Delphi 2.0 wish list
Bill Raike
>Suggestion: Only add simple requests to this list.
>
>1. In the BC++ IDE editor placing the cursor on either end of a { .. }
>pair and pressing Ctrl+q+] repositions the cursor to the other end. It
>would be nice if Delphi 2.0 had this feature for begin .. end pairs.
>
2. Allow static linking to .LIB files in addition to just .OBJ files.
Bill Raike
Jonathan W. Hendry
: rus...@belding.co.nz (Russell Belding) wrote:
3. A richer runtime support library, with more support for dynamic methods.
4. Objective-C style Categories, to allow adding methods to existing classes
without recompiling (don't create a Soundex component, add that capability
to _all_ the existing VCL controls)
5. Distributed Objects with a bundle of NeXT's D'OLE when it becomes
available.
Russell Belding
Jim Burns
appreciate hearing about it.... but I've found it extremely
irritating that I cannot seem to specify "final" output
(i.e. the .exe) to a seperate directory. I must say that I
love BC++'s option. They do allow it there.
Interestingly, and I know I am certainly biased, but in a
gui environment... what sense does it make to force me to
live with .exe's mixed in with .dcu's? It sounds so minor
a complaint but topically these files are distinct.
Please, let me put things where I want.
--
-- Jim Burns
Technology Dynamics
Pearland, Texas
Duncan Murdoch
>3. A richer runtime support library, with more support for dynamic methods.
Just curious: what sort of things do you want?
Duncan Murdoch
Mike Salisbury
Currently, when I try to step over a method that's recursively called
(via other objects in my case) control comes back to me in the middle
of the deeper instantiation. I just want to get control when I get
to the next line and the call stack is the same as it was when I
pressed F8.
-- Mike Salisbury
Jonathan W. Hendry
: hen...@MCS.COM (Jonathan W. Hendry) wrote:
: >3. A richer runtime support library, with more support for dynamic methods.
: Just curious: what sort of things do you want?
Things like these objective-C functions:
As the name implies, these let you add methods to a class at run time.
extern void class_addMethods(Class, struct objc_method_list *);
extern void class_removeMethods(Class, struct objc_method_list *);
This lets one class 'pose' as another. The runtime will send messages
intended for the original class to the imposter class. The imposter
takes the original's place in the inheritance hierarchy. This can
be highly nasty, but is useful at times.
extern Class class_poseAs(Class imposter, Class original);
extern unsigned method_getNumberOfArguments(Method);
extern unsigned method_getSizeOfArguments(Method);
extern unsigned method_getArgumentInfo(Method m, int arg, const char **type, int *offset);
While this sort of thing may not be needed by the average programmer,
there are people who can use this stuff to great effect.
Jasper Stil
2. I'd like a property "Transparent" (True/False) for all visual
components. Yes, even if that seems idiot right know. You never cab
tell and it could always be deafault False...
Jasper
"The only thing I know for sure
is that I don't know much... "
George Smith
>rus...@belding.co.nz (Russell Belding) wrote:
>>Suggestion: Only add simple requests to this list.
>>1. In the BC++ IDE editor placing the cursor on either end of a { .. }
>>pair and pressing Ctrl+q+] repositions the cursor to the other end. It
>>would be nice if Delphi 2.0 had this feature for begin .. end pairs.
>Better idea if borland borrowed a little more from there C++ compiler
>so...
>#define { begin
>#define } end
>much nicer :-)
They could change :- to := and call it Prolog to Pascal conversion.
Michael Snell
>Bill Raike (sra...@iconz.co.nz) wrote:
>: rus...@belding.co.nz (Russell Belding) wrote:
>
>: >Suggestion: Only add simple requests to this list.
>: >
>: >1. In the BC++ IDE editor placing the cursor on either end of a { .. }
>: >pair and pressing Ctrl+q+] repositions the cursor to the other end. It
>: >would be nice if Delphi 2.0 had this feature for begin .. end pairs.
>: >
>
>: 2. Allow static linking to .LIB files in addition to just .OBJ files.
>
>3. A richer runtime support library, with more support for dynamic methods.
>
>4. Objective-C style Categories, to allow adding methods to existing classes
>without recompiling (don't create a Soundex component, add that capability
>to _all_ the existing VCL controls)
>
>5. Distributed Objects with a bundle of NeXT's D'OLE when it becomes
>available.
6. A window that lists all declerations (functions, procedure etc) of
the current project in a outline, so that I can instantly jump to the
one I want, rather than BLOODY SEARCHING EVERY TIME.
-- Michael Snell
-- mic...@up.ac.za
-- http://penrose.up.ac.za/~michael
.. There are no atheists in the foxholes
Mike Frisch
: 3. Do not self-destruct the component pallette when a component doesn't load
: correctly. (Any body else out there try to start Delphi with Dwinsock on the
: pallette, and Winsock is not available?)
Agreed. I was pulling my hair out for hours before I booted into
Windows 3.1 (from Windows '95) which reported a DLL failed to load in the
toolbar. ARGH! I wish Windows '95 wouldn't hide these messages...
Mike.
--
==========================================================================
Mike Frisch Internet: mfr...@saturn.tlug.org
Northstar Technologies Compuserve: 76620,2534
Newmarket, Ontario, CANADA WWW: http://www.io.org/~mfrisch
Steve Teixeira
>
>3. Do not self-destruct the component pallette when a component doesn't load
>correctly. (Any body else out there try to start Delphi with Dwinsock on the
>pallette, and Winsock is not available?)
>
Unfortunately, that's how Windows works. COMPLIB.DCL is a dynamic link library. If it
contains static references to another DLL that Windows cannot find, Windows will not load
complib. This is the sign of a poorly-written component. A component should NEVER
statically link to a DLL. It should call LoadLibrary, and call RegisterComponents only
if LoadLibrary succeeds.
-Steve Teixeira
stei...@borland.com
Daniel Carter
obvious 'dynamic' candidates not defined as dynamic?
....
6. The Find/Replace dialog boxes copy selected text into the 'Find'
edit box.
7. A 'Step out of function' button in the IDE debugger (like in
MSVC++)
Jonathan W. Hendry
: without some operating system support. I know that there are clever (and
: useful) things one can do by *adding* methods but it is the sort of thing
: that can only be safe if you write *all* the code in the application or for
: certain rather specific kinds of object.
Why would adding methods be unsafe? If you don't override existing methods,
then the only code calling the added methods is your own code. You aren't
changing the behavior of existing methds, so preexisting or third-party
code would be unaffected. In fact, preexisting and third party code
won't touch your added methods since they won't know about them.
Basically, the only code that the added methods should affect is your own
code.
Brian Hirt
>hen...@MCS.COM (Jonathan W. Hendry) wrote:
>
>typed?
>
As far as I'm concerned, Pascal as we know it from borland isn't very
type'd any more.
I agree with Jonathan about posing and adding methods to existing classes
without having to subclass. Anyone who has done much work with Objective-C
will realize how extremely usefull it is.
-brian
gg...@lehigh.edu
ites:
>Suggestion: Only add simple requests to this list.
>
>1. In the BC++ IDE editor placing the cursor on either end of a { .. }
>pair and pressing Ctrl+q+] repositions the cursor to the other end. It
>would be nice if Delphi 2.0 had this feature for begin .. end pairs.
>
> 2. Allow static linking to .LIB files in addition to just .OBJ files.
>
>3. A richer runtime support library, with more support for dynamic methods.
>
>4. Objective-C style Categories, to allow adding methods to existing classes
>without recompiling (don't create a Soundex component, add that capability
>to _all_ the existing VCL controls)
>
>5. Distributed Objects with a bundle of NeXT's D'OLE when it becomes
>available.
>
6. Search (Find,Replace) through all units, without selecting units
exclusively and moving to the top of that page. I really miss VB's search
utilities.
7. Wouldn't be nice to see function's or procedure's name in a box, while
moving through the unit-page?
{cengiz}
Bengt Richter
>--
How about Delphi/Options/Project/Directories/OutputDirectory?
Please don't put in the wish list things you aren't pretty sure
don't exist already. I don't mean to sound negative ;-)
Regards,
Bengt Richter
Jim Robb
>Suggestion: Only add simple requests to this list.
>
>1. In the BC++ IDE editor placing the cursor on either end of a { .. }
>pair and pressing Ctrl+q+] repositions the cursor to the other end. It
>would be nice if Delphi 2.0 had this feature for begin .. end pairs.
What would also be nice is, if you place your cursor on one side of a
(..) pair, or a begin..end pair, both sides would change to a
different colour.
