Xojo Conferences
XDCMay2019MiamiUSA

OSX vs carbon/classic (Real Studio Plugins Mailinglist archive)

Back to the thread list
Previous thread: Frameworks 'n Stuff
Next thread: New FolderItem("ddd") or REALFSRefFromFolderItem return bogus in OSX


Re: class extensions...   -   Jan Erik Moström <
  OSX vs carbon/classic   -   Alfred Van Hoek
   Re: OSX vs carbon/classic   -   Kevin Ballard
    Re: OSX vs carbon/classic   -   Joseph J. Strout
    Re: OSX vs carbon/classic   -   Mars Saxman
   Re: OSX vs carbon/classic   -   Alfred Van Hoek
   Re: OSX vs carbon/classic   -   Alfred Van Hoek

OSX vs carbon/classic
Date: 10.09.02 16:21 (Tue, 10 Sep 2002 11:21:59 -0400)
From: Alfred Van Hoek
I seem to be on a track without an end. This track is the discovery of bugs
that only appear in RB-OSX, but not in RBCarbon9 or RBClassic. It is
important to let you all know so you can start chasing other potential
discrepancies (in the plugin SDK and in REALbasic).

The latest:

A plugin registers a class with 2 constructors:

class:
SampleClass

constructors:
SampleClass(f as folderItem)
SampleClass(s as string)

In OSX, RB complains with a type-mismatch.

I have not checked whether a sample class defined in RB with 2 constructors
does the same thing. I cannot wait to report, in case REAL wants to release
4.5.1 tomorrow.

(Other differences between RBOSX and RB9 are: you cannot create a
resourcefork or a binary stream, and REALFSRefFromFolderItem does not give
an appropriate result when returning a parentFSRef and a name of the to be
created file)

RS: a sample plugin is in the make (the same one I sent yesterday, but with
the addition of a second constructor), and believe me, the bug exists.

Alfred

---
A searchable archive of this list is available at:
<http://dbserver.realsoftware.com/KBDB/search.php>

Unsubscribe:
<mailto:<email address removed>>

Subscribe to the digest:
<mailto:<email address removed>>

Re: OSX vs carbon/classic
Date: 10.09.02 22:37 (Tue, 10 Sep 2002 17:37:57 -0400)
From: Kevin Ballard
I have a modification to this report:

If the constructors have different numbers of arguments, it works fine.

I have a class with multiple classes, 2 of which have constructors,
such as

constructors
CFString()
CFString(s As String)

And it works fine.

Hrm, after thinking about this, I believe this is a known bug, that in
constructors only, constructors require different numbers of parameters
to work. Hopefully this is fixed in the new compiler in v5, because I
think the problem was it wasn't easy to fix with the current compiler.

On Tuesday, September 10, 2002, at 11:21 AM, Alfred Van Hoek wrote:

> The latest:
>
> A plugin registers a class with 2 constructors:
>
> class:
> SampleClass
>
> constructors:
> SampleClass(f as folderItem)
> SampleClass(s as string)
>
> In OSX, RB complains with a type-mismatch.

Re: OSX vs carbon/classic
Date: 10.09.02 22:41 (Tue, 10 Sep 2002 14:41:53 -0700)
From: Joseph J. Strout
At 5:37 PM -0400 9/10/02, Kevin Ballard wrote:

>Hrm, after thinking about this, I believe this is a known bug, that
>in constructors only, constructors require different numbers of
>parameters to work. Hopefully this is fixed in the new compiler in
>v5, because I think the problem was it wasn't easy to fix with the
>current compiler.

Yes, that's all exactly right. (And feel free to test it in RBScript
if you want to prove to yourself that the new compiler handles it
correctly.)

Cheers,
- Joe

Re: OSX vs carbon/classic
Date: 10.09.02 22:49 (Tue, 10 Sep 2002 14:49:08 -0700)
From: Mars Saxman
kevin@sb.org wrote:

> I have a modification to this report:
>
> If the constructors have different numbers of arguments, it works fine.

Which version of REALbasic are you using? I worked on on this problem in 4.5
and have not heard any reports of its recurrence.

Mars Saxman
REAL Software

---
A searchable archive of this list is available at:
<http://dbserver.realsoftware.com/KBDB/search.php>

Unsubscribe:
<mailto:<email address removed>>

Subscribe to the digest:
<mailto:<email address removed>>

Re: OSX vs carbon/classic
Date: 10.09.02 16:58 (Tue, 10 Sep 2002 11:58:03 -0400)
From: Alfred Van Hoek
on 9/10/02 11:55 AM, Alfred Van Hoek at <email address removed> wrote:

> on 9/10/02 11:21 AM, Alfred Van Hoek at <email address removed> wrote:
>
>> I have not checked whether a sample class defined in RB with 2 constructors
>> does the same thing. I cannot wait to report, in case REAL wants to release
>> 4.5.1 tomorrow.
>
> The same holds when done in pure RB code,
>
> Joe, enclosed sample project and plugin with source...
>
> PS: if this will not be fixed (as well as the other issues) before 4.5.1 goes
> final I can simply forget to provide plugins for the OSX community; The
> plugin classes have overloading constructors, depend on the creation of files,
> and all of this simply does not work with REALbasic running in OSX..
>
> Best,
>
> Alfred
>

---
A searchable archive of this list is available at:
<http://dbserver.realsoftware.com/KBDB/search.php>

Unsubscribe:
<mailto:<email address removed>>

Subscribe to the digest:
<mailto:<email address removed>>

Re: OSX vs carbon/classic
Date: 11.09.02 15:59 (Wed, 11 Sep 2002 10:59:10 -0400)
From: Alfred Van Hoek
> Subject: OSX vs carbon/classic
> From: "Alfred Van Hoek" <<email address removed>>
> Date: Tue, 10 Sep 2002 11:21:59 -0400

I wrote:
>
> I seem to be on a track without an end. This track is the discovery of bugs
> that only appear in RB-OSX, but not in RBCarbon9 or RBClassic. It is
> important to let you all know so you can start chasing other potential
> discrepancies (in the plugin SDK and in REALbasic).

False Alarm. This happened because of a 5 day struggle to track down bugs
that really appeared one after another, by not paying attention that
4.5.1fc4 was not running, but 4.5.1fc2. Thus, overloading works, creation of
folderItems work, and functions in the plugin SDK work. Yes, I was
panicking...

Fortunately, David and Joe were very helpful to help me tracking the cause
of the problem(s).

Alfred

PS: my bottom line is to get rid of older versions of REALbasic in the OSX
environment; it does not serve me well, while wasting your valuable time.

---
A searchable archive of this list is available at:
<http://dbserver.realsoftware.com/KBDB/search.php>

Unsubscribe:
<mailto:<email address removed>>

Subscribe to the digest:
<mailto:<email address removed>>