Ian Joyner | 1 May 11:08 2010
Picon

Re: Third party development platforms

Excellent. I'm glad someone is using the approach with the Obj-C runtime. I've thought myself that would be
the way to do it, but lack time and motivation.

I still think to get Mac developers involved, the compiler has to be integrated with Xcode, which might
compel developers to look at EiffelStudio. However, Apple seems to be removing any desire for developers
to look at anything except Objective-C, which as I've said is a curious language (although I like it better
than C++ - but that wouldn't be hard would it!).

Ian

On 30 Apr 2010, at 23:43, Colin LeMahieu wrote:

> I'm working on extending the Objective C and Cocoa library started by Daniel
> Furrer and Manu. I'm taking a slightly different approach and doing the
> interop from ANSI C using the Objective C runtime instead of trying to
> integrate with Objective C without compiler support. I'm also tying in the
> eiffel trampoline library I wrote to do direct callbacks from Objective C in
> to Eiffel.
> 
> Right now I have some basic native Cocoa widgets working. Working on
> getting the Vision2 demo going and then I'd put the library out for anyone
> to use.
> 
> I was also disappointed to see the limitation for 3rd party languages. It
> seems this limitation would only really matter for Appstore applications;
> they don't have control over applications you put directly on the iPhone.
> 
> On Fri, Apr 30, 2010 at 6:50 AM, Ian Joyner <i.joyner@...> wrote:
> 
> > Cocoa is very nice, but large, so is not going to be perfect. But I am
(Continue reading)

Kevin McFarlane | 1 May 23:47 2010
Picon

Re: Third party development platforms

Ian, 

FWIW, the MonoTouch guys think that their toolset *seems* to comply with the new license - no confirmation
from Apple but some MonoTouch apps. have been approved in the App Store since the license change.

http://monotouch.net/

If it does then does the Eiffel support work in a similar manner?

Kevin

--- In eiffel_software@..., Ian Joyner <i.joyner <at> ...> wrote:
>
> You all might have seen Steve Jobs' broadside against Adobe and Flash today:
> 
> http://www.appleinsider.com/articles/10/04/29/apples_steve_jobs_publishes_public_thoughts_on_flash_letter.html
> 
> This might sound like specific Apple news and not of interest to an Eiffel forum, except it very much
affects third-party cross-platform development.
> 
> There are of course rebuttals:
> 
> http://www.osnews.com/story/23224/
> http://www.smh.com.au/technology/technology-news/adobe-hits-back-at-apple-let-the-people-decide-20100430-ty6h.html
> 
> although Narayan's line "We're saying let consumers pick. If this technology doesn't work, consumers
won't pick it." is somewhat laughable.
> 
> Anyway, this thinking seems to very much affect Eiffel's prospects and probably shows why we have never
got anywhere much getting Eiffel on the Mac. I'd very much like something better than Objective-C and
(Continue reading)

Guus Leeuw jr. | 3 May 23:38 2010

Vision2 & void safety

Hello,

Building a GUI app, having a class ABOUT_DIALOG as created by the wizard:

note

       description   : "About dialog box"

       author        : "Generated by the New Vision2 Application Wizard."

       date          : "$Date: 2010/5/2 9:45:18 $"

       revision      : "1.0.0"

class

       ABOUT_DIALOG

inherit

       EV_DIALOG

              redefine

                     initialize

              end

       INTERFACE_NAMES

(Continue reading)

Eric Bezault | 3 May 23:47 2010

Re: Vision2 & void safety

Guus Leeuw jr. wrote:
> How can we amend this, given that I cannot redefine default_create, as it is
> frozen in EV_ANY. Also, then, I would have to call Precursor, which should
> lead to the same error, really.

You might try to create your attributes before the call to
precursor in `initialize'.

--

-- 
Eric Bezault
mailto:ericb@...
http://www.gobosoft.com

------------------------------------

Guus Leeuw jr. | 3 May 23:52 2010

RE: Vision2 & void safety

The problem is, Eric, `initialize' is called from `default_create', which is
frozen in EV_ANY.

The generated MAIN_WINDOW class simply calls `create about_dialog', so it
will use the `default_create'.

The one thing that would work is if I were to utilize "On demand
void-safety", which is not what is going to be default for ec, I think.

Regards,

Guus

From: eiffel_software@...
[mailto:eiffel_software@...] On Behalf Of Eric Bezault
Sent: 03 May 2010 22:48
To: eiffel_software@...
Subject: Re: [eiffel_software] Vision2 & void safety

Guus Leeuw jr. wrote:
> How can we amend this, given that I cannot redefine default_create, as it
is
> frozen in EV_ANY. Also, then, I would have to call Precursor, which should
> lead to the same error, really.

You might try to create your attributes before the call to
precursor in `initialize'.

--

-- 
Eric Bezault
(Continue reading)

Ian King | 3 May 23:54 2010

Re: Vision2 & void safety

Hi Guus,

If you look at this page here:

http://docs.eiffel.com/book/solutions/converting-eiffelvision-2-systems-void-safety

<http://docs.eiffel.com/book/solutions/converting-eiffelvision-2-systems-void-safety>You
can see that you can redefine create_interface_objects to create the
attached types, and then use initialize to initialize them correctly for the
current widget.