--
Jim Robb "Standard disclaimers apply"
Fifth Avenue Collection Ltd.
Moose Jaw, SK, Canada Voice: 306-694-8188 Fax: 306-694-0610
Sioux Falls, SD, USA Voice: 605-335-0602 Fax: 605-335-0713
Jim Robb
>Suggestion: Only add simple requests to this list.
>
>1. In the BC++ IDE editor placing the cursor on either end of a { .. }
>pair and pressing Ctrl+q+] repositions the cursor to the other end. It
>would be nice if Delphi 2.0 had this feature for begin .. end pairs.
*. A utility that takes a project and builds a set of installation
disks (or disk images), complete with installation program.
David Taylor
>
>Jim Burns <jimb...@phoenix.net> wrote:
>
>>I'm new to Delphi, so if I've missed something I would sure
>>appreciate hearing about it.... but I've found it extremely
>>irritating that I cannot seem to specify "final" output
>>(i.e. the .exe) to a seperate directory. I must say that I
>>love BC++'s option. They do allow it there.
>>Interestingly, and I know I am certainly biased, but in a
>>gui environment... what sense does it make to force me to
>>live with .exe's mixed in with .dcu's? It sounds so minor
>>a complaint but topically these files are distinct.
>
>
>
>Please don't put in the wish list things you aren't pretty sure
>don't exist already. I don't mean to sound negative ;-)
>
Actually this does not do what he wants since it puts the .DCU and the
.EXE files in the ouput directory, and he wanted them separate. To
expand on the request I would like better control for Debug/Ship versions.
I would like to have a multiple target scheme where I can say if I want
the debug version or the final version built, each one going to separate
directories. Actually I would not even want it limited to debug/final
settings, but simply multiple target within the same project so I can
set any options I want on a per target basis. It would be great if you
could work on a DLL and an EXE within the same project and be able to
select which one you want built, or to have them all built, etc.
Dave
pi...@cam.org
Duncan Murdoch
>Function and operating overloading is definately one of the things I
>would put on my wish list. And I don't think it violate Pascal's
>philosophy at all.
When you say "function overloading", you have to be more specific.
For example, people often say that they want C++ style function
overloading. But that offers you almost nothing that you don't
already have: you can always figure out at compile time which C++
overloaded function will be called. The only difference from Delphi
is in the definition of the name of the function: it includes the
declared name only in Delphi, the declared name plus the argument
signature in C++.
I don't see any advantage to this kind of overloading. It's always
possible to figure out the function being called at compile time, but
not always easy. Source code should be extremely easy to figure out.
If you're going to make it less clear, you should gain something from
it. For example, if there were some kind of virtual method dispatch
function that depended on the run-time types of several arguments,
that might be worth having. (I'm not so sure though; the increased
obscurity might not be justified by the increased power.) However,
you gain no extra power in the C++ style of function overloading, so
there's no reason to make your code more obscure.
Duncan Murdoch
Edgar J. Young
1. Free, good access to M$ Access
or distribution rights to the Local Interbase database
for us who are writing stand alone Delphi applications.
2. A setup kit that would package a distribution (including
database components) in a single package with a setup.exe.
These aren't complaints, just an opportunity to make a truely
fantastic product yet better...
Bengt Richter
>In article <47e8el$8...@news.accessone.com>,
>Bengt Richter <bo...@accessone.com> wrote:
>>How about Delphi/Options/Project/Directories/OutputDirectory?
>>
>But that puts both the .EXE and the .DCU in the target
>directory, not just the .EXE, as Jim pointed out.
>-David
>--
>David Weisberger
>djwe...@insti.physics.sunysb.edu
Yes. I owe Jim an apology. I missed his point,
and my patronizing admonishment has boomeranged
deservedly. I am sorry. I actually would like
the feature he was putting on the wish list.
Bengt Richter
[Exit sheepishly wiping egg from face]
Jonathan W. Hendry
How about making each component a VCL, and loading each
individually?
--
Jonathan W. Hendry Steel Driving Software, Inc.
Delphi and NeXTSTEP consulting and software development.
Cincinnati, Ohio
Mathew Gyokers Eliot
Here's something I've been propounding for a while now:
If every time you type "begin", you've got to type "end", why doesn't
the editor do it for you and automatically indent, etc, as needed?
The same goes for any tokens that have a start and a finish.
--
Delphi and Paradox Consultant. Member: Borland Delphi Technical Support
Web Page being relocated- want to provide space?
Borland supplied disclaimer: The services I provide via this message
and all consulting services I may provide are in no way associated with
or sponsored by Borland International or any of its subsidiaries.
David Weisberger
Todd Dunnavant
>From: "Edgar J. Young" <te...@ix.netcom.com>
>Subject: Re: Delphi 2.0 wish list
>Date: 4 Nov 1995 13:17:34 GMT
How about a robust data structures library, a la BIDS in BC++?
------------------------------------------------------
Regards,
Todd Dunnavant Houston, Texas tdun...@neosoft.com
pr...@bix.com
>>: rus...@belding.co.nz (Russell Belding) wrote:
>>
>>: >Suggestion: Only add simple requests to this list.
>>: >
>>: >1. In the BC++ IDE editor placing the cursor on either end of a { .. }
>>: >pair and pressing Ctrl+q+] repositions the cursor to the other end. It
>>: >would be nice if Delphi 2.0 had this feature for begin .. end pairs.
>>
>>: 2. Allow static linking to .LIB files in addition to just .OBJ files.
>>
>>3. A richer runtime support library, with more support for dynamic
>methods.
>>
>>4. Objective-C style Categories, to allow adding methods to existing
>classes
>>without recompiling (don't create a Soundex component, add that capability
>
>>to _all_ the existing VCL controls)
>>
>>5. Distributed Objects with a bundle of NeXT's D'OLE when it becomes
>>available.
>
>the current project in a outline, so that I can instantly jump to the
>one I want, rather than BLOODY SEARCHING EVERY TIME.
>
7. A simplified set of functions. Things like @PrintCurrentForm, @Register
for DLLs, @ColumnCount and other data-aware math functions would go a long
way towards speeding up our app development.
Thanks,
Paul Rice
Rusmin Noer
Have my Object Inspector's CHANGED-properties's font to be BOLD,
it's easier for me to inspect them.
Amin,
mystique
>In article <47fp5e$d...@ixnews5.ix.netcom.com> "Edgar J. Young" <te...@ix.netcom.com> writes:
>>From: "Edgar J. Young" <te...@ix.netcom.com>
>>Subject: Re: Delphi 2.0 wish list
>>Date: 4 Nov 1995 13:17:34 GMT
>How about a robust data structures library, a la BIDS in BC++?
what I want is something that MS had in VB3.0 . Even though VB is
obvoiusly not as good as Delphi.... I would really like to have that
automatic syntax correction thingy so that it checks if you mispelled
something.
Mystique
Duncan Murdoch
<stee...@ix.netcom.com> wrote:
>Steve Teixeira wrote:
>>
>> Unfortunately, that's how Windows works. COMPLIB.DCL is a dynamic link library. If it
>> contains static references to another DLL that Windows cannot find, Windows will not load
>> complib. This is the sign of a poorly-written component. A component should NEVER
>> statically link to a DLL. It should call LoadLibrary, and call RegisterComponents only
>> if LoadLibrary succeeds.
>
>How about making each component a VCL, and loading each
>individually?
Did you mean to say "make each component a DLL"? I'm not sure what a
"VCL" is, if not.
If you did mean DLL, I don't think it's such a good idea. Yes, it
protects you from sloppy component writers, but it will cost in terms
of load time for Delphi and will use up lots of file handles for all
those open DLLs. Shouldn't you just demand that component writers do
it right, and not require everyone to pay a performance penalty
because some don't?
Duncan Murdoch
Sean Malloy
>gent...@sisna.com (The Gentleman) wrote:
>>3. Do not self-destruct the component pallette when a component doesn't load
>>correctly. (Any body else out there try to start Delphi with Dwinsock on the
>>pallette, and Winsock is not available?)
>Unfortunately, that's how Windows works. COMPLIB.DCL is a dynamic link library. If it
>contains static references to another DLL that Windows cannot find, Windows will not load
>complib. This is the sign of a poorly-written component. A component should NEVER
>statically link to a DLL. It should call LoadLibrary, and call RegisterComponents only
>if LoadLibrary succeeds.
