Board index » delphi » Re: Borland Delphi + 64Bits

Re: Borland Delphi + 64Bits


2005-03-31 01:21:57 AM
delphi220
Bob --
Great stuff.
--
Nick Hodges -- TeamB
Lemanix Corporation -- www.lemanix.com
Read my Blog -- www.lemanix.com/nick
 
 

Re: Borland Delphi + 64Bits

Jim Cooper writes:
Quote
I'm not alone in thinking that, BTW.
No, you certainly aren't.
--
Nick Hodges -- TeamB
Lemanix Corporation -- www.lemanix.com
Read my Blog -- www.lemanix.com/nick
 

Re: Borland Delphi + 64Bits

Bob Dawson writes:
Quote
There is a large potential market for linux software as well
Heck, /Linux/ is a great example. it is a very popular OS, but the
market for development tools for it is relatively small. Same for the
Mac as well.
--
Nick Hodges -- TeamB
Lemanix Corporation -- www.lemanix.com
Read my Blog -- www.lemanix.com/nick
 

Re: Borland Delphi + 64Bits

Jim Cooper writes:
Quote
>EEEK! Add Win32 evolution in there somewhere! (They are doing it)

Also, C++ is supposed to be integrated at some point, isn't it?
Yes, that is been announced.
--
Craig Stuntz [TeamB] ?Vertex Systems Corp. ?Columbus, OH
Delphi/InterBase Weblog : blogs.teamb.com/craigstuntz
Please read and follow Borland's rules for the user of their
news server: info.borland.com/newsgroups/guide.html
 

Re: Borland Delphi + 64Bits

Bob Dawson writes:
Quote
1. Despite compelling advantages, Win95 as an upgrade sale for Win31
actually didn't go very well--new machines carried Win95, not
upgrade sales. More broadly, the rate at which MS can move the market
has been tested several times (Win95, Win2K, XP), and in each case it
required the sale of new machines to build eventual momentum.
If memory serves me right, I think I saw figures claiming MS sold 70
million Win 3.x licenses in the time span from the release of Win 95 and
through 12 months.
--
Ingvar Nilsen
www.ingvarius.com
 

Re: Borland Delphi + 64Bits

Captain Jake writes:
Quote
It is a myth that you have to be first to a market to make a good
profit in that market.
And it is still quite prevalent, too.
--
John Kaster blogs.borland.com/johnk
Features and bugs: qc.borland.com
Get source: cc.borland.com
If it is not here, it is not happening: ec.borland.com
 

Re: Borland Delphi + 64Bits

"Michael Anonymous" <XXXX@XXXXX.COM>writes
Quote
John Jacobson writes:
The point remains that users could have used 32-bit software on
>their 32-bit PC's long before most of them actually did. Pretending that
>we must all arbitrarily limit our debate to Windows 95 and up is a silly
>tactic that quite frankly insults the intelligence of everyone reading
>this thread.

You are free to discuss whatever you want as long as you don't use your
arguments to
disprove my assertion that
"People wanted 32bit software when a 32bit Windows came out.
If this trend continues,
they behave the same way for 64bit Windows."
Actually I am free to use my arguments to dispute that assertion as well, as
indeed I have. NT was a "32-bit Windows", and yet there was not a lot of
demand for 32-bit software when it was first available. So even if we
arbitrarily limit the discussion just to Windows, you are still wrong.
 

Re: Borland Delphi + 64Bits

Arthur Hoornweg writes:
Quote
Rudy Velthuis [TeamB] schreef:

>Soon? Will in-house development also be affected? Who says so, BTW?

I have no reason to expect that in-house development will be
affected real soon. But I believe that application development
certainly will.

Rudy, you keep assuming that the market for 64-bit
applications is based on common sense. I believe that it
is not. But it is a market nonetheless.
Yes, but is it a market for most of the Delphi developers anytime soon?
--
Rudy Velthuis [TeamB] rvelthuis.bei.t-online.de
"Humor is a {*word*221} sword - it allows you to make a point without drawing
{*word*76}." -- Mary Hirsch
 

Re: Borland Delphi + 64Bits

John Jacobson writes:
Quote
NT was a "32-bit Windows", and yet there was
not a lot of demand for 32-bit software when it was first available.
Even in *business* which is the point being sidesteped by trying to limit
the discussion to *consumer* versions of Windows, et al.
--
Wayne Niddery - Logic Fundamentals, Inc. (www.logicfundamentals.com)
RADBooks: www.logicfundamentals.com/RADBooks.html
"True peace is not the absence of tension, but the presence of
justice." - Martin Luther King, Jr.
 

