[lazarus] LCL, general issues

[Stefan Hille]

Hi

On Wed, Jul 05, 2000 at 10:48:23AM +0200, Marc Weustink wrote:
> [...gtk errors...]
> + not sure if they don't have side-effects. So if it won't take you
> + too much time please fix these issues.
> 
> OK you're right. Since it is the last message you see, you expect the
> problem there while it is somewhere else.
> Fixing may take a while (or not)

Since we're all doing this in our spare time just take the time you need.

> since I was that fustrated yesterday about
> the new IDE code (while debugging) that I decided to do some rewrite/cleanup
> before it gets completely out of hand.
> At this moment I think the code is absolutely unreadable, uncomprehensable
> and not mainainable. [.......]

I never took a closer look at the IDE code but I think we're running in the same
direction: Some parts of the lazarus/lcl project suffer from their history
and need some cleanup.

Beeing at this point I think it's time to throw in some thoughts (I don't
wanna create a general debate here, but it may be good to look where we are
now):

1) What we want

    * a RAD environment 
       - supporting the Pascal language
       - supporting different development platforms (win / linux for now)
    * a component library
       - wrapping different widget-sets (gtk, qt, winapi ...)
       - compatible to the VCL where it can be achieved without too much
	 trouble

2) What's already there

   * a very good pascal compiler (Thanks to the fpc team, you're doing a
     great job)
   * the FCL with all those base classes (Thanks again to the fpc/fcl team)

   * basic code for the wrapper-interfaces around gtk, win & qt
   * a set of working (gtk-)components
   * basic code to integrate the compiler and an editor into one IDE

3) What's missing / tasks to be done

   a) code cleanups in a lot of files
   b) Most controls need much more testing, some still need much work.
   c) Nearly all controls need a review, so that we can update the status
      page. 
      (If somebody wants to spend some time on this: TEdit could be a good
       starting point for tests, cleanups and enhancements)
   d) all those database controls are missing completly
      (TDataset and some descendants are already in fcl but TDataSource and 
       the edit-controls are missing. Is there anybody willing to implement 
       parts of this?)
   e) a form designer
      (Cliff, are you still working on this? What about glade?)
   f) Much work has to be done on the IDE which will glue together all the
      stuff.
      
   From my point of view we should currently focus on a, b and c whereas d, e
   and f have lower priority, because they all depend on a working set of
   visual components. To be clear about this: I don't wanna keep anybody
   from working on the ide/designer but instead want to encourage anybody
   to get the basic components stable. This can also mean that you simply
   point out which controls have bugs or need more work. One of our major
   problems right now seems to be that nobody knows exactly what's working 
   and what's not.

bye,

  Stefan
-- 
---------------------------------------------------------
Stefan Hille          email: stoppok at osibisa.ms.sub.org 
48155 Muenster	      voice: 0251/664695