RE: GtkAda compile errors (Apex Ada)

RE: GtkAda compile errors (Apex Ada)

[Lionel.DRAGHI]

| -----Message d'origine-----
| De: Lionel.DRAGHI@fr.thalesgroup.com
...
| 
| During the preparation of the 2001 Ada-France, to ease the 

of the 2001 Ada-France programming contest

RE: GtkAda compile errors (Apex Ada)

[Lionel.DRAGHI]

| -----Message d'origine-----
| De: Jeff C, [mailto:jcreem@yahoo.com]
...
| I have never heard of anyone trying GtkAda with Apex. 

During the preparation of the 2001 Ada-France, to ease the production of
compiler agnostic software with sexy GUI, we tried to compile GtkAda with
ObjectAda and Apex.

We cleanup the code from GNAT specific attributes (and other stuff I don't
remember), but we stopped this effort due to more serious problem
(compilation errors due to different RM interpretation).

Note that GtkAda team was ready to include patch improving compiler
compatibility.
Rational people (can someone confirm?) told us that the compilation error
was due to an ongoing   AI, and that they will retry later to compile.
For ObjectAda, there where also compilation errors, but I don't remember
what.

Note also that Martin Dowie's project "GNAT for ObjectAda" (if I remember
well, I don't have access to the web here) may also ease the porting.

So, maybe this porting will be easier now (but I won't bet on it).
What is clear for me, is that having a portable among compiler GtkAda would
be great for the Ada community. 

-- 
Lionel Draghi

Re: GtkAda compile errors (Apex Ada)

[Nelson Chandler]

I think Apex may be interpreting the RM differently than GNAT as well.
It's very frustrating.  Thanks for the reply.

-Nelson
<Lionel.DRAGHI@fr.thalesgroup.com> wrote in message
news:mailman.65.1078478969.327.comp.lang.ada@ada-france.org...
>
>
> | -----Message d'origine-----
> | De: Jeff C, [mailto:jcreem@yahoo.com]
> ...
> | I have never heard of anyone trying GtkAda with Apex.
[...]
> We cleanup the code from GNAT specific attributes (and other stuff I don't
> remember), but we stopped this effort due to more serious problem
> (compilation errors due to different RM interpretation).
[...]
>What is clear for me, is that having a portable among compiler GtkAda would
>be great for the Ada community.
>
> -- 
> Lionel Draghi

Re: GtkAda compile errors (Apex Ada)

[Dan Eilers]

Lionel.DRAGHI@fr.thalesgroup.com wrote in message news:<mailman.65.1078478969.327.comp.lang.ada@ada-france.org>...
> 
> During the preparation of the 2001 Ada-France, to ease the production of
> compiler agnostic software with sexy GUI, we tried to compile GtkAda with
> ObjectAda and Apex.
> 
> We cleanup the code from GNAT specific attributes (and other stuff I don't
> remember), but we stopped this effort due to more serious problem
> (compilation errors due to different RM interpretation).
> 
> Note that GtkAda team was ready to include patch improving compiler
> compatibility.

I found that gtk-marshallers.adb has a problem with respect to AI-299,
and reported this two years ago
      http://lists.act-europe.fr/pipermail/gtkada/2002-February/001240.html
but have yet to receive a reply.

Also, gtk-handlers.adb, line 154, has a problem with respect to
AC-014,
where Signal_Id is used for both a parameter name and the function
return type.

And glib-convert.ads has a problem where glib.convert is used both as 
a child package name and the name of a nested function.

	-- Dan Eilers

Re: GtkAda compile errors (Apex Ada)

[Nelson Chandler]

Sorry, the part that says "pragma c" should read "pragma convention (c,
<type>)"

Re: GtkAda compile errors (Apex Ada)

[Jeff C,]

"Nelson Chandler" <nelson.c.chandlerSPAM@FILTERboeing.com> wrote in message
news:Hu0GnL.BMz@news.boeing.com...
> Sorry, the part that says "pragma c" should read "pragma convention (c,
> <type>)"
>
>

Hmm..Neither my newsreader nor google groups shows the original message that
contained the rest of
your message. You might want to try again.

Re: GtkAda compile errors (Apex Ada)

[Nelson Chandler]

"Jeff C," <jcreem@yahoo.com> wrote in message
news:HFt1c.32832$ko6.308078@attbi_s02...
>
> "Nelson Chandler" <nelson.c.chandlerSPAM@FILTERboeing.com> wrote in
message
> news:Hu0GnL.BMz@news.boeing.com...
> > Sorry, the part that says "pragma c" should read "pragma convention (c,
> > <type>)"
> >
> >
>
> Hmm..Neither my newsreader nor google groups shows the original message
that
> contained the rest of
> your message. You might want to try again.
>

Ok, here's the original (corrected):

Hi, I'm using Rational Apex Ada NT 4.2.0c and I'm trying to make GtkAda work
with it.

So far, I've converted all of the gnat style file names to rational apex ada
style file names and made a sub-system and working area for all of the
gtkAda code (2.2.0).  But when I actually try to use it, I keep getting
errors.  Usually it starts with a convention c imported function having one
of its componant types not of convention c.  And that componant type unable
to be used with pragma convention (c, <type>) because
"<type> is (or depends on) an incompletely defined private type" or
"<type> is neither unconstrained nor statically constrained" errors.

Has anyone else had these problems or gotten any of it to work?

-Nelson

Re: GtkAda compile errors (Apex Ada)

[Jeff C,]

"Nelson Chandler" <nelson.c.chandlerNOSPAM@DELETECAPSboeing.com> wrote in
message news:Hu2BCw.9MM@news.boeing.com...
>> Hi, I'm using Rational Apex Ada NT 4.2.0c and I'm trying to make GtkAda
work
> with it.
>
> So far, I've converted all of the gnat style file names to rational apex
ada
> style file names and made a sub-system and working area for all of the
> gtkAda code (2.2.0).  But when I actually try to use it, I keep getting
> errors.  Usually it starts with a convention c imported function having
one
> of its componant types not of convention c.  And that componant type
unable
> to be used with pragma convention (c, <type>) because
> "<type> is (or depends on) an incompletely defined private type" or
> "<type> is neither unconstrained nor statically constrained" errors.
>
> Has anyone else had these problems or gotten any of it to work?
>
> -Nelson
>


I have never tried it. There are a lot of things going against you here. Gtk
on
NT still being somewhat of a second class citizen to the Unix ports (though
it
clearly has gotten a lot better than it used to be)..

I have never heard of anyone trying GtkAda with Apex. What MIGHT work would
be to take all of the GtkAda code and pass it through gcc -gnatR which would
spit
out rep-spec versions for each of the record types. Then you could eliminate
all of the
pragma conventions C things and replace them with rep-specs...Not sure if it
would
would work.


Note that I would be surprised if you got this to work all the way without
some real effort. It would certainly be easier if you used GNAT. It is
probably
not the "standard tool" but then again I suspect GtkAda is not either..