Delphi 2005 Update
Michael Swindell (Borland)
improve memory usage among some other issues that have been raised post
ship. Thank you for bringing some of these to our attention, your early
feedback has been very helpful. We expect to have this update available
after Thanksgiving, sometime in early December. Although there are some
common issues raised here in these forums, the majority of early customer
feedback has been that Delphi 2005 is a high quality Delphi release with an
amazing feature set. With more features than we ever put in a single
release, there will inevitably be some bugs that come up that we need to fix
and we're focused on keeping Delphi 2005 quality level up to it's feature
level - the highest ever :o)
Thanks again for your feedback and happy holidays,
Michael
Win the "Ultimate Delphi Email Address":
http://homepages.borland.com/mswindell/
Alexander Tereshchenko
------------------------------------------
Alexander Tereshchenko
al...@knives-out.org
http://www.knives-out.org/
"Michael Swindell (Borland)" <no-...@borland.com> wrote in message
news:41a4...@newsgroups.borland.com...
Ilkka Hyvärinen
> Just wanted to let you know that we do have an update underway
Thank you for the news, Michael. Much appreciated.
--
ilkka hyvärinen
chrisC
"Michael Swindell (Borland)" <no-...@borland.com> wrote in message
news:41a4...@newsgroups.borland.com...
Way to go! Thanks Michael for the new Borland Spirit! I hope the BCB guys
will hear good news too by Dec. 15?
Chris
Ingvar Nilsen
> Just wanted to let you know that we do have an update underway
If this isn't great news, I am Abraham Lincoln.
Am I Abraham Lincoln? No.
IOW: Great News!!
--
Ingvar Nilsen
Borut Sfiligoj
newsreader I noticed that you have already replied to it. However, I hope it
can still be of some value, so I'm sending it anyway.
And yes, D2005 is (appart from some beginning glitches) superb. I hope (and
believe) this direction you (Borland) have taken in the last couple of
months continues.
Just one question, though: is Borland going to implement named parameters in
BDP? The current solution with using ? for parameter holders breaks the
current (SQL) code (BDE, ADO, IBX) and is very cumbersome to use.
-----------------------------------------------
I am using Delphi 2005 Architect English (purchased, not trial). I run it on
my 3 machines: Pentium 4 (2 GHz - 512 MB RAM, 2.2 GHz - 512 MB and 3 GHz - 2
GB). All three have Windows XP SP2 installed.
So far my experience with using D2005 is pretty positive, regarding that
this is the first time Borland has put 3 products in one IDE. There is
however one major problem, which other people have already been talking
about - memory leak in IDE.
I have been trying to pinpoint the problem but have been exactly successful.
However I can now reproduce it on all 3 machines, so I believe it can be
reproduced anywhere.
The easyest way to see the problem is when using large .pas files (the
problem also exist with .cs files):
1. Open Windows Task Manager, jump to page Processes add a column VM Size
(View / Select Columns: Virtual Memory Size).
2. Enable all Code Insight features (6), disable Together Support in Tools /
Otions Dialog. Use Default layout.
3. Create a new project in D2005 (File / New / VCL Forms Application -
Delphi for Win32).
4. Add a file C:\Program Files\Borland\Delphi7\Source\Vcl\ComCtrls.pas
(about 665 KB long) to the project (there must be at least 2 files in a
project.
5. Jump to the end of the ComCtrl.pas and add one line (press Enter).
6. Press Space and wait a couple of seconds.
7. Repeat step 6.
The VM sizes of bds.exe on my machine are:
- Start Delphi with a blank screen (no project loaded):
89 MB
- Create New Win32 Application
100 MB
- Add (and open in editor) ComCtrls.pas to the new project 121 MB
- Add a new line to ComCtrls.pas
139 MB
- Add space to the end of the file
148 MB
- Add another space to the end of the file (and wait 5 sec) 158
MB
- Add another space to the end of the file
131 MB
- Add another space to the end of the file
132 MB
- Add another space to the end of the file
142 MB
Each keystroke adds 1 - 10 MB to bds.exe's VM Size. However IDE seems to do
some kind of garbage collection here and there and the VM size usualy also
decreases after a while.
