[PATCH v4 07/10] shell32/autocomplete: Redesign the window proc to trigger on key presses instead of key release

[Huw Davies]

On Tue, Sep 11, 2018 at 08:04:50PM +0300, Gabriel Ivăncescu wrote:
> On Tue, Sep 11, 2018 at 2:54 PM, Huw Davies <huw at codeweavers.com> wrote:
> >
> > No.  I'm not signing off on this.  You're going to have to come on with
> > a much cleaner way to do this.
> >
> 
> What specifically? The goto after the displayall stuff, or the three
> nested ifs there? Obviously, my first thought was to just break off on
> the first if, but then it would change the indentation and the patch
> would be much larger (same as with my initial patches with heap_alloc
> failure). I can't split it either, since otherwise the label wouldn't
> exist and it would fail to compile. (even if it did compile being
> split, it would amount of wrong usage results, half of it WM_KEYDOWN
> and no WM_CHAR etc)
> 
> If it's the goto... I don't know how to handle it differently without
> some other ugly complication, or a lot more code / changes.

Specifically the gotos, but I've already commented on the high
identation level in this function (I know that's not your fault).

It seems to me that you want the WM_KEYDOWN handler to call the
WM_CHAR handler.  At the risk of repeating myself, having the handlers
as a helper functions would make this easier.

Huw.