This brings up a question specific to the Winsock library that I'm not
familiar enough with the internals of Windows to answer (and currently
don't have the time to experiment with):
Netscape Communications made the file MOZOCK.DLL available for FTP,
which you are supposed to rename to WINSOCK.DLL and stick in your
WINDOWS\SYSTEM directory to let you run Winsock programs without
needing to have a Winsock connection established.
Does Windows rely on the _filename_ of the DLL once it's loaded, or
just the internal library names? If the filename wasn't relevant once
the DLL loaded, then you could design your component so that it called
LoadLibrary on WINSOCK.DLL, and if that failed, to call LoadLibrary on
MOZOCK.DLL -- giving you a fallback to let you do development without
having to establish a Winsock connection -- and only fail without
calling RegisterComponent if MOZOCK.DLL was missing.
--
Sean R. Malloy | American Non Sequitur
Naval Medical Center | Society
San Diego, CA 92134-5000 |
mal...@cris.com | "We may not make sense,
srma...@snd10.med.navy.mil | but we do like pizza"
Aurelio Joaniquet
See the connections between elements in a Form/s , e.g.: In complex
forms you can have multiple connection between components (often in
Database entry Forms), TDBEdit, TDBGrid, TTable, TQuery, etc.. What I
would like is somehing that would show me that connections in a
similiar way other visual tools do (like IBM Visual Age, Netware
AppWare,...). Preferably, this has to be an optional feature (maybe
Speed Menu), because it can go too messy.
Phil Scadden
beefed up in many ways. While I dont have a mechanism, it would be
nice for a language to allow an Extend object (like a user extended
Tcanvas) to be used instead of the supplied one but I found this to
much more difficult than expected with Tcanvas being a PROPERY of so
many things.
Anyway, an extended Tcanvas should cover things like:
WinG access
Polygon,polyline methods that allow dynamic arrays to by passed
Encapulation of common WinAPI calls like:
SetbkMode
Regions
InvalidateRect and updateWindow (combine both - use a parameter)
SetStretchmode (add it to CopyRect)
And perhaps a cleanup of palette handling. How many questions about
256 colour handling have been asked here? I wonder if palettes should
be made a property of a Form (or perhaps a canvas)? There are hints that
this might have been the initial conception.
--------------------------------------------------------------------
Phil Scadden, Scadden Research, P.Sc...@gns.cri.nz
55 Buick St, Petone, Lower Hutt
New Zealand, Ph ++64 4 5687190, fax ++64 4 5695016
Harry Culpan
>hcu...@ix.netcom.com (Harry Culpan) wrote:
I'm afraid I have to disagree with you on this. Function overloading
as you defined it does add considerable power, and I think leads to
more, rather than less, readable code. As an example, I had a
situation wherein I passed a parameter by reference, so that I could
then fill the parameter with a value I got from another source (a
Btrieve database, although that is unimportant). Since I was copying
data byte by byte, I needed to know whether the data was a string (in
which case I ignore the first byte) or a number. There are of course
several solutions to this problem, and I chose to create two functions
with suffixes that specified the data type. With function
overloading, however, I could have used the same name. Keeping in
mind that I wrote this for others to use, it would have made it
simpler for them to consistently call one function regardless to type
rather than to call two different functions.
As for anyone going back through my code, a quick look at the list of
prototypes should suffice to tell them how I am performing this
"magic" of being able to pass different parameter types. Thus, at
least in my opinion, my code would be no less readable than before,
while the code produced by others using my component would be more
readable.
Harry W. Culpan
Kent Tong
>procedure TObserver.Body; { Descendant of TProcess }
> function ThereAreCars: Boolean; far;
> Begin
> ThereAreCars := NumberOfCars > 0;
> End;
>Begin
> ....
> WaitUntil(@ThereAreCars); { Pass the adress of the local function }
> ...
>End;
A better solution should be:
type
TMonitor = object
begin
public
procedure WaitUntil; virtual;
end;
TMonitorCars = object(TMonitor)
begin
public
procedure WaitUntil;
end;
procedure TObserver.Body(m: TMonitor); { Descendant of TProcess }
Begin
....
m.WaitUntil;
...
End;
---
Kent Tong
Win95 version is out!!!
Freeman Installer ==> http://www.netnet.net/users/freeman/
Olaf Rauch
>I wish to:
To follow up your idea it would also be nice to have the non-visual
components (especially TTable, TDataSource) on some kind of separate
window.
On forms with a lot of such components on it confusion starts coming
up.
/
/ Olaf Rauch - Student of Computer Science - Germany
/ i2...@informatik.fh-wuerzburg.de
/
Olaf Rauch
>Bill Raike (sra...@iconz.co.nz) wrote:
>: rus...@belding.co.nz (Russell Belding) wrote:
>: >Suggestion: Only add simple requests to this list.
>: >
>: >1. In the BC++ IDE editor placing the cursor on either end of a { .. }
>: >pair and pressing Ctrl+q+] repositions the cursor to the other end. It
>: >would be nice if Delphi 2.0 had this feature for begin .. end pairs.
>: >
>: 2. Allow static linking to .LIB files in addition to just .OBJ files.
>3. A richer runtime support library, with more support for dynamic methods.
>4. Objective-C style Categories, to allow adding methods to existing classes
>without recompiling (don't create a Soundex component, add that capability
>to _all_ the existing VCL controls)
>5. Distributed Objects with a bundle of NeXT's D'OLE when it becomes
>available.
* Some kind of group/ungroup function in the form designer.
* Lock controls or groups of controls separately
Kerry Sanders
}obvoiusly not as good as Delphi.... I would really like to have that
}automatic syntax correction thingy so that it checks if you mispelled
}something.
Will not happen, because Delphi is a compiled language. It doesn't interpret
what you type during design time like VB does. Heck... the compile times for
Delphi are so quick, it doesn't really matter anyway.
----------------------------------------------------------------------
Kerry Sanders | "Cause when ya stop dreaming
E-Mail: ksan...@mindspring.com | its time to die" - CHANGE
----------------------------------------------------------------------
Tony Lownds
>dmur...@mast.queensu.ca (Duncan Murdoch) wrote:
>>hcu...@ix.netcom.com (Harry Culpan) wrote:
>>>Function and operating overloading is definately one of the things I
>>>would put on my wish list. And I don't think it violate Pascal's
>>>philosophy at all.
I agree! Add templates, assignment operators, named and optional
parameters and deand I'll (almost) never be tempted to use C++ again.
>>When you say "function overloading", you have to be more specific.
>>For example, people often say that they want C++ style function
>>overloading. But that offers you almost nothing that you don't
>>already have: you can always figure out at compile time which C++
>>overloaded function will be called. The only difference from Delphi
>>is in the definition of the name of the function: it includes the
>>declared name only in Delphi, the declared name plus the argument
>>signature in C++.
I would NOT define the return type as a part of the function signature
as in C++.
>>I don't see any advantage to this kind of overloading. It's always
>>possible to figure out the function being called at compile time, but
>>not always easy. Source code should be extremely easy to figure out.
>>If you're going to make it less clear, you should gain something from
>>it. For example, if there were some kind of virtual method dispatch
>>function that depended on the run-time types of several arguments,
>>that might be worth having. (I'm not so sure though; the increased
>>obscurity might not be justified by the increased power.) However,
>>you gain no extra power in the C++ style of function overloading, so
>>there's no reason to make your code more obscure.
Then don't use it. I would propose that overloaded functions be tagged
as such with a keyword (such as overloaded):
Unit System;
{...}
procedure Frac(x: real): Integer;
{...}
end;
Unit RatMath;
type
Rational={...};
{...}
procedure Frac(x:Rational): Integer; overloaded;
{...}
end;
>I'm afraid I have to disagree with you on this. Function overloading
>as you defined it does add considerable power, and I think leads to
>more, rather than less, readable code. As an example, I had a
>situation wherein I passed a parameter by reference, so that I could
>then fill the parameter with a value I got from another source (a
>Btrieve database, although that is unimportant). Since I was copying
>data byte by byte, I needed to know whether the data was a string (in
>which case I ignore the first byte) or a number. There are of course
>several solutions to this problem, and I chose to create two functions
>with suffixes that specified the data type. With function
>overloading, however, I could have used the same name. Keeping in
>mind that I wrote this for others to use, it would have made it
>simpler for them to consistently call one function regardless to type
>rather than to call two different functions.
>As for anyone going back through my code, a quick look at the list of
>prototypes should suffice to tell them how I am performing this
>"magic" of being able to pass different parameter types. Thus, at
>least in my opinion, my code would be no less readable than before,
>while the code produced by others using my component would be more
>readable.