Now do exactly the same as in the previous example, except steps 3. and 4.:
3. Create a new package in D2005 (File / New / Package - Delphi for Win32).
4. Add AT LEAST 2 .pas files to this package: C:\Program
Files\Borland\Delphi7\Source\Vcl\ComCtrls.pas and C:\Program
Files\Borland\Delphi7\Source\Vcl\Controls.pas
Now after each keystroke (you must wait a couple of seconds after each
keystroke) in ComCtrls.pas the VM Size increases about 10 - 15 MB! and never
decreases. Also CPU utilization jumps to over 50 % for bds.exe for a couple
of seconds.
After reaching a certain limit of used VM (about 500 MB on my 512 MB RAM
machine) , which takes about 100 or less keystrokes, the IDE starts
responding slowly and finally virtually freezes which makes further work
impossible.
If you rename ComCtrls.pas to say ComCtrl.txt or ComCtrl.sql and try editing
the file the VM size doesn't change at all. So the problem probably lies in
background compilation (Code Insights).
I tried to exclude some Code Insight features and noticed that after
excluding Error Insight the situation changes drammatically. The IDE still
consumes about 50 - 100 KB after each keystroke (even if you exclude all
Code Insights), which means there are more memory leaks but nevertheless the
situation is much better.
Something very similar happens if you have Together support for the current
project enabled. The memory increases in 5 - 10 MB chunks for each
keystroke. And never decreases.
So in the current version I cannot use Error Insight (which is unfortunate)
and Together support (which I don't use anyway), because most of my projects
include large files and after half an hour or so the memory consuption of
bds.exe even on my 2 GB machine is so large, I have to shut down the IDE.
I tried the same thing in Delphi 7 and there the memory never increases when
editing a file (with D7's Code Insights enabled).
I managed to transport all of my components to D2005 (the only thing that
gave me real problems was the changed behaviour of TForm showing - in D2005
the TForm recreates its window whenever you call ShowModal - which is
described in readme, but it also doesn't use Left and Top positionings if
the form has no owner and TForm.Position is poDefault or poDefaultPosOnly
and always displays it in the left-top corner of the screen (coordinates 0,
0)).
I have (since friday, when I got my D2005) been working almost exclusively
with Win32 part of it, so I cannot comment other personalities (except that
I have tried reproducing this memory leak in D.Net and CS, where it also
exists).
I hope Borland can solve this problem quickly, because I would really like
to propose a transition from D7 to D2005 in our company which currently I
cannot. I can't say I haven't expected any problems with the new version of
Delphi (especially since a lot of new stuff was added to it), however I
certainly expect frequent and fast SPs for D2005.
Borut.
Holger Flick
would be true ;-) *g*
not ( not T ) => T
T => T
If you assign True to "great". Okay, enough of this.
But it's good to hear that an update will be there soon.
Regards,
Holger
Kristofer Skaug
> Am I Abraham Lincoln? No.
relieved to hear this, I must admit to being caught in a moment of
doubt... <g>
--
Kristofer
Liz
<no-...@borland.com> wrote:
>Just wanted to let you know that we do have an update underway that will
>improve memory usage among some other issues that have been raised post
>ship.
Will there be any chance that this will be a freely downloadable one?
so people with the trial can play with the trial a little more
efficiently too?
Kristofer Skaug
> Just wanted to let you know that we do have an update underway that
> will improve memory usage among some other issues
- The refactoring failures?
- The exception on shutdown?
- The WinXP SP2 "Open/Save file" popup?
--
Kristofer
Ingvar Nilsen
> Ingvar Nilsen wrote:
>
>>Am I Abraham Lincoln? No.
>
>
> relieved to hear this, I must admit to being caught in a moment of
> doubt... <g>
>
LOL.. actually, I'm in doubt sometimes myself <g>
--
Ingvar Nilsen
Daniel Wischnewski
>
Allen Bauer mentioned in his blog that this one will be in the next update
as he had the solution obviously a few days after rtm...
http://blogs.borland.com/abauer/archive/2004/11/04.aspx
--
Regards,
Daniel
Co-Admin: www.delphipraxis.net
* * The German Delphi Forum * *
Carl
Derekl Davidson
> improve memory usage among some other issues that have been raised post
> ship. Thank you for bringing some of these to our attention, your early
> feedback has been very helpful.
Ah good. It's excellent to know that a fix is underway.
> the majority of early customer
> feedback has been that Delphi 2005 is a high quality Delphi release
Oh, for crying out LOUD! Are you Borland guys in total denial? This
must be the worst Delphi release since D4/D8. That you're moving fast
to fix it is laudatory. To claim that it's a high quality release
flies in the face of the facts.
