[DUG] Unicode [redux]

[Jolyon Smith]

> The 64bit version will be interesting...

aiui on x86-64, "Integer" remains 32-bit "in the metal" so I think 64-bit
shouldn't have that much impact.  The only immediately obvious issue I see
with 64-bit will be where people are using "Integer" types to store pointer
values.  Int64 becomes hardware supported rather than requiring compiler
magic (which is - or was - sometimes broken!).

But, now that pointer arithmetic is directly supported, using Integers to
wrap pointers shouldn't be necessary any more as this is often where/why
such things are done.

I hope Embarcadero(*) don't capitulate to the backward-compatability crowd
by migrating Integer to 64-bits just to ease the transition of such code
(except in so far as adjusting things like the TComponent.Tag property, but
that should be handled by making that property Int64, not making Integer
itself 64-bit).



(*) see next email....



-----Original Message-----
From: delphi-bounces at delphi.org.nz [mailto:delphi-bounces at delphi.org.nz] On
Behalf Of Jeremy North
Sent: Friday, 16 October 2009 12:09 p.m.
To: NZ Borland Developers Group - Delphi List
Subject: Re: [DUG] Unicode [redux]

Under what circumstances would you want to use the $H- option now? For
recompiling old code? Perhaps it should now be deprecated as well.
Time to update that really old code or just leave it in an older
version. I personally have a number of projects that I won't upgrade
for various reasons.

I find it frustrating how sometimes Borland/CodeGear/Embarcadero seem
to always favour backward compatiblity at the expense of doing things
better or right. However there is really only one group to blame for
that. Users. People complain when they have to change a line of code
when using a new version. They expect all these new things to be
added, and expect not to have to change anything. This is what leads
to compromises. The 64bit version will be interesting...

When unicode was introduced, it was going to cause some upgrading
issues, so they (Embarcadero) should have broken it properly, instead
of leaving a lot of legacy smells around. It is a balancing act and
you'll never please everyone.

I'd certainly settle for Embarcadero giving Delphi 2007 to upgraders
of Delphi 2010. Anything that gets people off the pre-Galileo IDEs is
a blessing. Although it is pretty easy to give product away when it
isn't yours to give in the first place though.

A "clean-up" release would be nice but might be a while off and folks
don't generally like those because they don't bring new features. A
bit like the Delphi 2007 release. I hated that because of the
non-breaking nature of the release, but many other developers thought
it was really good because upgrading was really easy (from 2006).

Anyway, I've gone off on a little tangent there...
_______________________________________________
NZ Borland Developers Group - Delphi mailing list
Post: delphi at delphi.org.nz
Admin: http://delphi.org.nz/mailman/listinfo/delphi
Unsubscribe: send an email to delphi-request at delphi.org.nz with Subject:
unsubscribe