Pen Widths, and patch etiquette

[Rob Brown]

Hi,

I'm trying to get Protel 99SE working under Wine. It's nearly there, in
fact it's totally usable, but some slight niggles remain (as can be seen in
the appDB). I'm a coder, and I am actively looking at the code to see what
I can do. I'm also a total newbie to Wine development, and while I'm
reading the ML archives and generally trying to sponge up as much as I can
by myself, I'd prefer a bit of advice!

My first job was to find a problem where some lines are drawn a single
pixel wide, when they should be several pixels wide. After some digging, I
found that the DC is in MM_ISOTROPIC mapping mode, and that a line with
width 1 is not being transformed. Wider lines do get transformed, and lines
of width 0 are correctly drawn a single pixel wide.

The culprit in this case is line 67 of dlls/winex11.drv/pen.c, which says

   physDev->pen.width = logpen.lopnWidth.x;
** if ((logpen.lopnStyle & PS_GEOMETRIC) || (physDev->pen.width > 1))
   {
      physDev->pen.width = X11DRV_XWStoDS( physDev, physDev->pen.width );
      if (physDev->pen.width < 0) physDev->pen.width = -physDev->pen.width;
   }

If I change that line to

   if ((logpen.lopnStyle & PS_GEOMETRIC) || (physDev->pen.width > 0))

my problem goes away, but I'm concerned that it isn't appropriate for other
mapping modes. I can do something a bit more conservative like

   if ((logpen.lopnStyle & PS_GEOMETRIC) || (physDev->pen.width > 1) ||
      ((GetMapMode ( physDev->hdc ) == MM_ISOTROPIC) &&
       (physDev->pen.width == 1)))

and it Works For Me (tm).

I've tried to do some tests in Windows to see what happens in different
mapping modes, and I'm getting there, but this is new stuff for me so it's
fairly slow going. In particular, it seems that the line with in
anisotropic map mode will depend on whether the line is vertical or
horizontal etc.

So, I have a couple of questions:

1) Is the mailing list or bugzilla a better place to discuss this sort of
thing?

2) Should I continue to investigate the issue in Windows, or is there some
gdi guru out there that can save me the effort by telling me all about pen
width mapping?

3) What is the normal flow of things here: do people fix their bug
specifically and then move on, or is it preferred to investigate the issue
fully and make sure the solution is truly correct?

Thanks for your time!
Rob.

[Joris Huizer]

You may want to try and write a regression test (a small amount of C 
code calling the function/functions in such a way that the problem you 
see can be seen as output value) to show what windows does in this case; 
if then the values are corrected under wine with your changes, it shows 
you're going in the right direction :-)

HTH,

Joris

[Kuba Ober]

On Tuesday 11 July 2006 18:01, Rob Brown wrote:
> Hi,
>
> I'm trying to get Protel 99SE working under Wine. It's nearly there, in
> fact it's totally usable, but some slight niggles remain (as can be seen in
> the appDB).

Does it install out of the box, or does it need some Windows components? 

Did you use the installer, or copy the installation from a Windows machine?


I have written a unix utility to convert a .ddb to a .zip, if you're 
interested let me know and I'll sanitize it a bit and can post it under some 
free software license.

Cheers, Kuba

[Huw Davies]

As has been mentioned already a Wine regression test would be ideal
here.  In this case it's slightly tricky as you're testing a graphical
component however it's still achievable.  I'd suggest you add a test
to gdi/tests/pen.c that creates a 1bpp dibsection, selects that into a
dc, draws lines on it with a selection of pens in different graphics
modes.  After each line get the bits from the dibsection and compare
them to what you've found out Windows gives you and then clear the
dibsection before drawing with the next pen/graphics mode.  If you
need any help with this then please let me know.

Huw.
-- 
Huw Davies
huw(a)codeweavers.com