Hope this helps

Ian King
Eiffel Software

On Mon, May 3, 2010 at 2:38 PM, Guus Leeuw jr. <guus.leeuw@...>wrote:

>
>
> Hello,
>
> Building a GUI app, having a class ABOUT_DIALOG as created by the wizard:
>
> note
>
> description : "About dialog box"
>
> author : "Generated by the New Vision2 Application Wizard."
>
(Continue reading)

Guus Leeuw jr. | 4 May 01:10 2010

RE: Vision2 & void safety

Ian,

Thanks for that, Ian.

Now we just need to make sure that the Vision2 Wizard pops out a correct
void-safe class to start with :-D

Regards,
Guus

> -----Original Message-----
> From: eiffel_software@...
> [mailto:eiffel_software@...] On Behalf Of Ian King
> Sent: 03 May 2010 22:54
> To: eiffel_software@...
> Subject: Re: [eiffel_software] Vision2 & void safety
> 
> Hi Guus,
> 
> If you look at this page here:
> 
> http://docs.eiffel.com/book/solutions/converting-eiffelvision-2-
> systems-void-safety
> 
> 
> <http://docs.eiffel.com/book/solutions/converting-eiffelvision-2-
> systems-void-safety>You
> can see that you can redefine create_interface_objects to create the
> attached types, and then use initialize to initialize them correctly
> for the
(Continue reading)

Guus Leeuw jr. | 4 May 08:58 2010

Time Library patch

Hello,

Working with the time library, I figured out today that the default
*English* time format for DATE_TIME_TOOLS is actually en-US
([0]mm/[0]dd/yyyy).

Hence, I created a patch against the time library that will:

.         Include en-US en en-GB as new formats

.         Obsolete DATE_TIME_TOOLS from the "english" directory

.         Still include "english" in time.ecf and time-safe.ecf (so as to
not break old systems)

.         Correctly (de)select en-US / en-GB in various *.ecf files

Patch will be submitted to Jocelyn & Manu for inclusion in SVN...

What we should think about is including the i18n work into EiffelTime, so
that we can rid the library of the various format/*language* directories and
make a more intuitive solution to the internationalization problem.

Regards,

Guus

[Non-text portions of this message have been removed]

------------------------------------
(Continue reading)

Guus Leeuw jr. | 4 May 09:05 2010

RE: [eiffelstudio] Commit - Revision 83053

Hello,

From the commit message from fivaa:
> +/* Constants for describing the type of event, using the first 3 bits 
> +of a byte */
> +
> +#define TYPE_MASK 0xF0
> +
> +#define OUTCALL 0x00	/* A C routine is called from Eiffel */
> +#define OUTRET  0x10	/* A C routine returns to Eiffel */
> +#define OUTEXC  0x20	/* An exception occured in C and gets propagated back to Eiffel */
> +
> +#define INCALL  0x30	/* An Eiffel routine is called from the C side */
> +#define INRET   0x40	/* An Eiffel routine returns to C */
> +
> +#define NEWOBJ	0x50	/* A Eiffel object is created in C */
> +#define NEWSPL	0x60
> +#define NEWTPL	0x70
> +#define NEWBIT	0x80
> +
> +#define MEMMUT  0x90    /* Memory area mutation */
> +
> +/* The other 5 bits can be used for argument count etc. */
> +
> +#define ARG_MASK 0x0F

This smells to me as though you would 4 bits for describing the event and 4 bits for argument count. Maybe the
comments are wrong / old?

Regards,
(Continue reading)

Jocelyn Fiat [ES] | 4 May 09:26 2010

Re: RE: [eiffelstudio] Commit - Revision 83053

I think you should post such message on the http://forum.eiffel.com/13
(i.e: http://eiffelstudio.origo.ethz.ch/forum/13)
which is more focused on EiffelStudio's development.

Otherwise, I let Arno (the author of the commit) answer your comment

Regards,
-- Jocelyn

On 5/4/2010 9:05 AM, Guus Leeuw jr. wrote:
> Hello,
>
> >From the commit message from fivaa:
>   
>> +/* Constants for describing the type of event, using the first 3 bits 
>> +of a byte */
>> +
>> +#define TYPE_MASK 0xF0
>> +
>> +#define OUTCALL 0x00	/* A C routine is called from Eiffel */
>> +#define OUTRET  0x10	/* A C routine returns to Eiffel */
>> +#define OUTEXC  0x20	/* An exception occured in C and gets propagated back to Eiffel */
>> +
>> +#define INCALL  0x30	/* An Eiffel routine is called from the C side */
>> +#define INRET   0x40	/* An Eiffel routine returns to C */
>> +
>> +#define NEWOBJ	0x50	/* A Eiffel object is created in C */
>> +#define NEWSPL	0x60
>> +#define NEWTPL	0x70
>> +#define NEWBIT	0x80
(Continue reading)


Gmane