Re: [2/3] comctl32: Never draw item text with OPAQUE.

[Felix Nawothnig]

(CCing wine-devel again)

Dimi Paun wrote:
> > Problem is that DrawText() for the subitems will also use the
> > background color - which is wrong.
>
> Why is it wrong?

Because the "background color" isn't the background color of the list. 
And it's definitly not the text bkclr in gdi-context either - I hacked 
together an app to prove that:

http://flexo.popuserv.org/test.c

Now it seems getting rid of the OPAQUE completly is wrong, but for 
different reasons.
> It seems to me that this is a bit of a problem:
>
> 3783     if (nSubItem == 0 || (cdmode & CDRF_NOTIFYITEMDRAW))
> 3784         prepaint_setup(infoPtr, hdc, &nmlvcd);
> 3785
> 3786     /* in full row select, subitems, will just use main item's colors */
> 3787     if (nSubItem && uView == LVS_REPORT && (infoPtr->dwLvExStyle & LVS_EX_FULLROWSELECT))
> 3788         nmlvcd.clrTextBk = CLR_NONE;
>
> I guess the call to prepaint_setup should come _after_ the if
> with FULLROWSELECT, no?

Depends. Right now this code does nothing (and I was going to remove it 
anyway) - the only place where clrTextBk is read from is in 
prepaint_setup() - but prepaint_setup() is not called for subitems (and 
the CLR_NONE override is only done for subitems).
> Also, I can't remember, but there was a reason why prepain_setup
> was called just for the main item, why did you changed that?

Because subitems can use different colors (see test app) and 
prepaint_setup() is where the colors are set up...

Note that I don't change the notifications being passed or anything of 
their context (with the exception of the GDI context in the postpaint 
notification - but the GDI context should be saved & restored in between 
anyway, see the failing test I commited a few days ago) so I can't 
imagine how this could break anything.

Felix

[2/3] comctl32: Never draw item text with OPAQUE.

[Dimi Paun]

On Fri, March 16, 2007 16:37, Felix Nawothnig wrote:
> > It seems to me that this is a bit of a problem:
> >
> > 3783     if (nSubItem == 0 || (cdmode & CDRF_NOTIFYITEMDRAW))
> > 3784         prepaint_setup(infoPtr, hdc, &nmlvcd);
> > 3785
> > 3786     /* in full row select, subitems, will just use main item's colors */
> > 3787     if (nSubItem && uView == LVS_REPORT && (infoPtr->dwLvExStyle & LVS_EX_FULLROWSELECT))
> > 3788         nmlvcd.clrTextBk = CLR_NONE;
> >
> > I guess the call to prepaint_setup should come _after_ the if
> > with FULLROWSELECT, no?
>
> Depends. Right now this code does nothing (and I was going to remove it
> anyway) - the only place where clrTextBk is read from is in
> prepaint_setup() - but prepaint_setup() is not called for subitems (and
> the CLR_NONE override is only done for subitems).

Well, but that's the point: now that prepaint_setup() is always called
(I agree with that change), this would do what we want if only we said
the CLR_NONE _before_ the call to prepaint_setup() where it's being
acted on.

Wouldn't that solve the problem?

[2/3] comctl32: Never draw item text with OPAQUE.

[Felix Nawothnig]

Dimi Paun wrote:
> > Depends. Right now this code does nothing (and I was going to remove it
> > anyway) - the only place where clrTextBk is read from is in
> > prepaint_setup() - but prepaint_setup() is not called for subitems (and
> > the CLR_NONE override is only done for subitems).
>
> Well, but that's the point: now that prepaint_setup() is always called
> (I agree with that change), this would do what we want if only we said
> the CLR_NONE _before_ the call to prepaint_setup() where it's being
> acted on.

Doesn't work because prepaint_setup() overrides the color again. Anyway, 
I'm revising my patches...

Felix