin
Received on Tue Nov 2 10:23:10 2004

This archive was generated by hypermail 2.1.8 : Tue Nov 02 2004 - 10:23:10 CET