Function overloading of tagged function based on parameters would be,
IMHO, a pretty benign addition to OP. What is less clear is the issue
of operator overloading. I am not sure it should be included as a
'simple suggestion'.
-Tony Lownds
Tony Lownds
>myst...@TheRealmOfImagination.com (mystique) wrote:
>}obvoiusly not as good as Delphi.... I would really like to have that
>}automatic syntax correction thingy so that it checks if you mispelled
>}something.
>Will not happen, because Delphi is a compiled language. It doesn't interpret
>what you type during design time like VB does. Heck... the compile times for
>Delphi are so quick, it doesn't really matter anyway.
Delphi is a one-pass language so it could happen - and it would be an
incredibly useful feature.
Bengt Richter
>rus...@belding.co.nz (Russell Belding) wrote:
>>Suggestion: Only add simple requests to this list.
>>
>>1. In the BC++ IDE editor placing the cursor on either end of a { .. }
>>pair and pressing Ctrl+q+] repositions the cursor to the other end. It
>>would be nice if Delphi 2.0 had this feature for begin .. end pairs.
>>
>2. Allow static linking to .LIB files in addition to just .OBJ files.
>Bill Raike
I would like to see names of units disassociated from the names
of their containers (files). IMHO that was a big mistake, however
expedient it might have been. It is an inappropriate confusion of
name spaces that ought to be distinct. The IDE can maintain
a table of where to find program entities, perhaps even as URLs.
Doing this should also ease platform migration.
Regards,
Bengt Richter
Slobodan Celenkovic
>possible to figure out the function being called at compile time, but
>not always easy. Source code should be extremely easy to figure out.
>If you're going to make it less clear, you should gain something from
>it. For example, if there were some kind of virtual method dispatch
>function that depended on the run-time types of several arguments,
>that might be worth having. (I'm not so sure though; the increased
>obscurity might not be justified by the increased power.) However,
>you gain no extra power in the C++ style of function overloading, so
>there's no reason to make your code more obscure.
It doesn't have to be obscure. Here is an example. I was writing a
dynamic array object. I want a method which removes an item from the
array. In C++ I wrote:
void Remove(TIndex Index);
void Remove(TIndex Index, unsigned Count);
void Remove();
1st Removes the specified item, second removes the specified number of
items starting at Index, third removes all items. So in code you would
write:
Arr.Remove( i );
Arr.Remove( i, 10 );
Arr.Remove();
Now to me that doesn't look obsure, on the contrary it looks quite
natural. In Pascal I created 2 version, Remove (same as #1) and
RemoveMulti (same as #2).
Similarly operator overloading can be just as effective.
Of course, overloading can be dangerous if not used selectivly and
wisely. What looks clear/obvious to me, might look obscure to others. So
it is a "dangerous feature" and indeed great care must be taken when
using it.
Bob
Harry Culpan
>>I'm afraid I have to disagree with you on this. Function overloading
>>as you defined it does add considerable power,
>I'm using "power" to mean that converting a program that uses this
>feature to one that doesn't takes a lot of work. C++ function
>overloading adds no power, because converting is as simple as
>including the argument signature as part of the name: it can be done
>mechanically.
>> and I think leads to
>>more, rather than less, readable code.
>You gave an example of one function taking a string, another taking an
>integer. But what if I wanted one for words, one for integers? Is 34
>a word or an integer (or a byte or a shortint or a real or a
>double...)? C++ has a complicated set of rules for resolving such
>ambiguities. Delphi has a very simple set of rules, since the
>ambiguity can't happen. I think that makes Delphi more readable.
Without trying to be argumentative, I should point out that I did not
make any claim that C++ was easier to read than Delphi. My only point
was that function overloading as implemented in C++ can lead to code
that is as readable - if not more so - than current Delphi syntax. In
addition, as my example was intended to demonstrate, it can make a
component or unit easier for others to use.
Of course, we have thus far ignored operator overloading, but I think
the same arguments apply, but even more so. While you might say that
these add no "power" (i.e., an overloaded operator can be mechanically
replaced by a function), it almost certainly leads to more readable
code - assuming one uses operator overloading well.
Harry W. Culpan
Kendall P. Bullen
Slobodan Celenkovic <slob...@cs.unh.edu> wrote:
>void Remove(TIndex Index);
>void Remove(TIndex Index, unsigned Count);
>void Remove();
>
>1st Removes the specified item, second removes the specified number of
>items starting at Index, third removes all items. So in code you would
>write:
>
> Arr.Remove( i );
> Arr.Remove( i, 10 );
> Arr.Remove();
>
>Now to me that doesn't look obsure, on the contrary it looks quite
>natural. In Pascal I created 2 version, Remove (same as #1) and
>RemoveMulti (same as #2).
No offense, but a 'need' for this seems sloppy. You should have one function
to do the cases of removing one, some, or all. Having three functions to do
one thing (remove 1-n items from an array) seems inefficient.
Is this a good example of what some people are wishing for? If so, it seems
very kludgy (sp?).
No offense intended. Perhaps I'm completely misunderstanding, or perhaps we
just have different programming styles, or both. :)
Kendall
Shamsul
1) Procedure Overloading
2) Function Overloading
3) Operator Overloading
IMHO, method overloading captures the basic (pascal ;->) essence
of OOP.
Boris Ingram
>From: joan...@eq.upc.es (Aurelio Joaniquet)
>Subject: Re: Delphi 2.0 wish list
>I wish to:
>See the connections between elements in a Form/s , e.g.: In complex
>forms you can have multiple connection between components (often in
>Database entry Forms), TDBEdit, TDBGrid, TTable, TQuery, etc.. What I
>would like is somehing that would show me that connections in a
>similiar way other visual tools do (like IBM Visual Age, Netware
>AppWare,...). Preferably, this has to be an optional feature (maybe
>Speed Menu), because it can go too messy.
I found a thing called tconnections, which i donloaded from somewhere - can't
remember. The author is : Kenneth W. Clubb
and his e-mail is : 73047...@compuserve.com.
Try to get a copy
Boris Ingram, Cyborg Software
bori...@iafrica.com
http://www.pcb.co.za/users/borising/cyborg.htm
Eric Begin
- Controls array like Visual Basic
- An TDateTime class, not only a record
- New easter eggs :)
Is it possible to get some feedbacks from Borland about
our suggestions?
+---------------------+----------------------------------------+
| Eric Begin | Eric....@absolu.com |
| Software designer | beg...@roselin.dmi.usherb.ca |
| Absolu Technologies | |
+---------------------+----------------------------------------+
Harry Culpan
Actually, in my opinion, default parameter values (a la C++) would be
a better way to handle this. In the above case, you could rewrite it
to be:
void Remove(TIndex Index = <clear>, unsigned Count = 1)
The "<clear>" above would be whatever value would equate to Clear of
the TIndex type.
In this way, you could write exactly the same code as the original
poster presented, but still have one function to handle it.
Harry W. Culpan
Phil Pickard
wrote:
>Suggestion: Only add simple requests to this list.
>
>1. In the BC++ IDE editor placing the cursor on either end of a { .. }
>pair and pressing Ctrl+q+] repositions the cursor to the other end. It
>would be nice if Delphi 2.0 had this feature for begin .. end pairs.
>
>
Can we have nested comments? {Nested comments {PLEASE!!}}
Steve Teixeira
>My wish for Delphi 2.0
>
>- Controls array like Visual Basic
>- An TDateTime class, not only a record
>- New easter eggs :)
>
>Is it possible to get some feedbacks from Borland about
>our suggestions?
>
While we can't comment on specific features of Delphi32 at
this time, we are paying attention to the suggestions put
forth in this thread.
-Steve Teixeira
stei...@borland.com
David R. Fentem
utilize custom hardware I've designed, and the "Windows Way" of doing I/O is
through a Windows Device Driver. I'd envision that a VxD would be done
similar to the way a DLL is done now, with those exceptions that make sense
(e.g. very limited runtime library, etc.)
The current way of developing VxD's requires Microsoft's SDK/DDK. This is
somewhat reminiscent of the way we used to have to write applications using
the SDK... thankfully that's history now. Maybe Borland can fill this void,
also.