> ... there will inevitably be some bugs that come up that we need to fix
Undoubtedly. But, I'd feel far happier knowing that you were doing
something positive to stop this type of show stopping bug occuring
again. For example: A public beta would probably have caught this. How
about public betas?
--
Derek Davidson
El Lobo
Thank you for the great work. Now, how is this diferent from a public Beta?
:-P
I mean, the same feedback you are getting now, you would get with a public
beta.
Regards
"Michael Swindell (Borland)" <no-...@borland.com> wrote in message
news:41a4...@newsgroups.borland.com...
Larry Drews
news:41a47b05$1...@newsgroups.borland.com:
> Hello
>
> Thank you for the great work. Now, how is this diferent from a public
> Beta?
>:-P
> I mean, the same feedback you are getting now, you would get with a
> public beta.
> Regards
>
How is any release that is subsequently followed by an update different
from a public beta? Since almost all software that I know about has
subsequent releases, it must all be public beta, right?
El Lobo
"Larry Drews" <ldrews...@tampabay.rr.com> wrote in message ? Since
> subsequent releases, it must all be public beta, right?
You must be kidding, right?
Yannis
> How is any release that is subsequently followed by an update different
> from a public beta? Since almost all software that I know about has
> subsequent releases, it must all be public beta, right?
I do not see a grin anywhere on your post are you trying to be funny?
regards
Yannis
Liz
<ldrews...@tampabay.rr.com> wrote:
>Release public beta => followed by release of updated version known as
>formal release
>
>Release of formal release => followed by release of updated version known
>as update or service pack or ?
>
>So only the names used are different. So what is the functional
>difference?
Some people who may not be able to purchase the formal release can do
the public beta?
Larry Drews
@newsgroups.borland.com:
No, actually I'm serious. Doesn't all released software that you know
about have some bugs? Doesn't the vendor usually send out updates? Isn't
this the same pattern that occurs with "public betas"?
Kristofer Skaug
> So only the names used are different. So what is the functional
> difference?
A few thousand $$$ for the beta tester, plus saving face for Borland.
--
Kristofer
Paul Dolen
>after Thanksgiving, sometime in early December.
Great. Look forward to it.
>Although there are some
>common issues raised here in these forums, the majority of early customer
>feedback has been that Delphi 2005 is a high quality Delphi release with an
>amazing feature set.
Don't get me wrong, I like D2005, and I'm sure that I will all the
more so after you fix some of these bugs. But, I find it highly
unlikely that anybody starting to use D2005 for any serious work
hasn't run into these problems. It is not yet a "high quality Delphi
release".
Paul Dolen
>Am I Abraham Lincoln? No.
>IOW: Great News!!
Just reminded me of a silly story. As you know, a lot of call centers
and tech support is being done in India. And a lot of these people
adopt American names to sound less foreign. Some guest on Jay Leno
said he got a tech support person in India that picked an American
sounding name, Abraham Lincoln.
David Farrell-Garcia
> Some people who may not be able to purchase the formal release can do
> the public beta?
Doesn't sound like a target customer to me :-)
--
David Farrell-Garcia
Whidbey Island Software LLC
Posted with XanaNews 1.16.5.1
Allen Bauer
>> will improve memory usage among some other issues
>
> - The refactoring failures?
Yes. Are they in QC? If so, please highlight the requisite numbers here so
that a QC sysop can push it(them) to our internal database.
> - The exception on shutdown?
Yes.
> - The WinXP SP2 "Open/Save file" popup?
Definately, Yes.
--
Allen Bauer
Delphi/C#Builder Principal Architect
Borland Software Corporation.
http://blogs.borland.com/abauer
Rick Carter
>so people with the trial can play with the trial a little more
>efficiently too?
I don't think you'll be able to apply any updates to the trial version.
Unless Borland decides to roll out a revised version of the trial, I
think WYSIWYG.
Rick Carter
cart...@despammed.com
Chair, Delphi/Paradox SIG, Cincinnati PC Users Group
--- posted by geoForum on http://delphi.newswhat.com
Ingvar Nilsen
> And a lot of these people adopt American names to sound less foreign.
>
melodic names, although not necessarily easy to remember..
One of my best friends' first name: Thillaiampalam
> that picked an American sounding name, Abraham Lincoln.
Maybe Abraham is easier to remember, preferably combined with another
last name than Lincoln though :-)
--
Ingvar Nilsen
Bob Swart
> Just wanted to let you know that we do have an update underway that will
> improve memory usage among some other issues that have been raised post
> ship.
