Subject Re: [IBO] Re: IBO Version 4_5
Author Helen Borrie
At 02:01 PM 16/12/2004 +0000, you wrote:


>--- In IBObjects@yahoogroups.com, "Jason Wharton" <jwharton@i...> wrote:
> > > What happened to versions 4_5_Aa thru 4_5_Af?
> >
> > These are beta revisions, they are not to be considered production
>quality
> > and therefore they are not persistently made available.
> >
> > Back in town finally from some meetings...
> >
> > Jason Wharton
> > www.ibobjects.com
>
>And 4.4 series?

Can I step in here and partly answer this...there were no 4.4 releases for
Delphi, though some 4.4 artifacts are probably in the VCS. The IBO-Kylix
conversion will need to step into the 4.4 area for its next round of betas,
if it possibly can. We're currently confronting the task of converting and
resynching the sources of the 005 ibo-k betas, which were converted from
the 4.2H Delphi sources, to and with the 4.3 and beyond sources.

At the binary level you don't notice that Jason made a major transition
between 4.2 and 4.3, to coalesce the code from previously distinct .INT,
.IMP and .pas units into larger, self-contained .pas units, leaving behind
only stubs for the old .INT and .IMP units. This presents a stumbling
block for the Kylix sources, since it makes it difficult to upgrade and
convert simultaneously via diff files, as we did previously.

We need to step into the 4.4 series for our transitional ibo-k beta
sources, since it's not a quick job to integrate them directly from 4.2H to
whatever happens to be the latest 4.5 release (a moving target). We'll
probably need to stay in 4.4 for quite a while, since we also have
conversions of many of the native IBO controls in a very alpha state,
alongside two strains of non-visual componentry supporting not just the
TIBO/TIB dichotomy, but also the X/non-X (or Console) dichotomy.

These comments are made on the ~uninformed assumption that Jason did mean
to let me earmark 4.4 for Kylix. It could well be that Jason already has
his .NET stuff occupying this space in the tree...and I will have to
rethink what to do about the transitional IBO-K stuff - and how we sync
.NET and Mono.

Whether it will *ever* be possible to keep the Delphi and Kylix versions in
sync is entirely moot. IBO for Delphi is built on the VCL, IBO-K on
CLX. The VCL presumes Windows and .NET, CLX is stripped down both ways to
provide classes that are implementable on either Linux or Windows and
(possibly) Mono. It might be that, in future, Jason will need to step up
Kylix sub-releases by even numbers and Delphi by odd numbers (my
perspective only).

Helen