+---------------------------------------------------------------------+
| David R. Fentem | david....@gtri.gatech.edu |
| Georgia Tech Research Institute | CompuServe: 70674,3306 |
| GTRI/ELSYS/ESMD | Voice: (404) 894-7045 |
| Atlanta, GA 30332-0800 | FAX: (404) 894-7080 |
+---------------------------------------------------------------------+
The Lord of Darkness
pop up menu with cut/copy/paste etc
now it's got lots more (no complaint), but I'd like the
cut/copy/paste duplicated on the menu...might be possible to
add these with resource workshop...Hmmm
Also the search/replace doesn't work the same (very poorly done
now- doesn't start at the top when you ask for entire scope)
John B
Z Solutions, LLC
>While we can't comment on specific features of Delphi32 at
>this time, we are paying attention to the suggestions put
>forth in this thread.
> -Steve Teixeira
> stei...@borland.com
I'd like to see the component.cursor to work. Presently you have to
use screen.cursor. Example: I want the cursor to change from an
hourglass to an arrow when the user moves the cursor over say a button
component, then have it go back to an hourglass when the cursor moves
off the button. I can't get it to work with version 1.0.
-------------------------
Steve Kammerer
Z Solutions, LLC
zs...@mindspring.com
*******************************************************
http://www.mindspring.com/~zsol/nnintro.html
*******************************************************
p...@pobox.com
>>would put on my wish list. And I don't think it violate Pascal's
>>philosophy at all.
>
>For example, people often say that they want C++ style function
>overloading. But that offers you almost nothing that you don't
OK Time to disagree. Here is why I want Objective Overloading or
whatever somebody called it.
I want to add a property (or whatever) to every component in my lib.
Currently I have to descend from EACH one. Why not be able to just add to
TComponent? (Remember, you may not have access to the source code, and even
if you do, interdependencies will foil that usually anyways.)
--------------------------------------------------------
|Chad Z. Hower - P...@pobox.com - http://pobox.com/~pbe |
|Independent Consultant as Phoenix Business Enterprises|
| & |
|Senior Analyst Programmer - Seltmann, Cobb & Bryant |
--------------------------------------------------------
_
Pennsylvania: Communistwealth of Pennsylvania
>> SQUID - The Ultimate Usenet Utility #BLYGJOMK
Special Compile: 1.032/1.033B (Beta)
David Brabant [SNI]
> or distribution rights to the Local Interbase database
> for us who are writing stand alone Delphi applications.
>2. A setup kit that would package a distribution (including
> database components) in a single package with a setup.exe.
>
>These aren't complaints, just an opportunity to make a truely
>fantastic product yet better...
For marketing reasons, I would add: a good VB to Delphi converter
distributed for free with the package.
David
--
David Brabant, | E-mail: David....@csl.sni.be
Siemens Nixdorf (SNI), | CIS: 100337,1733
Centre Software de Liège, | X-400: C=BE;A=RTT;P=SCN;O=SNI;OU1=LGG1;OU2=S1
2, rue des Fories, | S=BRABANT;G=DAVID
4020 Liège (BELGIUM) | HTTP: www.sni.de www.csl.sni.be/~david
alexander m nolasco
Dear Borland:
How about adding generics to Delphi?
PS: Pssss, VB 4.0 does not have that "yet"...
Marc Chapman
A TQuery component for SQL, or at least one that has an option to pass a
query through without checking and still give a live result set.
Russell Belding
i2...@informatik.fh-wuerzburg.de says...
>
>>: >Suggestion: Only add simple requests to this list.
>>: >
>>: >1. In the BC++ IDE editor placing the cursor on either end of a { ..
}
>>: >pair and pressing Ctrl+q+] repositions the cursor to the other end.
It
>>: >would be nice if Delphi 2.0 had this feature for begin .. end pairs.
>
>>: 2. Allow static linking to .LIB files in addition to just .OBJ files.
>
>>3. A richer runtime support library, with more support for dynamic
methods.
>
>>4. Objective-C style Categories, to allow adding methods to existing
classes
>>without recompiling (don't create a Soundex component, add that
capability
>>to _all_ the existing VCL controls)
>
>>5. Distributed Objects with a bundle of NeXT's D'OLE when it becomes
>>available.
>
>* Some kind of group/ungroup function in the form designer.
>
>* Lock controls or groups of controls separately
Eg. In a recent development session I got "Error : Some internal limit
(See context)" and that was all the message said.
7. Have someone independent go through each help screen and produce a
report on all errors in the hypertext links (in 1.02). Borland to correct.
8. Have the compiler find more than one error before it stops. BC++ does
this with skill. Perhaps too hard.
9. Put content and context into the help system examples. No one liners.
10. Include an install system which allows developers to package their
apps and which packages the BDE with the developers app, if requested.
To pay for the extra development costs charge more for Delphi 2.0, if you
have to. It is and will be worth it.
Guy Hussussian
Dear Santa,
I would like the borland container class libs in delphi. This should
be made available with the standard Delphi distribution.
C YA,
-guy
Christopher Fynn
probably be very popular (sprites and things).
Better and more complete documentation including
a really coprehensive one on Object Pascal with
lots of examples.
--
Christopher J Fynn <cf...@sahaja.demon.co.uk>
Kendall P. Bullen
rus...@belding.co.nz (Russell Belding) wrote:
>8. Have the compiler find more than one error before it stops. BC++ does
>this with skill. Perhaps too hard.
I think it is not just too hard, but not possible. The first error may be
what's causing the "next" one -- a misplaced semicolon or 'END' (shudder) can
mess up a lot of stuff, and you'd get false errors all over the place. I'd
prefer to find my errors one at a time and know there's a good chance they're
actual errors, thankyouverymuch. ;)
Kendall
Paul Sleigh
I LOVE a captive audience...
My own personal wish list, off the top of my balding head...
Operator overloading as follows:
operator "+"(const AList: TList; const AItem: TObject): TList;
begin ... end;
(Nice and easy, adds one reserved word, breaks no existing code,
may be safely ignored.)
Function overloading as follows:
function Put(const AString: string);
begin ... end;
function Put(const AInteger: LongInt);
begin ... end;
function Put(const AWord: Word);
begin ... end;
(* this last will not compile; ambiguous resolution error! *)
(Also fairly easy, with C++ style internal name mangling; breaks some code,
so maybe it'd be better if some directive like "overloaded" were
required for ALL instances, INCLUDING the first one)
program ForExample;
begin
for A: Integer := 1 to 10 do
WriteLn(A,' Hello!');
end.
(In other words, don't require that a for loop control variable be
predefined -- let it be declared just for the loop, then disappear out of
scope immediately afterward. Easy to parse, breaks no existing code)
That'll do for now. Comments?
: Eric :
--
Razors pain you; rivers are damp; : Paul Sleigh
Acids stain you; and drugs cause cramp. : (Eric the Fruitbat)
Guns aren't lawful; nooses give; : fruitbat@canberra
Gas smells awful; you might as well live. - D Parker : .DIALix.oz.au
David Taylor
>
>Can we have nested comments? {Nested comments {PLEASE!!}}
>
Actually, I would be happy simply with an EOL style comment, such
as the C++ style of //. In fact I am not sure I would like to see
nested comments, I think using two different comment types is
better, such as (* { } *) for nested comments. If you only use
{} comments in your code you can use the (* *) comments to block
out whole sections.
Dave
pi...@cam.org
Marc Chapman
mind posting it? I have missed the first few, and they are now expired.
Also, I have a couple requests and would like to add to the list, but
not repeat. Thank you very much.
Marcus.
Robert Small
Add the abilty to the search and search/replace functions in the IDE
to operate over all units in the project.
The options would be
selected text
current unit
all units
There have been quite a few times when I have needed to make "global"
name changes, or simply to find some text (and have forgotten which
unit it was in).
BTY, Steve, just got your book. Excellent.
But the CD doesn't work. I got my book via PC Tecniques and got the US
version. Can you tell me where I can ftp the Oz versions of the
printing examples?
==
Bob Small
Jon Jenkins
the basic Delphi is preety good as is (hopefully a decent
debugger and a few bug fixes will make it even better).
I understand that the 32 bit version will generate standard
obj and that C++ and Pascal will share a back end code
generator, if so great. This would facilitate seamless
integration of Pascal and C++ objects.
I would also like to see a completely integrated development
package which not only includes the code generator
but install/uninstall/registry as well help authoring
and image processing tools/components.
Perhaps as Borland seem to be jumping on the Java or HTML
wagon you could break from the MS imposed RTF/.hlp format
and include a Java/HTML based help facility either instead
of or as well as the MS standard.
Also useful to a smaller number of coders might be
VxD/device driver template components.
Inclusion of common interface components such as
COMM access, Winsock objects etc
Thats enough for now !!
