Re: [ros-dev] [ros-diffs] [mkupfer] 39871: Sascha Clausen <r4v3r AT hotmail DOT de> - Draw bullet for menu radio group. - See issue #4193 for details.

[Ged]

Looking at the patch only this looks wrong to me.
Assuming this is radio button it should use the marlett font.

Ged.

-----Original Message-----
From: [e-mail] [mailto:[e-mail]] On Behalf Of [e-mail]
Sent: 04 March 2009 16:30
To: [e-mail]
Subject: [ros-diffs] [mkupfer] 39871: Sascha Clausen <r4v3r AT hotmail DOT de> - Draw bullet for menu radio group. - See issue #4193 for details.

Author: mkupfer
Date: Wed Mar  4 19:29:57 2009
New Revision: 39871

URL: http://svn.reactos.org/svn/reactos?rev=39871&view=rev
Log:
Sascha Clausen <r4v3r AT hotmail DOT de>
- Draw bullet for menu radio group.
- See issue #4193 for details.

Modified:
    trunk/reactos/dll/win32/user32/windows/draw.c

Modified: trunk/reactos/dll/win32/user32/windows/draw.c
URL: http://svn.reactos.org/svn/reactos/trunk/reactos/dll/win32/user32/windows/dr...
==============================================================================
--- trunk/reactos/dll/win32/user32/windows/draw.c [iso-8859-1] (original)
+++ trunk/reactos/dll/win32/user32/windows/draw.c [iso-8859-1] Wed Mar  4 19:29:57 2009
@@ -1276,11 +1276,12 @@
             yc = myr.top  + SmallDiam - SmallDiam/2;
             i = 234*SmallDiam/750;
             i = i < 1 ? 1 : i;
-            myr.left   = xc - i+i/2;
+            myr.left   = xc - i/2;
             myr.right  = xc + i/2;
-            myr.top    = yc - i+i/2;
+            myr.top    = yc - i/2;
             myr.bottom = yc + i/2;
-            Pie(dc, myr.left, myr.top, myr.right, myr.bottom, xe, ye, xe, ye);
+            // if the start and the end point are equal, Pie() only draws a single line, so start one pixel lower
+            Pie(dc, myr.left, myr.top, myr.right, myr.bottom, xe, ye+1, xe, ye);
             break;

case DFCS_MENUCHECK:

[James Tabor]

Hi!
This is true, it should be marlett font and not the old wine hack
using pie or even eclipse.
Thanks,
James

On Wed, Mar 4, 2009 at 11:56 AM, Ged [e-mail] wrote:
> Looking at the patch only this looks wrong to me.
> Assuming this is radio button it should use the marlett font.
>
> Ged.

[Matthias Kupfer]

Hello,

I would like to make clear some things:
- we generally don't use marlett font for drawing gui elements in draw.c
- I asked some years ago to do this that way (with marlett or even prepared 
bitmaps) and it was rejected by several devs
- It's not the radio button element, the function is similar to radio button, 
but the element looks different (marlett may contain it nevertheless)

So we can change all elements step by step even in draw.c and elsewhere...

Matthias

Am Donnerstag, 5. März 2009 07:57:56 schrieb James Tabor:
> Hi!
> This is true, it should be marlett font and not the old wine hack
> using pie or even eclipse.
> Thanks,
> James
>
> On Wed, Mar 4, 2009 at 11:56 AM, Ged [e-mail] wrote:
> > Looking at the patch only this looks wrong to me.
> > Assuming this is radio button it should use the marlett font.
> >
> > Ged.
>
> Ros-dev mailing list
> [e-mail]
> http://www.reactos.org/mailman/listinfo/ros-dev
-- 
Matthias Kupfer          Telefon [phone]
Wilhelm-Firl-Straße 21   Mobil [phone]
09122 Chemnitz

[Steven Edwards]

Hi,

On Thu, Mar 5, 2009 at 5:16 AM, Matthias Kupfer [e-mail] wrote:
> I would like to make clear some things:
> we generally don't use marlett font for drawing gui elements in draw.c
> I asked some years ago to do this that way (with marlett or even prepared
>
> bitmaps) and it was rejected by several devs
> It's not the radio button element, the function is similar to radio button,
>
> but the element looks different (marlett may contain it nevertheless)
>
> So we can change all elements step by step even in draw.c and elsewhere...

Sorry for being so late bringing this up. I was wondering why you did
not start by importing the latest Wine code for the menus and then
from there start your development. I ask because I am curious if your
new code passes the regression tests. We were using a very old
implementation based off winehq from what 2002? I am sure it had its
share of bugs and I agree the current Winehq implementation still only
uses marlett in one place so its not the 'correct' solution but I know
as far as compatibility goes its currently pretty good.

Anyway it just seems to me that having the tests pass as close to 100%
is the most important thing so starting from a known stable point
seems like it would be a better method, but I admit I've not looked at
either your implementation or the Wine code in depth and I understand
your aiming for the most technically correct solution.

Thanks
-- 
Steven Edwards

"There is one thing stronger than all the armies in the world, and
that is an idea whose time has come." - Victor Hugo