Re: gdiplus: GdipTranslateLineTransform: make stub lie and return Ok

[Dan Kegel]

On Sat, Feb 11, 2012 at 3:29 PM, Dan Kegel <dank(a)kegel.com> wrote:
> This lets many .net apps run, see bug 23759.

I see this patch is marked pending.

Since tweaking the stub to lie and return success lets a bunch
of .net apps run, it's tempting to say the benefits of the patch
outweigh the risk... wine has done this kind of hack before,
though maybe not during code freeze.

Can someone who knows more about gdiplus comment?
- Dan

gdiplus: GdipTranslateLineTransform: make stub lie and return Ok

[Vincent Povirk]

Well, it'll probably just mean we render the brush incorrectly, and
the program goes on with what it was doing.

It's theoretically possible for a program to break in a case where
winforms currently catches the exception in .NET (instead drawing the
famous red X) but will instead move on and break in a worse way than
rendering incorrectly. That's not a likely case, but my understanding
is that this is the rationale for not allowing new features during
code freeze.

On Tue, Feb 14, 2012 at 11:11 PM, Dan Kegel <dank(a)kegel.com> wrote:
> On Sat, Feb 11, 2012 at 3:29 PM, Dan Kegel <dank(a)kegel.com> wrote:
> > This lets many .net apps run, see bug 23759.
>
> I see this patch is marked pending.
>
> Since tweaking the stub to lie and return success lets a bunch
> of .net apps run, it's tempting to say the benefits of the patch
> outweigh the risk... wine has done this kind of hack before,
> though maybe not during code freeze.
>
> Can someone who knows more about gdiplus comment?
> - Dan

gdiplus: GdipTranslateLineTransform: make stub lie and return Ok

[Dan Kegel]

On Wed, Feb 15, 2012 at 7:37 AM, Vincent Povirk <madewokherd(a)gmail.com> wrote:
> Well, it'll probably just mean we render the brush incorrectly, and
> the program goes on with what it was doing.
>
> It's theoretically possible for a program to break in a case where
> winforms currently catches the exception in .NET (instead drawing the
> famous red X) but will instead move on and break in a worse way than
> rendering incorrectly. That's not a likely case, but my understanding
> is that this is the rationale for not allowing new features during
> code freeze.

Just to be clear: without this patch (or a real fix),
.net apps that turn on theming crash immediately.

I suppose Alexandre's thinking is that if users are going to load
native .net, they may as well also load native gdiplus, at least
for now.
- Dan