Jon
--
----------------------------------------------------------------------
Name: Dr Jon Jenkins Location: Digital Equipment Corporation NaC
Voice/Fax: 61-7-55-75-0151/100 Burnett Place, Research Park,
Inet: jenk...@ozy.dec.com Bond University, Gold Coast
Close Proximity: "HEY YOU !!!" QLD, AUSTRALIA 4229
"Daddy, what's outside the Universe?" (My 5 year old.....)
-----------------------------------------------------------------------
Casey Charlton
easier (well actually 2).
GFA Basic (a very old program on the Atari ST) had a wonderful feature
that I haven't seen elsewhere. It had the ability within the editor to
'fold' procedures up, so that only the definition was showing. The
ability to fold up procedures and function to a single line removes a
great deal of clutter from the screen and enables very quick
navigation of a large program (or unit).
On the same sort of subject I would really appreciate a proper macro
capability within the editor/IDE. Even a glorified recording facility
would be useful!
Kendall P. Bullen
frui...@canberra.DIALix.oz.au (Paul Sleigh) wrote:
>Bonus! Borland are paying attention to this thread!
>I LOVE a captive audience...
Well, hardly captive. They can leave anytime they like, they can ignore
anything they like, etc. ;)
>Operator overloading as follows:
>
> operator "+"(const AList: TList; const AItem: TObject): TList;
> begin ... end;
>
>(Nice and easy, adds one reserved word, breaks no existing code,
>may be safely ignored.)
For the unoverloaded among us, um, what's this mean (operator overloading,
that is)?
>program ForExample;
>begin
> for A: Integer := 1 to 10 do
> WriteLn(A,' Hello!');
>end.
>
>(In other words, don't require that a for loop control variable be
>predefined -- let it be declared just for the loop, then disappear out of
>scope immediately afterward. Easy to parse, breaks no existing code)
I think it's a bad idea; it kinda goes against the basis of Pascal as a
strongly typed & defined language. Also promotes sloppy programming, IMNSHO.
;)
Kendall
Kendall P. Bullen
>Newsgroups: comp.lang.pascal.delphi.misc
>Subject: Re: Delphi 2.0 wish list
>Organization: Not much
..
>GFA Basic (a very old program on the Atari ST) had a wonderful feature
>that I haven't seen elsewhere. It had the ability within the editor to
>'fold' procedures up, so that only the definition was showing. The
>ability to fold up procedures and function to a single line removes a
>great deal of clutter from the screen and enables very quick
>navigation of a large program (or unit).
That's a superb idea! Wow, I'm surprised I haven't heard of it before (or
thought of it before, for that matter ;).
>On the same sort of subject I would really appreciate a proper macro
>capability within the editor/IDE. Even a glorified recording facility
>would be useful!
Yes, the perennial request. (sigh)
Kendall
Medge
>>
>>1. In the BC++ IDE editor placing the cursor on either end of a { .. }
>>pair and pressing Ctrl+q+] repositions the cursor to the other end. It
>>would be nice if Delphi 2.0 had this feature for begin .. end pairs.
>
>disks (or disk images), complete with installation program.
2. A simple database component like VB. I have not had time to play with
Delphi and no one is going to be paying the bills for me to do it, so I
might be missing something here. The VB OpenDatabase and CreateDynaset etc
I find extremely easy, all these aliases in Delphi hmmmm... Actaully what
would be nice would be a good database tutorial somewhere.
Eric Begin
Programmer: "Allow to disabled items in a TGroupBox"
Amen
u821...@cc.nctu.edu.tw
: On 31 Oct 1995 01:23:36 GMT, rus...@belding.co.nz (Russell Belding)
: wrote:
: >Suggestion: Only add simple requests to this list.
: >
: >1. In the BC++ IDE editor placing the cursor on either end of a { .. }
: >pair and pressing Ctrl+q+] repositions the cursor to the other end. It
: >would be nice if Delphi 2.0 had this feature for begin .. end pairs.
: >
: >
1: Emacs-style editor is strong recommanded!! (is that simple??)
2: TQuery -> "primary table" can be edited
Slobodan Celenkovic
>ken...@his.com (Kendall P. Bullen) wrote:
>
>>No offense, but a 'need' for this seems sloppy. You should have one function
>>to do the cases of removing one, some, or all. Having three functions to do
>>one thing (remove 1-n items from an array) seems inefficient.
Indeed the 1st 2 may be fused into one, using default value of 1 for Count. However
the 3rd one is very different (implementation), hence should be a different
function.
>
>>Is this a good example of what some people are wishing for? If so, it seems
>>very kludgy (sp?).
>
>>No offense intended. Perhaps I'm completely misunderstanding, or perhaps we
>>just have different programming styles, or both. :)
>
>Actually, in my opinion, default parameter values (a la C++) would be
>a better way to handle this. In the above case, you could rewrite it
>to be:
>
>void Remove(TIndex Index = <clear>, unsigned Count = 1)
>
>The "<clear>" above would be whatever value would equate to Clear of
>the TIndex type.
>
>In this way, you could write exactly the same code as the original
>poster presented, but still have one function to handle it.
>
>Harry W. Culpan
>
Ok, so the example isn't that good. Perhaps better example are constructors.
Constructors often have to be overloaded because of the many different ways of
creating an object (look at the classes bundled with GNU's C/C++ compiler). Most
of them have very different implementations, hence are placed into different
functions. Yet functionally they do the same thing: create an object, just in
different ways.
Delphi makes an exception for constructors, allowing them to be overloaded but
using different identifiers. Why not allow overloading of any method, not just
constructors?
P.S.: Default parameter values would be neat, too.
Jim Eadie
>On 31 Oct 1995 01:23:36 GMT, rus...@belding.co.nz (Russell Belding)
>wrote:
>>Suggestion: Only add simple requests to this list.
>>
>>1. In the BC++ IDE editor placing the cursor on either end of a { .. }
>>pair and pressing Ctrl+q+] repositions the cursor to the other end. It
>>would be nice if Delphi 2.0 had this feature for begin .. end pairs.
>>
>>
>Can we have nested comments? {Nested comments {PLEASE!!}}
Seeing it's close to XMas .... here goes my wish list :
1. simple commenting ala C++
// this is a comment and i dont have to put sumpting at end
2. Editor window button that pops up the method/func/procs in the current
file so you can quickly jump to it
3. Project-like facility as in IDE files and utility of BC4.x .
This would allow one to see how large projects hang together
4. Function /operator overloading ,etc as described many times above
Hmmm ... I see where this is all leading!!!!!
Make it into Delphi C++ . Forget Pascal!
Before everyone jumps on me, keep in mind that any Delphi Pascal/VCL code
is NOT portable to anywhere I know of, it is a SINGLE platform/operating
system (so far) so there is no restriction on keeping the "Pascal" to any
"standard". And even if it does eventually migrate chances are it will be
under Windows NT. Also as the whole point Delphi is to make WINDOWS
programming easier there is no reason to NOT have a WINDOWS specific
language. Thus under a WINDOWS NT world Delphi could become the standard.
NB: I am not advocating NT or anything like it - just following a train of
thought
Jim
p...@pobox.com
>>void Remove(TIndex Index, unsigned Count);
>>void Remove();
>>
>> Arr.Remove( i );
>> Arr.Remove( i, 10 );
>> Arr.Remove();
>>
>No offense, but a 'need' for this seems sloppy. You should have one function
>to do the cases of removing one, some, or all. Having three functions to do
>one thing (remove 1-n items from an array) seems inefficient.
What would be a better solution is to use OPEN arrays.
Procedure Remove(iIn: open array);
begin
if High(iIn) = 0 then....
end;
You get the picture, this would then translate to:
Arr.Remove([i]);
Arr.Remove([i,1]);
Arr.Remove([]); (Is this legal?? You'll have to try this one out)
If the last is not legal, you can use an UNTYPED parameter and do it
that way.
--------------------------------------------------------
|Chad Z. Hower - P...@pobox.com - http://pobox.com/~pbe |
|Independent Consultant as Phoenix Business Enterprises|
| & |
|Senior Analyst Programmer - Seltmann, Cobb & Bryant |
--------------------------------------------------------
_
Pennsylvania: Is this is a road or a bombing range?
p...@pobox.com
>of their containers (files). IMHO that was a big mistake, however
>expedient it might have been. It is an inappropriate confusion of
>name spaces that ought to be distinct. The IDE can maintain
>a table of where to find program entities, perhaps even as URLs.
>Doing this should also ease platform migration.
