Graeme Geldenhuys wrote:
> Hi,
>
> Was any study (or prototypes) done to help make the choice on how
> Lazarus would implement widget sets?
THere was no study, however during the years, the inner workings of the
LCL <-> Widgetset changed (and is still changing).
> For example:
>
> 1) Wrapping existing native widgets with the LCL (as is currently
> done) vs Writing a native (object pascal) widget sets from scratch
> (like fpGUI or MSEgui) and implementing the same compatible VCL
> interface for that widget set.
As Florian already said, we choose for a native widgetset. That was at
that time one of the key points why I joined toe project.
> 2) Ease of supporting new platforms. Using the current LCL (wrapping
> native widgets) vs having a native (object pascal) widget set, and
> only having to port the painting backend.
This is the same as 1). The ease of supporting new platforms was not an
issue at that time. The whole widgetset interface code has grown bigger
than we imagined at that time.
> I can think of many more such test cases... One vs Another. My main
There is only one testcase and you are using it. You cannot conclude a
thing from writing only a hello world app. The advantages will only show
if you did the complete thing. But that doesn't matter, native widgets
was one of the keypoints.
> reason for asking, is to find out if there was a definitive reason why
> the Lazarus project decided on rather wrapping existing platform
> widgets in a LCL compared to writing a widget set from scratch.
>
> Our company is trying to determine if it will be a viable option to
> hire a few developers to finish the fpGUI project and interface fpGUI
> with Lazarus. Obviously to the extent that Lazarus itself compiles
> and runs using the fpGUI widget set. We don't want to branch Lazarus
> though.
Coding own crossplatform widgetset is one, adding it to lazarus is
another thing. However with fulltime devels, who know the widgetset and
the LCL it should be doable (note that I'm originally a win16/32 api
person, without gtk or carbon experience, so I need to read a lot on
those interfaces)
> This is primary to get Lazarus ready for our commercial
> projects. I got pretty frustrated with Lazarus today and how the
> behaviour of widget sets differ between platforms. [It's been one of
> THOSE days!]
Document those differences !
If they are unknown they won't get solved.
> The other obvious choice for would be to hire some full time
> developers to only work on Lazarus and sort out the issues we are
> experiencing. I can justify the expense compared to buying licenses
> of Delphi. Lazarus gives us a lot more flexibility. IDE wise and
> supported platforms.
>
> In my research so far I think Qt got it right. Implementing their own
> widget set, allowing them to handle all widget internals - events,
> painting, behaviour on different platform, etc.. With the fpGUI
> project I am experience that same result.
Yes, and thats for me one of the reasons why I try to avoid Qt apps
under windows if I have a choice. They feel different/primitive
Marc
_________________________________________________________________
To unsubscribe: mail [EMAIL PROTECTED] with
"unsubscribe" as the Subject
archives at http://www.lazarus.freepascal.org/mailarchives