Re: Borland Delphi + 64Bits

Arthur Hoornweg writes:
Quote
This situation is weird. Here I sit, actually begging Borland
to please take my good money for a product. And they don't even
listen.
You are not telling Borland. You are telling us. And that seems to be
the problem. If enough people would actually tell Borland, if possible
in an organized way, they might even listen.
--
Rudy Velthuis [TeamB] rvelthuis.bei.t-online.de
"A printer consists of three main parts: the case, the jammed paper
tray and the blinking red light" -- unknown
 

Re: Borland Delphi + 64Bits

"Henrick Hellström [StreamSec]" writes:
Quote

Well, I actually have to represent values up to at least 2^16384.
Since I doubt you will add an Int16384 type, I will have to represent
these values as arrays of some base type. Furthermore, operations on
such numbers will be painfully slow, so it certainly does matter
which base type I choose for the arrays.
Byte, Word, DWord, and Int64 are fixed size across platforms.
-Danny
--
Delphi Compiler Core: blogs.borland.com/dcc
 

Re: Borland Delphi + 64Bits

Rudy Velthuis [TeamB] writes:
Quote

Actually, THandle is defined as LongWord in System.pas. I know it is a
pointer in C, and it should have been a pointer in Delphi altogether,
IMO. So simply make it a pointer in Win64, and you'd get nice error
messages when assigning to an integer, or when comparing with 0.

Why should THandle be a pointer type? You can not dereference it.
THandle is a pointer-sized integer type. There's a big difference
between "pointer-sized" and "pointer type".
My point earlier is that because THandle is assignment compatible with
Integer in the current 32 bit code, there is a lot of code out there in
the wild that makes unsafe assumptions about the equivalence of these
types. Assigning a THandle property to an integer local variable is
currently valid, but it will be invalid in a Win64 environment.
-Danny
--
Delphi Compiler Core: blogs.borland.com/dcc
 

Re: Borland Delphi + 64Bits

Eric Grange writes:
Quote
>AFAIK, some of Danny's research included investigation of codegen
>for (IIRC) Intel's offering(s).

Dave, are you not mixing up what was done for the Itanium
with what isn't done for x86-64?
I'm not mixing anything up. His research was research. I expect it also
wasn't restricted to that alone.
--
Dave Nottage [TeamB]
 

Re: Borland Delphi + 64Bits

"Danny Thorpe" <XXXX@XXXXX.COM>writes
Quote
Why should THandle be a pointer type? You can not dereference it.

THandle is a pointer-sized integer type. There's a big difference
between "pointer-sized" and "pointer type".
This suggests that "THandle" is a bad name for it, since it implies a class
and therefore a pointer. I guess "TDateTime" falls into this same category
as well. It is not a pointer to a class, though the name suggests it.
 

Re: Borland Delphi + 64Bits

Danny Thorpe writes:
Quote
Why should THandle be a pointer type? You can not dereference it.
I know it should be treated as an opaque type, but it IS originally a
pointer type in C.
In winnt.h:
typedef PVOID HANDLE;
#define DECLARE_HANDLE(name) typedef HANDLE name
In windef.h
DECLARE_HANDLE(HWND);
Quote
THandle is a pointer-sized integer type. There's a big difference
between "pointer-sized" and "pointer type".
The original handles, like HWND and HDC are declared as pointers, in C.
The fact that they are not declared as pointers in Delphi can cause
some hard to solve problems when interfacing with BCB. I will have to
find the example, but it was discussed in the BCB groups.
Quote
My point earlier is that because THandle is assignment compatible with
Integer in the current 32 bit code, there is a lot of code out there
in the wild that makes unsafe assumptions about the equivalence of
these types.
Oh, indeed. That is why it might actually be useful to make it a
pointer type. you will never see that problem happen.
--
Rudy Velthuis [TeamB] rvelthuis.bei.t-online.de
"I have made this letter longer than usual because I lack
the time to make it shorter." -- Blaise Pascal