Delphi 32 allows aliasing, which will probably allow you to do as you
wish. You can also make WINDOWS=WindProcs, Wintypes so that you can just say:
Uses
Windows;
--------------------------------------------------------
|Chad Z. Hower - P...@pobox.com - http://pobox.com/~pbe |
|Independent Consultant as Phoenix Business Enterprises|
| & |
|Senior Analyst Programmer - Seltmann, Cobb & Bryant |
--------------------------------------------------------
_
Pennsylvania ---> Taxsylvania
Robert Fabiszak
Actually, from what I've seen published (by Borland) about Delphi 32, it will,
in fact, detect and announce all of the compile errors, just like most C
compilers do.
I agree that a simple error like a misplaced END could "cascade" and cause
many spurious errors to be listed, but you can always just fix the first error
and recompile. I mean, with Delphpi it's not like you have to wait five
minutes for the compiler to chug away, like you do with a C++ compiler <g>.
Bob
Ron Paulsen
I believe that Paul Sleigh was enumerating a wish list that comes from C++ (correct
me if I'm wrong, Paul). Briefly:
Operator overloading: the ability to redefine operators so that they can operate on
custom data types. For instance, in a particular application it might make sense to
"add" two enumerated types (tangerine + grapefruit = tangelo).
Locally-defined loop variable (or any variable for that matter): a matter of style -
C++ folks have this technique available and like to use it.
Every language has its advantages in shortcomings. I believe that Paul would like to
see Delphi become more C++-like. We can all benefit from studying and using other
languages and C++ is an excellent choice for a Delphi user try.
-Ron Paulsen
-----------------------------------------------------------------------------------
ken...@his.com (Kendall P. Bullen) wrote:
Aurelio Joaniquet
>There's only one request that I can think off that would make my life
>easier (well actually 2).
>GFA Basic (a very old program on the Atari ST) had a wonderful feature
>that I haven't seen elsewhere. It had the ability within the editor to
>'fold' procedures up, so that only the definition was showing. The
>ability to fold up procedures and function to a single line removes a
>great deal of clutter from the screen and enables very quick
>navigation of a large program (or unit).
Very useful idea !
Eivind Bakkestuen
(F> many spurious errors to be listed, but you can always just fix t
(F> and recompile. I mean, with Delphpi it's not like you have to wa
(F> minutes for the compiler to chug away, like you do with a C++ co
I would be surprised if its not possible to choose which compile/error
reporting strategy to use.
--
Mr. HillBilly (eiv...@infolink.no)
Programmers BBS kjører kurs i grafikkprogrammering! Både C og Pascal!
"In C it is possible to shoot yourself in the
foot. With C++ you blow your whole leg off"
°HyperSoft HyperReader á 0.9°
Bob Mallett
In article <47tute$s...@firehose.mindspring.com>, Z Solutions, LLC (zs...@mindspring.com) writes:
>Steve Teixeira <stei...@borland.com> wrote:
>
>
>>While we can't comment on specific features of Delphi32 at
>>this time, we are paying attention to the suggestions put
>>forth in this thread.
>
>> -Steve Teixeira
>> stei...@borland.com
>
What I would love to have is drag and drop in the editor.
Highlight my text, click, hold and drag the block to a new place.
If WordPerfect programmers were able to do this then Borland
should be able to!
---
Bob Mallett
Remember: While the Arc was built by amateurs, it was professionals
who built the Titanic - Richard Needham M.P.
Jonathan W. Hendry
> Every language has its advantages in shortcomings. I believe that Paul would like to
> see Delphi become more C++-like. We can all benefit from studying and using other
> languages and C++ is an excellent choice for a Delphi user try.
I know it's a religious issue, but I don't think Delphi needs
the added complexity. Adding complexity is a bad thing. Ask
anyone (like Taligent) who's working on a C++ project which
is being developed on a geologic time scale. There's lots of
them.
IMHO, I'd like to see Delphi become more Objective-C -like.
(for the umpteenth time I've said so here). It's powerful, great
for object-oriented programming, but without being too complex.
--
Jonathan W. Hendry Steel Driving Software, Inc.
Delphi and NeXTSTEP consulting and software development.
Cincinnati, Ohio
Jonathan W. Hendry
>
>
> In article <47tute$s...@firehose.mindspring.com>, Z Solutions, LLC (zs...@mindspring.com) writes:
> >Steve Teixeira <stei...@borland.com> wrote:
> >
> >
> >>While we can't comment on specific features of Delphi32 at
> >>this time, we are paying attention to the suggestions put
> >>forth in this thread.
> >
> >> -Steve Teixeira
> >> stei...@borland.com
> >
>
> What I would love to have is drag and drop in the editor.
> Highlight my text, click, hold and drag the block to a new place.
> If WordPerfect programmers were able to do this then Borland
> should be able to!
Yuck. I'm forever moving text by accident in Lotus Word Pro,
which implements this. It's a pain, IMHO.
------james------
>In article <8162707...@larouss.demon.co.uk> ca...@larouss.demon.co.uk writes:
>
>>GFA Basic (a very old program on the Atari ST) had a wonderful feature
>>that I haven't seen elsewhere. It had the ability within the editor to
>>'fold' procedures up, so that only the definition was showing. The
>>ability to fold up procedures and function to a single line removes a
>>great deal of clutter from the screen and enables very quick
>>navigation of a large program (or unit).
>
How about having the ability to fold up Begin..End pairs?
Both reduces clutter and helps in debugging.
<----SNIP---->
--
..Jim TIP#587
....jbu...@bates.edu
......http://abacus.bates.edu:80/~jburke/
........For best results, squeeze from the bottom and flatten as you go up.
SPIS & NZ Science Monthly
>GFA Basic (a very old program on the Atari ST) had a wonderful feature
>that I haven't seen elsewhere. It had the ability within the editor to
>'fold' procedures up, so that only the definition was showing. The
>ability to fold up procedures and function to a single line removes a
>great deal of clutter from the screen and enables very quick
>navigation of a large program (or unit).
DAMNED good idea!
>On the same sort of subject I would really appreciate a proper macro
>capability within the editor/IDE. Even a glorified recording facility
>would be useful!
And ditto. I'd forgotten how frustrating I find some aspects of editing in
the IDE. Though I'd need more than just a recording facility, I'm afraid.
--
cheers,
peter
--
==============================================================================
Peter Hyde, South Pacific Information Services Ltd, Christchurch, NZ
Author of TCompress File and Database Compression Components
<<< Humour, science, skeptics and SCA at: http://www.southern.co.nz/spis >>>
JMM
to load and 20 more to print.:) It is nice that Borland bundles ReportSMith
with Delphi, but still does not make sense to bundle THAT report generator as
to the fact that most users of the non clientserver Delphi program on regular
pc's and cannot afford the luxury of sitting for a minute each time they want
to print a report.
JEff
Mark Watts
>There's only one request that I can think off that would make my life
>easier (well actually 2).
>On the same sort of subject I would really appreciate a proper macro
>capability within the editor/IDE. Even a glorified recording facility
>would be useful!
The IDE has a basic recording facility already. Ctrl-Shift-R to start
and stop and Ctrl-Shift-P to replay under both the default and classic
key setups. There is no way to save these however much less have
complex macros. I agree that this should be added.
alexander m nolasco
Dear Borland,
How about adding generics(templates) to Delphi?
Anyone who has worked with Ada would certainly appretiate it..
Leroy Casterline
>> What I would love to have is drag and drop in the editor.
>> Highlight my text, click, hold and drag the block to a new place.
>> If WordPerfect programmers were able to do this then Borland
>> should be able to!
>
>Yuck. I'm forever moving text by accident in Lotus Word Pro,
>which implements this. It's a pain, IMHO.
>--
>
>Jonathan W. Hendry Steel Driving Software, Inc.
>Delphi and NeXTSTEP consulting and software development.
>Cincinnati, Ohio
Agree. If this feature is implemented, it should be an option.
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
= Leroy Casterline ~ Cahill Casterline Ltd. ~ 970/484-2212 ~ cast...@csn.org =
= Digital electronic & firmware design ~ PC-based telephony system design =
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Bengt Richter
>In article <47r27e$q...@sol.zynet.net>, p...@trigon.zynet.co.uk says...
>>
>>Can we have nested comments? {Nested comments {PLEASE!!}}
>>
>Actually, I would be happy simply with an EOL style comment, such
>as the C++ style of //. In fact I am not sure I would like to see
>nested comments, I think using two different comment types is
>better, such as (* { } *) for nested comments. If you only use
>{} comments in your code you can use the (* *) comments to block
>out whole sections.
>Dave
>pi...@cam.org
I think I would like to see a compiler directive that would allow
you to control pre-processing at the comment-detecting level,
although it might have to be limited to prevent too much mischief.