Thanks for letting us know!
> feedback has been that Delphi 2005 is a high quality Delphi release with an
> amazing feature set.
You can say that again! Delphi 2005 Rocks!
Groetjes,
Bob Swart (aka Dr.Bob - www.DrBob42.com)
--
Bob Swart Training & Consultancy (eBob42) - Borland Technology Partner
Delphi Win32 & .NET training and support - IntraWeb Authorized Trainer
Liz
<dav...@orcasoftware.com> wrote:
>Doesn't sound like a target customer to me :-)
I meant that not everyone can afford to buy a new release the moment
its out. But do intend to do so, but can afford to use time for a
beta.
Liz
wrote:
>You can say that again! Delphi 2005 Rocks!
It sure does.
Yes its frustrating when it barfs (as I found with my win32
refactoring, it shows the code in the window a treat, but fails to
actually make the change as it barfs with long long barfs) however,
the product itself, combining the 3 products way better than VS ever
did (I gave up around 98 on visual studio not that I ever used it
through choice)
Once I can afford Im very much hoping to upgrade my home use version
of delphi to 2005, I just cant afford it. Freeware apps do that, no
money coming in, means no pot to pull on when you want to upgrade.
Im going to mourn the day my trial runs out even though I cant use it
to develop my app as I have found work arounds to many of my issues,
only left with one now, and no responses anywhere on it which is a
shame..
(OT techy bit: For those who missed it, indy 9 telnet component, if
you programattically close it the thread stays open, if the MDI form
closes the thread and socket its ok, however, if you click on the x
for the mdi form so bypass any capturings in that sense, even
terminating the thread and closing the socket is still leaving me with
the thread open - eg its not stopping it at request and the app hangs)
Kristofer Skaug
>>
>> - The refactoring failures?
>
> Yes. Are they in QC? If so, please highlight the requisite numbers
> here so that a QC sysop can push it(them) to our internal database.
QC9586.
--
Kristofer
DS
> Just wanted to let you know that we do have an update underway that will
> Thanks again for your feedback and happy holidays,
> Michael
>
Michael,
Thank you for the good news. I'm still playing with trial and I decided
to go after Arch version (but I'm waiting until next week when there's
discount at local presentation ;-). Great product it is!
Anyways, just to mention some quirks I encountered:
1. Importing C# projects from VS.NET:
Simple stuff imports OK. But I have solution, which spawns across
several subdirectiories which are separate sub-namespaces, e.g.
folder namespace
SMTH App
SMTH\Tools App.Tools
SMTH\Extras App.Extras
Now, .CSPROJ gets converted to .BDSPROJ but it doesn't compile. It says
that resources are missing. When I create new .CS file I see its .resx
in resource tree as subitem, whilst while importing it's on the same
level, so probably there's the problem. I can send you reproducible case
if you want.
2. Character issues:
I have a form or two (again, VC#) which contain some extended characters
and I saved them from VS.NET with selecting UTF encoding in "Save As"
dialog. They import succesfully in D2005 but non-ASCII characters are
replaced with garbage. I can send you reproducibla case.
3. Importing VS.NET 2005 (beta) .CSPROJ:
D2005 doesn't complain (although project format is changed from VS2003)
but upon compile it dies with "Internal error".
4. This one is probably due my misunderstading:
I create UML ECO model. When I reopen the project and try to view UML
via "Model view" >> "CoreClassesPackage" it opens quicky. If instead I
first doubleclick "Project manager" >> "Model support" >>
"CoreClassesPackage" >> "CoreClassesPackage.txvpck" CPU utilization goes
100% for quite some minutes and ObjectInspector flashes like crazy.
5. Copy/Paste
Sometimes (quite often) keyboard shortcuts Ctrl-C/Ctrl-V stop to work
and I have to select Copy/Paste via right-click menu.
Again, let me state that I don't want to sound negative. I'm *very*
enthusiastic about the product. I've been with Borland since TP3 and
only in last year I switched to M$. It looks like it's time to come home
again.
I'm quite sure you'll polish the small anoyances. Thank you for great
product.
LP,
Dejan
Lord Crc
<aba...@spicedham.borland.com> wrote:
>If so, please highlight the requisite numbers here so
>that a QC sysop can push it(them) to our internal database.
Just entered 9610.
Btw is the "Extract Method" supposed to work on regular
procedures/functions, or only methods?
