[Development] Removing Wacom support in Qt5

[Boudewijn Rempt]

On Sun, 9 Sep 2012, Thiago Macieira wrote:

> On domingo, 9 de setembro de 2012 10.55.10, Frank Hemer wrote:
>>> Note that this also applies for Qt 4.8. If the support is broken there and
>>> causing issues like Ariel is describing, it's definitely a P1 that needs
>>> fixing. If we can't muster the knowledge of how to fix it properly, we
>>> should disable.
>>>
>>> Now, it appears that in Qt 5, the support is already non-existent. So
>>> there's nothing we need to do now.
>>
>> You cannot be serious about that. Its not only that some commercial products
>> depend on wacom support, but with a growing qml usage, more and more
>> applications may need it. Its a definite showstopper for one of my products
>> forcing me to stick with qt4 for the time being.
>
> What were you referring to when you said "you cannot be serious about that"?
>
> If you meant my proposal to disable in Qt 4, then let me explain: the bugs
> that Ariel is reporting are definitely P1. From past history, we know that
> Wacom support breaks every other release. We need to do something about it.
> Preferably, it should be fixing it once and for all.

Well, given that Digia actually seems to have a patch for the bug in 
question, I'd say that the choice is easy. Use the bug fix. If Qt goes 
down the way of "it's better to break existing applications than to allow
in a patch not everybody likes", then Qt will lose its reputation for
dependability completely.

> The other extreme is to
> disable it completely.

Which is of course even more unacceptable. Breaking applications by 
removing a feature that's been around for many years is even worse.

> A middle-ground is to figure out what is different from Ariel's scenario, maybe
> blacklisting some driver versions or device versions or something.
>
> If you meant that the Qt 5 support is non-existent, I was serious. It's a fact
> and I wasn't the one who reported it. All of the input method, pointer and
> window management systems were rewritten for Qt 5 and, apparently, Wacom
> support on Windows wasn't re-added yet.
>
>> I'm a commercial license owner (hello digia, whats your standpoint?), but my
>> resources are limited so I cannot support with active contribution. But
>> anyway - simply stating that the feature is not supported (it used to be
>> supported in the qt4 days and it was never mentioned to be deprecated or
>> alike) is imho the very wrong direction.
>
> The feature doesn't *exist*. We can't support code that isn't there.

If there's api for something, it is unreasonable to say that the feature 
doesn't exist. Not having tablet support on Windows and OSX while having 
QTabletEvent in the api is a bug. There's a promise that something should 
work, and it doesn't work.

> We need someone to write it again. Someone who understands that code.
> Hopefully, it should be properly integrated into the QPA input system so that
> it doesn't break as often as the Qt 4 solution did.

Boudewijn