Perhaps only allow new definitions, not re-definitions, which could
be silently ignored. I'm mainly interested in EOL-style comments,
comment beginning and ending, and whitespace, wanting to turn
unused character sequences into any of these. You could specify
strings or characters [singly or sets] in the directive.
For example, {$PREP CommentToEOL '//'} would make // start an
EOL-style comment. {$PREP Whitespace [#0..#255]} would say turn
anything unused into whitespace. You could allow some interesting
pretty-printing by turning single-line box characters into whitespace.
Similarly you could turn left and right edge double line box
characters into begin/end comment tokens. (That does require being
able to use the same character for both, like string quoting).
That would let you embed boxed comments in the code without curls.
I might use
{$PREP CommentToEOL '{{'}
{$PREP CommentBegin [#186,#199..#201,#204,#211,#214]}
{$PREP CommentEnd [#182,#183,#185..#189]}
{$PREP Whitespace [#0..#255]} {everything remaining}
Regards,
Bengt Richter
Bengt Richter
>>From: ca...@larouss.demon.co.uk (Casey Charlton)
>>Newsgroups: comp.lang.pascal.delphi.misc
>>Subject: Re: Delphi 2.0 wish list
>>Date: Mon, 13 Nov 1995 13:48:47 GMT
>>Organization: Not much
>..
>>GFA Basic (a very old program on the Atari ST) had a wonderful feature
>>that I haven't seen elsewhere. It had the ability within the editor to
>>'fold' procedures up, so that only the definition was showing. The
>>ability to fold up procedures and function to a single line removes a
>>great deal of clutter from the screen and enables very quick
>>navigation of a large program (or unit).
>That's a superb idea! Wow, I'm surprised I haven't heard of it before (or
>thought of it before, for that matter ;).
Yes, that definitely gets my vote too.
Actually, I suspect the lisp people may have been there first, and
more generally, since about everything you could browse there has
parentheses whose contents could be collapsed to ... in a browser.
Hm, maybe that is worth thinking about (i.e., collapsing other
program entities besides procedures).
>>On the same sort of subject I would really appreciate a proper macro
>>capability within the editor/IDE. Even a glorified recording facility
>>would be useful!
I still like the M editor I got with Microsoft C 5.1 (or was it 4.x).
Sorry if that offends anyone, but it has a good regular expression
search and search/replace. And it has the ability to cut and paste
two-dimensional chunks of your text page, which is really useful.
Just being able to copy a column of blanks and paste it to move that
many lines to the right (or to the left by cutting columns) all at
once, is nice. I still go back to it when I want to do massive
text surgery. It also would add parentheses or curly brackets for
you, to balance what preceded, while highligting the balancee each
time you hit the key for another. I'd like that again.
If you want editor features, TECO had some nice ones. Particularly
in the ability to to do more than simple cut&paste: You could cut
or copy and have it paste-appended to a named buffer, without having
to switch and paste, so you could just walk through text and snip
and copy everything you needed, then switch to the other buffer
and have everything there to operate on. Another thing was the ability
to split/join lines with a search/replace operation. And the macro
capability was hard to beat. Sufficient to write emacs as a macro,
I believe. You could execute macros from files, have macros create
and execute other macros, and on and on. Probably a little much for
a mainstream ide editor, but if such things could be available without
intruding...
>Yes, the perennial request. (sigh)
>Kendall
Regards,
Bengt Richter
BTW, is the Delphi editor replaceable? I.e., is there a file that
communicates compilation results to direct the editor to the errors
(whoops, 'error' - hope that changes with Delphi32) in a standard way?
[posted to comp.lang.pascal.delphi.misc]
Hakon Baunsgaard
>Something else for the editor: bulk indenting and unindenting.
>That is, the ability to select a block of text and move every
>line, say, 3 spaces left or right.
Try using Ctrl-K I or U. That should do the trick...
Mvh Hakon
Jonathan W. Hendry
> Improvements:
> -------------
> -Ability to write to SELF parameter
Huh?
> -Use of ! to differntiate an object from a property
Ermm. I think that'd create a lot of confusion with C's !.
A naming convention might be better.
> -Allow for better use of external editors. Establish some sort of API for this.
Agreed.
> -Create another string type that acts just like the STRING, but allows length up to 64K (or preferrably a full 32 bits).
(This can be done to a certain degree
This is already in the works, with up to 4 Gigs, I believe.
> -Allow for STATIC variables - Currently you need to declare a global.
>This is not a biggie, but would allow for better structure.
Class variables would be nice. Or else the ability to 'fake'
class variables using static variables (like in ObjC)
> -Option for a runtime dll. (ex: VCL.DLL) I often write an app
>with several EXEs, the current scheme wastes A LOT of memory.
The ability to create shared library DLL's would be a boon.
Borland should let the Delphi runtime know about DLL's, and
how to access classes in them.
> -Ability to "wrap" a class. ie, I should be able to "Wrap"
>TComponent, or "insert" my object between TComponent and any if
>it's ancestors
Something like NeXT Categories would allow something like this.
It would allow programmers to add methods to an existing class,
like TComponent, without recompiling the class. It works by
appending methods to the dynamic method table. The new methods
are inherited by all subclasses of the Categoried class.
(For example, I could write a TimeDate class, and have all
TField classes work with it, by adding two methods.)
Very useful, very simple, very elegant.
Mathew Gyokers Eliot
> If WordPerfect programmers were able to do this then Borland
> should be able to!
If I could just make a little joke: they could do it, but it'd
mean adding 10 megs to the package while using 20% more resources and
make delphi 50% slower if Borland followed the wordperfect way of doing things.
--
Delphi and Paradox Consultant. Member: Borland Delphi Technical Support
Web Page being relocated- want to provide space?
Borland supplied disclaimer: The services I provide via this message
and all consulting services I may provide are in no way associated with
or sponsored by Borland International or any of its subsidiaries.
Harry Culpan
>Actually, if you hold down the ALT key before selecting text you will
>find that you can select a 'box' of text rather than whole lines.
>Select a box 'x' number of lines deep, 3 characters wide, and press
>delete.
>Voila.
True, but how can you block indent? In VB, if you block some code,
you can tab in and out to indent. This was a feature I missed quite a
bit moving to Delphi.
>On 15 Nov 1995 20:32:48 GMT, djwe...@insti.physics.sunysb.edu (David
>Weisberger) wrote:
>>In article <8...@kryton.win-uk.net>,
>>Bob Mallett <bmal...@kryton.win-uk.net> wrote:
>>>What I would love to have is drag and drop in the editor.
>>
>>Something else for the editor: bulk indenting and unindenting.
>>That is, the ability to select a block of text and move every
>>line, say, 3 spaces left or right.
>>
>>-David
>>
>>--
>>David Weisberger
>>djwe...@insti.physics.sunysb.edu
>>
>>
Harry W. Culpan
Eric Begin
using the C++ Language?
They could Name it Athena :)
If any Borland employee read it, could he give me an answer?
-Eric-
Kevin Goldstein
>From: djwe...@insti.physics.sunysb.edu (David Weisberger)
>Subject: Re: Delphi 2.0 wish list
>In article <8...@kryton.win-uk.net>,
>Bob Mallett <bmal...@kryton.win-uk.net> wrote:
>>What I would love to have is drag and drop in the editor.
>Something else for the editor: bulk indenting and unindenting.
>That is, the ability to select a block of text and move every
>line, say, 3 spaces left or right.
>-David
>--
>David Weisberger
>djwe...@insti.physics.sunysb.edu
Yes! I second that idea!
-Kevin
Kendall P. Bullen
"Jonathan W. Hendry" <stee...@ix.netcom.com> wrote:
>Yuck. I'm forever moving text by accident in Lotus Word Pro,
>which implements this. It's a pain, IMHO.
Good word processors and text editors make this an option, not a requirement;
clearly, drag & drop text editing in Delphi would be best implemented as an
option.
Kendall
Casey Charlton
>ca...@larouss.demon.co.uk (Casey Charlton) wrote:
>
>>Actually, if you hold down the ALT key before selecting text you will
>>find that you can select a 'box' of text rather than whole lines.
>>Select a box 'x' number of lines deep, 3 characters wide, and press
>>delete.
>
>>Voila.
>
>True, but how can you block indent? In VB, if you block some code,
>you can tab in and out to indent. This was a feature I missed quite a
>bit moving to Delphi.
Select a block (without ALT) and use CTRL-K I, or CTRL-K U.
Voila!