- Asbjørn
Will DeWitt Jr.
> Btw is the "Extract Method" supposed to work on regular
> procedures/functions, or only methods?
I would hope the former-- and I've noticed the menu item isn't disabled
for regular procedures/functions (though that may be a bug in itself if
it's only meant for methods).
Will
--
Want a 64-bit Delphi compiler for AMD64 / IA-32e? Vote here--
Holger Flick
it's called "Extract Method". So it'll work on... ;-)
But I understand the point you make. However, functions and procedures
w/o an assigned to them (what distinguishes them from methods...) are
not to be considered in OOP (you cannot even do it in Java, e.g.). So
for this reason I do not think it's possible.
--
Holger
Lord Crc
wrote:
>But I understand the point you make. However, functions and procedures
>w/o an assigned to them (what distinguishes them from methods...) are
>not to be considered in OOP (you cannot even do it in Java, e.g.). So
>for this reason I do not think it's possible.
Oh it's perfectly possible, it just doesnt generate correct code at
the moment, which is why i asked. There's a bug, i'm just trying to
find out if its this or that ;)
- Asbjørn
Rick Beerendonk
my colleagues and I will never ever going to advise to use Delphi again. Our
asses are on the line here. Thx!!!
Feel left alone.
Rick
David Smith
>
> Thanks again for your feedback and happy holidays,
> Michael
Very nice to hear from you. I was little worried about the memory
reports, because the memory problems can be somewhat hard to find. Other
than that (I'm willing to accept the quite long startup time, because no
other tool supports three environments), this release really showed how
good things can be.
When I was working today, I would have killed to have the refactoring
features in D7, because I had a lot of code modifications. If you can
patch the most visible problems in D2005, then you'll have the best .NET
IDE in your hands. VS2003 has no match.
--
David S.
Delphi programming : http://www.borland.com/delphi/
David Smith
Of course they will fix the D8 issue first. The fix is already in D2005,
so it's just back porting. And they have said that the update is now on
it's way.
Kristofer Skaug
>
> But I understand the point you make. However, functions and procedures
> w/o an assigned to them (what distinguishes them from methods...) are
> not to be considered in OOP (you cannot even do it in Java, e.g.). So
> for this reason I do not think it's possible.
Two points regarding your comment:
1) There's nothing inherently "OOP" about refactoring.
OK, so "method" is an OOP term, and some refactorings (e.g. extract
superclass) are intrinsically OOP.
But refactoring in general is about keeping your code in good shape,
avoiding duplication, finding good semantics, etc - equally necessary in
code that happens to live outside of a class. I don't see any reason at
all, therefore, why "extract method" shouldn't be completely applicable
to global functions and procedures.
and now for a slight digression:
2) "functions... are not to be considered in OOP". That's a very
orthodox position, which I completely disagree with. Functions and
procedures that do not share global state or data are not in violation
of fundamental OOP goals; they've got every bit as much reusability
potential, information hiding, etc. as the same code written as a method
of a class. The whole Delphi RTL/VCL is built on a sound mix of globally
accessible utility functions and a powerful class library. I wouldn't
want it any other way - the Java way for example, where you need to
create and access a class just to do some simple date/time arithmetic...
IMHO, it's pointless.
--
Kristofer
David Clegg
> Maybe Abraham is easier to remember, preferably combined with another
> last name than Lincoln though :-)
Simpson?
--
Cheers,
David Clegg
dcl...@gmail.com
Vote 1 http://cc.borland.com/codecentral/ccweb.exe/listing?id=21489 :-)
Now supports Google Groups searching with Dyna-extend(tm) technology!
QualityCentral. The best way to bug Borland about bugs.
http://qc.borland.com
"Ahhh... sweet pity. Where would my love life be without it?" - Homer
Simpson
JED
> I wouldn't
> want it any other way - the Java way for example, where you need to
> create and access a class just to do some simple date/time
> arithmetic... IMHO, it's pointless.
While I agree with what you are saying, you could use class methods to
do this so you don't need to create a class instance.
--
Alpha 3.1 - Released October 13, 2004
JED, QC - Win32 Client for QualityCentral:
http://www.alphalink.com.au/~jed/QC/
JED, QC Blog: http://jedqc.blogspot.com/
Marcelo Carvalho
Since I saw the new features and knew that they would come also for
Win32 in the Borcon presentation, I saw that this would be a
"must-have" version! I'm already working with the Trial version, just
waiting to my already ordered box arrive.
