Re: [PATCH] user32/tests: Add tests for LockWindowUpdate.

[Alexandre Julliard]

Lauri Kenttä  writes:
> v2: Win8 and Win10 give different results (sometimes even
> different different results), so mark some cases as broken.
> Also simplify the patch overall.

I don't think that painting on the desktop window is a very useful test,
and it can fail for various reasons. If the goal is to test
LockWindowUpdate, you should do that with top-level windows instead.

[PATCH] user32/tests: Add tests for LockWindowUpdate.

[Lauri Kenttä]

On 2016-12-08 17:58, Alexandre Julliard wrote:
> Lauri Kenttä  writes:
> > v2: Win8 and Win10 give different results (sometimes even
> > different different results), so mark some cases as broken.
> > Also simplify the patch overall.
>
> I don't think that painting on the desktop window is a very useful 
> test,
> and it can fail for various reasons. If the goal is to test
> LockWindowUpdate, you should do that with top-level windows instead.

If painting to desktop is not interesting, then bug 52 (the oldest
open bug) should probably be closed.

[PATCH] user32/tests: Add tests for LockWindowUpdate.

[Alexandre Julliard]

Lauri Kenttä  writes:
> On 2016-12-08 17:58, Alexandre Julliard wrote:
> > Lauri Kenttä  writes:
> > > v2: Win8 and Win10 give different results (sometimes even
> > > different different results), so mark some cases as broken.
> > > Also simplify the patch overall.
> >
> > I don't think that painting on the desktop window is a very useful
> > test,
> > and it can fail for various reasons. If the goal is to test
> > LockWindowUpdate, you should do that with top-level windows instead.
>
> If painting to desktop is not interesting, then bug 52 (the oldest
> open bug) should probably be closed.

As I mentioned in a comment there, that's mostly a WONTFIX. There may be
a use for LockWindowUpdate in other cases, but scribbling over the
desktop is not a particularly interesting case.

[PATCH] user32/tests: Add tests for LockWindowUpdate.

[Bruno Jesus]

On Thursday, December 8, 2016, Alexandre Julliard 
wrote:
> Lauri Kenttä <[e-mail] javascript:;> writes:
> > On 2016-12-08 17:58, Alexandre Julliard wrote:
> > > Lauri Kenttä <[e-mail] javascript:;> writes:
> > > > v2: Win8 and Win10 give different results (sometimes even
> > > > different different results), so mark some cases as broken.
> > > > Also simplify the patch overall.
> > >
> > > I don't think that painting on the desktop window is a very useful
> > > test,
> > > and it can fail for various reasons. If the goal is to test
> > > LockWindowUpdate, you should do that with top-level windows instead.
> >
> > If painting to desktop is not interesting, then bug 52 (the oldest
> > open bug) should probably be closed.
>
> As I mentioned in a comment there, that's mostly a WONTFIX. There may be
> a use for LockWindowUpdate in other cases, but scribbling over the
> desktop is not a particularly interesting case.

A program use case: IDA 5 Free relies on this while doing heavy loading, to
reproduce just try to open any exe file and notice the LockWindowUpdate
calls.

[PATCH] user32/tests: Add tests for LockWindowUpdate.

[Lauri Kenttä]

On 2016-12-08 18:14, Alexandre Julliard wrote:
> Lauri Kenttä  writes:
> > On 2016-12-08 17:58, Alexandre Julliard wrote:
> > > Lauri Kenttä  writes:
> > > > v2: Win8 and Win10 give different results (sometimes even
> > > > different different results), so mark some cases as broken.
> > > > Also simplify the patch overall.
> > >
> > > I don't think that painting on the desktop window is a very useful
> > > test,
> > > and it can fail for various reasons. If the goal is to test
> > > LockWindowUpdate, you should do that with top-level windows instead.
> >
> > If painting to desktop is not interesting, then bug 52 (the oldest
> > open bug) should probably be closed.
>
> As I mentioned in a comment there, that's mostly a WONTFIX. There may 
> be
> a use for LockWindowUpdate in other cases, but scribbling over the
> desktop is not a particularly interesting case.

I thought your comment was about virtual desktop vs. WM integration.
Are you now extending that WONTFIX to virtual desktop as well?
In that case, I'd really close bug 52, which is about the desktop.

I shortly tested IDA 5 Free (as suggested by Bruno), and I didn't notice
any weird glitches. Let someone open a new bug if there's a real 
(useful)
application that really needs this function.

Anyway, I'm sending v3 with parent+child tests instead of desktop tests,
just in case. If that doesn't work out, I'll just move on to another 
bug.

[PATCH] user32/tests: Add tests for LockWindowUpdate.

[Alexandre Julliard]

Lauri Kenttä  writes:
> On 2016-12-08 18:14, Alexandre Julliard wrote:
> > Lauri Kenttä  writes:
> > > On 2016-12-08 17:58, Alexandre Julliard wrote:
> > > > Lauri Kenttä  writes:
> > > > > v2: Win8 and Win10 give different results (sometimes even
> > > > > different different results), so mark some cases as broken.
> > > > > Also simplify the patch overall.
> > > >
> > > > I don't think that painting on the desktop window is a very useful
> > > > test,
> > > > and it can fail for various reasons. If the goal is to test
> > > > LockWindowUpdate, you should do that with top-level windows instead.
> > >
> > > If painting to desktop is not interesting, then bug 52 (the oldest
> > > open bug) should probably be closed.
> >
> > As I mentioned in a comment there, that's mostly a WONTFIX. There
> > may be
> > a use for LockWindowUpdate in other cases, but scribbling over the
> > desktop is not a particularly interesting case.
>
> I thought your comment was about virtual desktop vs. WM integration.
> Are you now extending that WONTFIX to virtual desktop as well?
> In that case, I'd really close bug 52, which is about the desktop.

I don't mind if we keep it open for virtual desktops, but I don't think
any effort should be spent on it until we find a real application that
needs it (and TBH, fixing it would be quite hard, so it would need to
cause a bad problem to make it worth the effort).