Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.lang.clipper.visual-objects > #4490 > unrolled thread

About VO 2.8 SP4

Started byCarlos Rocha <jazzda1307@gmail.com>
First post2012-12-18 18:14 -0800
Last post2012-12-23 23:34 +0000
Articles 8 — 4 participants

Back to article view | Back to comp.lang.clipper.visual-objects


Contents

  About VO 2.8 SP4 Carlos Rocha <jazzda1307@gmail.com> - 2012-12-18 18:14 -0800
    Re: About VO 2.8 SP4 Carlos Rocha <jazzda1307@gmail.com> - 2012-12-19 10:16 -0800
      Re: About VO 2.8 SP4 James Martin <jmmicmc@cyg.net> - 2012-12-19 11:05 -0800
    Re: About VO 2.8 SP4 Carlos Rocha <carlos.rocha@doossier.com.removethis> - 2012-12-22 19:31 +0000
      Re: About VO 2.8 SP4 Carlos Rocha <carlos.rocha@doossier.com.removethis> - 2012-12-22 22:35 +0000
        Re: About VO 2.8 SP4 Carlos Rocha <jazzda1307@gmail.com> - 2012-12-22 17:42 -0800
    Re: About VO 2.8 SP4 igsofterror@gmail.com - 2012-12-23 13:57 -0800
      Re: About VO 2.8 SP4 Carlos Rocha <carlos.rocha@doossier.com.removethis> - 2012-12-23 23:34 +0000

#4490 — About VO 2.8 SP4

FromCarlos Rocha <jazzda1307@gmail.com>
Date2012-12-18 18:14 -0800
SubjectAbout VO 2.8 SP4
Message-ID<143c7a9e-df41-48ef-b63a-758ebbb59f46@googlegroups.com>
Hi all,

I'm finishing my last project with SP3 and am now almost ready for SP4.

But before I make the move I would like to know what exactly will I win, specially regarding to COM dealing, like:

1. Support for ByteArrays. With SP3 we have the option to receive them as Strings, but it's not possible to send them back as ByteArrays.

2. Support for Variants passed by reference in events. Actually they are received as Usuals and therefore unusable.

Of course other matters interest me too, like stability, Debug...

I've read the "WhatsNew" paper. I don't use most of the changed stuff, but there are some interesting new things in Tools.

Finally...what new bugs were introduced in SP4, if any?

Thanks

[toc] | [next] | [standalone]


#4492

FromCarlos Rocha <jazzda1307@gmail.com>
Date2012-12-19 10:16 -0800
Message-ID<0f597054-c624-4845-a2ce-d02245d6610e@googlegroups.com>
In reply to#4490
I see...
I bought it anyway

[toc] | [prev] | [next] | [standalone]


#4493

FromJames Martin <jmmicmc@cyg.net>
Date2012-12-19 11:05 -0800
Message-ID<af93e082-32a3-4a97-be2e-55921545f53e@googlegroups.com>
In reply to#4492
I am interested in your evaluation. I am considering it too, even though I don't need any of the new features. I just want to support the platform and get the best product.

James

[toc] | [prev] | [next] | [standalone]


#4511

FromCarlos Rocha <carlos.rocha@doossier.com.removethis>
Date2012-12-22 19:31 +0000
Message-ID<kb51sa$q68$1@speranza.aioe.org>
In reply to#4490
First impressions:

>
> 1. Support for ByteArrays. With SP3 we have the option to receive them as Strings, but it's not possible to send them back as ByteArrays.
> 2. Support for Variants passed by reference in events. Actually they are received as Usuals and therefore unusable.

No. Maybe someday some user "asks" for it :/

>
> Of course other matters interest me too, like stability, Debug...

Seems better overall, but I still get the "Sorry..." button if I use the 
debug.

With the new pair OLE runtime/OleAutoObjectEx I found an old obscure bug 
in my code:

When some Access/Method that returns another OleAutoObject doesn't 
return the "real" object I usually cast it like

local oRealObject as TheRealObject
oRealObject := TheRealObject{MyGeneratedClass:SomeAccess}
oRealObject:Color := Rgb(R,G,B)

This way I can use the specific properties of TheRealObject that might 
not exist in the original (generated) uRetVale, and is also useful to 
call declared methods for early binding.

I was casting an object to a wrong one by mistake (since VO 2.7 ?), but 
with SP4 the runtime detected it (with an exception, of course).

> Finally...what new bugs were introduced in SP4, if any?
>

The "Automation Server..." generated repeated code for different objects 
in a few occasions. It's easy to detect, and deleting and generating 
again solves it. Never happened to me with previous versions.

The compiler is really much faster, but sometimes one need to touch some 
modules/applications by hand or the binary (.DBG) isn't updated accordingly.

I tried to correct the introduced bug in RGB() - now returns a Dword, 
like nobody else :/

_dll FUNCTION RgbAsDWord(bR, bG, bB as byte) as DWORD PASCAL:VO28RUN.Rgb

function RGB(bR, bG, bB as byte) as long pascal
return long(_cast, RgbAsDWord(bR, bG, bB))

but I get a runtime message stating that "the entry point to RGB was not 
found in VO28RUN" if running from the IDE, but the EXE runs fine.