I skipped D8, and for a while thought that Borland abandoned us. Now
I think that the spirit of partnership is back! Despite some few
enhances in VCL, most are in IDE (all very good), this 3 in 1 product
can offer the soft transition to .Net world that D8 couldn't do.
All of us understand that bugs are part of this job, and having
Borland team near, listening us about these found bugs, taking care of
solving them fast, and also listening our new features claims, is all
we want!
Of course all of us work for money, and Borland is a commercial
company that must have profits. But the spirit of community must be
preserved as well. Of course customers must respond buying the products
to turn it sustainable.
I hope it continues this way. Please, Borland, don't shy if you need
to release a patch per month, if they become necessary to turn the
product really as solid as a rock, and thousands of customers really
happy. We will not complain about having many patches, but we will
complain if keeping having bugs...
What most of us disliked was when, in some recent past, you Borland
used to throw the product "as is", and with very rare and delayed
patches, and we should live with a sort of problems - leading us to
question if we have made the right choice. Partnership means recognize
some mistakes, as told before inherent to our job, and concentrate
efforts about solving them, just as we are seeing now!
Thanks,
Marcelo.
Michael Swindell (Borland) wrote:
> Just wanted to let you know that we do have an update underway that
> will improve memory usage among some other issues that have been
> raised post ship. Thank you for bringing some of these to our
> attention, your early feedback has been very helpful. We expect to
> have this update available after Thanksgiving, sometime in early
> December. Although there are some common issues raised here in these
> forums, the majority of early customer feedback has been that Delphi
> 2005 is a high quality Delphi release with an amazing feature set.
> With more features than we ever put in a single release, there will
> inevitably be some bugs that come up that we need to fix and we're
> focused on keeping Delphi 2005 quality level up to it's feature level
> - the highest ever :o)
>
> Thanks again for your feedback and happy holidays,
> Michael
>
> Win the "Ultimate Delphi Email Address":
> http://homepages.borland.com/mswindell/
Kristofer Skaug
>
> While I agree with what you are saying, you could use class methods to
> do this so you don't need to create a class instance.
Of course you can, but what is then the point of having a class?
Just a way to bundle otherwise unrelated functions together?
Or just a check-in-the-box formality, so one can claim to be "100%
Object-Oriented"?
--
Kristofer
Thomas Edison
> You can say that again! Delphi 2005 Rocks!
It's funny how the "Borland Technology Partners" are always so
enthusiastic. <g>
Hrvoje Brozovic
>However, functions and procedures w/o an assigned
>to them (what distinguishes them from methods...)
are not to be considered in OOP
This is complete nonsense. Objects are invented to
group data and it's own manipulation together.
There is absolutely no sense in object that groups
manipulation with nothing else together.
It is just add nothing to manipulation of nothing.
So added value is ... nothing.
But there lot of value lost, in simplicity,
implementation efficiency, and conceptual clarity.
Objects provide context for code.
If there is no context, there is no need for objects.
Square root of number 4 is number 2.
Factorial of number 3 is number 6.
Today is friday, and tomorow will not be.
Putting those methods in some class will
only create false illusion that those base
trues are somehow dependant of anything else
than our known universe. And they are not.
Bob Dawson
> Objects provide context for code.
State (object data) is not the sole definition of context, there are also
the matters of domain and scope. Classes are not simply state containers,
they are also classificatory.
> If there is no context, there is no need for objects.
I disagree.
> Square root of number 4 is number 2.
> Factorial of number 3 is number 6.
> Today is friday, and tomorow will not be.
The first two of these are conceptually/domain related in a way that neither
is from the last, so I'd find the groupings
class function TMathUtils.SquareRootOf(const Value: double) : double;
class function TMathUtils.FactorialOf(const Value: integer) : int64;
class function TDateUtils.DayOfWeek(const FormatStr : string; const aDate :
TDateTime) : string;
quite natural, not indicative at all of any
> [...] false illusion that those base
> trues are somehow dependant of anything else
> than our known universe.
Grouping global functions into meaningful classes is really no different
than what Delphi has always done at the unit level (e.g., providing separate
Math and DateUtil units). Delphi for .NET automates this process for us
(creating a default unit class for any 'global' methods), or allows us to
allows us to define our own more explicitly.
bobD
Hrvoje Brozovic
> "Hrvoje Brozovic" wrote
>
> > Objects provide context for code.
>
> State (object data) is not the sole definition of context, there are also
> the matters of domain and scope.
What domain and scope. Code
x, y: integer;
y := 2;
x := y + 3;
Will always produce number 5 in the variable x,
no matter how class in which code executes is called.
If you allowed to do that, (any you are) why on earth
you have to impose anything for other simple calculations.
>
> Grouping global functions into meaningful classes is really no different
> than what Delphi has always done at the unit level (e.g., providing separate
> Math and DateUtil units). Delphi for .NET automates this process for us
> (creating a default unit class for any 'global' methods), or allows us to
> allows us to define our own more explicitly.
On the first page of any OOP book you will find
sentence like "object's are created to mimic nature"
Nature of our universe is that for any number
you can calculate one and only one sine and
cosine value, and all you need for calculation
is number itself. There is no contex, no problem
domain, nothing.
Only if you don't need that feature, you don't have to
use sin() and cos() functions, so you can help linker
not to bother with them at all with not including their
unit in uses clause.
Promoting unit with global functions to class with
methods is hack to enable 2005 years old style
calcuations in super modern code.
There is nothing what can be said to
defend need for embedding sin() and cos()
in a TTrigonometry class.
It is pure nonsense and violence over common sense.
If you intend to create application that calculate
mathematical functions in 7-dimensional
space, you are welcome to produce sutch a class,
or simple create your own unit Math7, and use it
with dcc32 -aMath=Math7 compiler switch.
Marcelo Carvalho
About the expected first patch, is IDE speed a bug to be included in
the patch?
For me, the the IDE is working well (TRIAL version), but VERY slow
if compared with D7 even with single form and small projects.
Is is a bug to be fixed, or some thing related to .Net technology
used on it? (every new thing from M$ is always heavier and slower)
Thanks and regards,
Virgil Dupras
improve speed is the memory leaks are making you run on the swap.
--
Virgil Dupras
Hardcoded Software
http://www.hardcoded.net
Marcelo Carvalho
Bob Dawson
> Will always produce number 5 in the variable x,
> no matter how class in which code executes is called.
The class container is for the convenience of the programmer, not the
compiler. While you're quite right that Pi, Avogadro's Number and Planck's
constant are 'universe specific,' people still think of them and categorize
them as facts of geometry, chemistry, and physics. It's really that simple.
Facts tend to group in ways other than the scope of their validity.
> space, you are welcome to produce sutch a class,
> or simple create your own unit Math7, and use it
As I pointed out, these end up being roughly equivilent actions in DFDN.
bobD
Holger Flick
I truly accept your criticism, however, I am a little bit mad at your
choice of words. To call it 'complete non-sense' is a little bit harsh.
Kristofer made a good point, I totally support. I agree that it is
stupid having to use static class methods in order to get the current
time like it is in .NET and Java.
However, to call my statement non-sense is not right. OOP clearly does
NOT include any functions or procedures, which are means of structural
programming. In order to not get a 'flamatory' remark: I also agree
that Delphi found a nice way between "correct OOP" and "structural
means".
Regards,
Holger
Hrvoje Brozovic
>
> I truly accept your criticism, however, I am a little bit mad at your
> choice of words. To call it 'complete non-sense' is a little bit harsh.
I am aware of that. It justs happened that your remark triggered
my irritation with altitude that Delphi is in minus compared with Java
and C# only because it has capabilities those two lack.
Anyway, sorry, I apologise.
Holger Flick
if I'd prefer Java's strict OOP, I'd use Java. Furthermore, what makes
me hesitate to use .NET is just that. There a 1000 possibilities to get
the current date and time using different static methods depending if
you use a special encoding, localization etc. "Pure Delphi" is so much
easier and prefered by me.
Thanx for you apology...
--
Holger
Holger Flick
even though our names might mean the same, it was not intended and a
mistake. Sorry for that :-)
Holger Flick wrote:
> Hi Holger,
Kristofer Skaug
>
> Putting those methods in some class will
> only create false illusion that those base
> trues are somehow dependant of anything else
> than our known universe. And they are not.
He, I think I'll put that post in a solid frame on the office wall, and
occasionally use it to beat some of my OOP-overzealous colleagues over
the head with
<G> - IOW, well put!
--
Kristofer
Kristofer Skaug
>
> Btw is the "Extract Method" supposed to work on regular
> procedures/functions, or only methods?
For the record, yes it does!
--
Kristofer