In conclusion, overall SP4 is better than SP3 and my program apparently 
runs faster and is smaller (gotta check why - 18.9 MB to 15.6 MB seems 
too much).

-- 
Carlos Rocha

[toc] | [prev] | [next] | [standalone]


#4512

FromCarlos Rocha <carlos.rocha@doossier.com.removethis>
Date2012-12-22 22:35 +0000
Message-ID<kb5cli$l48$1@speranza.aioe.org>
In reply to#4511
>
> _dll FUNCTION RgbAsDWord(bR, bG, bB as byte) as DWORD PASCAL:VO28RUN.Rgb
>
> function RGB(bR, bG, bB as byte) as long pascal
> return long(_cast, RgbAsDWord(bR, bG, bB))
>
> but I get a runtime message stating that "the entry point to RGB was not
> found in VO28RUN" if running from the IDE, but the EXE runs fine.
>

Now it works also in IDE. Go figure...
It seems related to the faulty auto-touching process. Any thoughts?

-- 
Carlos Rocha

[toc] | [prev] | [next] | [standalone]


#4513

FromCarlos Rocha <jazzda1307@gmail.com>
Date2012-12-22 17:42 -0800
Message-ID<85a88c44-d4e7-4e24-bd9d-016b26e71c87@googlegroups.com>
In reply to#4512
On Saturday, December 22, 2012 10:35:19 PM UTC, Carlos Rocha wrote:
> >
> 
> > _dll FUNCTION RgbAsDWord(bR, bG, bB as byte) as DWORD PASCAL:VO28RUN.Rgb
> 
> >
> 
> > function RGB(bR, bG, bB as byte) as long pascal
> 
> > return long(_cast, RgbAsDWord(bR, bG, bB))
> 
> >
> 
> > but I get a runtime message stating that "the entry point to RGB was not
> 
> > found in VO28RUN" if running from the IDE, but the EXE runs fine.
> 
> >
> 
> 
> 
> Now it works also in IDE. Go figure...
> 
> It seems related to the faulty auto-touching process. Any thoughts?
> 

Yes it is a bug, but maybe not in the auto-touching. The reason is that the .DBG file is not updated, but if I delete it by hand a new one is created and works as expected, with no "touching" (visually speaking, but maybe it happens internally).

I found that if the changed library is included in the "Default App" the .DBG is updated. This makes no sense and SP3 doesn't need it.

Is there a solution?

-- 
Carlos Rocha

[toc] | [prev] | [next] | [standalone]


#4514

Fromigsofterror@gmail.com
Date2012-12-23 13:57 -0800
Message-ID<fee7939c-3c3d-4c1f-bb37-c5eb0e396385@googlegroups.com>
In reply to#4490
Hi Carlos,

I have tried it. I get a lot of warning error that weren't present before:
- any method without return cause a warning
- all conversion are now strongly typed: 
- for next => always dword
- cannot declare d as int and the in the code d := day(mydate) because day() return dword

I can turn off the warning and the app runs, but I prefer to correct all before...We don't take care of that before, but now we must...

Jean-Luc

[toc] | [prev] | [next] | [standalone]


#4515

FromCarlos Rocha <carlos.rocha@doossier.com.removethis>
Date2012-12-23 23:34 +0000
Message-ID<kb84gq$nsv$1@speranza.aioe.org>
In reply to#4514
Jean-Luc,

>
> I have tried it. I get a lot of warning error that weren't present before:
> - any method without return cause a warning

I see no problem with that, just add "return nil"

> - all conversion are now strongly typed:
> - for next => always dword
> - cannot declare d as int and the in the code d := day(mydate) because day() return dword

Well, that's easy too. Just declare the variables as dword

>
> I can turn off the warning and the app runs, but I prefer to correct all before...We don't take care of that before, but now we must...
>

Absolutely, and maybe there's a good reason for that, dunno...
But Rgb returning dword... it's sacrilege in Win32 world, imo (the 
workaround I posted just avoids changing lots of code lines, but the 
same could easily have been done do change it in the other way, without 
spending developers time. I don't get it... )

I believe that the real reason for all this is that VO has been 
progressively adapted to be as much compatible with Vulcan as possible 
(I'm guessing here) over Win32, where it really belongs, and most recent 
changes were required (and tested only) by Vulcan users (please don't 
crucify me this time, it's Christmas!!)

VO is a little better with SP4, no doubts about it, but it lost ground 
in recent years (since Vulcan/VOPS) if compared with other Win32 
languages (like Delphi...), specially in the ability to use COM, and COM 
is needed even to use .NET stuff. I need to wrap almost all 3rd party 
components in VB6 because of that unfinished 0.5%. Sad, to say the least :/

-- 
Carlos Rocha

[toc] | [prev] | [standalone]


Back to top | Article view | comp.lang.clipper.visual-objects


csiph-web