AdaCore Community 2018 Bugs

AdaCore Community 2018 Bugs

[Roger]

I'm often getting this:
+===========================GNAT BUG DETECTED==============================+
| Community 2018 (20180523-73) (x86_64-apple-darwin16.7.0) GCC error:      |
| in gnat_to_gnu_entity, at ada/gcc-interface/decl.c:2985                  |
| Error detected at gl.ads:27:53                                           |
| Please submit a bug report by email to report@adacore.com.               |

I usually get around it by rearranging code or finding just where in my code the bug is getting generated and working on something else for a while.
The bug request asks me to provide a large number of files which I'm not happy to do as in the past of had little success in submitting such bug reports.
Any recommendations.
Perhaps use a different Ada compiler?

Re: AdaCore Community 2018 Bugs

[Shark8]

On Friday, August 10, 2018 at 7:06:52 PM UTC-6, Roger wrote:
> I'm often getting this:
> +===========================GNAT BUG DETECTED==============================+
> | Community 2018 (20180523-73) (x86_64-apple-darwin16.7.0) GCC error:      |
> | in gnat_to_gnu_entity, at ada/gcc-interface/decl.c:2985                  |
> | Error detected at gl.ads:27:53                                           |
> | Please submit a bug report by email to reportadacore.com.               |
> 
> I usually get around it by rearranging code or finding just where in my code the bug is getting generated and working on something else for a while.
> The bug request asks me to provide a large number of files which I'm not happy to do as in the past of had little success in submitting such bug reports.
> Any recommendations.
> Perhaps use a different Ada compiler?


If you're trying to use Ada 2012 there's no other compiler extant. If 2005 there's more options, but none are free TTBOMK.

Randy's Janus/Ada is Ada 95 w/ bits of succeeding standards, but it's the only one that I know of that's in the "hobbyist" price-range.

My own Byron, while currently stalled, is an attempt to write an Ada 2012 in Ada 2012.

Re: AdaCore Community 2018 Bugs

[Roger]

On Saturday, August 11, 2018 at 1:11:31 PM UTC+10, Shark8 wrote:
> On Friday, August 10, 2018 at 7:06:52 PM UTC-6, Roger wrote:
> > I'm often getting this:
> > +===========================GNAT BUG DETECTED==============================+
> > | Community 2018 (20180523-73) (x86_64-apple-darwin16.7.0) GCC error:      |
> > | in gnat_to_gnu_entity, at ada/gcc-interface/decl.c:2985                  |
> > | Error detected at gl.ads:27:53                                           |
> > | Please submit a bug report by email to reportadacore.com.               |
> > 
> > I usually get around it by rearranging code or finding just where in my code the bug is getting generated and working on something else for a while.
> > The bug request asks me to provide a large number of files which I'm not happy to do as in the past of had little success in submitting such bug reports.
> > Any recommendations.
> > Perhaps use a different Ada compiler?
> 
> 
> If you're trying to use Ada 2012 there's no other compiler extant. If 2005 there's more options, but none are free TTBOMK.
> 
> Randy's Janus/Ada is Ada 95 w/ bits of succeeding standards, but it's the only one that I know of that's in the "hobbyist" price-range.
> 
> My own Byron, while currently stalled, is an attempt to write an Ada 2012 in Ada 2012.

I enquired about Randy's Janus/Ada  and have been advised that there is no OSX version and that there is no home-use/hobbyist  pricing. Only qualified student pricing.

Re: AdaCore Community 2018 Bugs

[Jacob Sparre Andersen]

Roger wrote:

> I enquired about Randy's Janus/Ada and have been advised that there is
> no OSX version and that there is no home-use/hobbyist pricing. Only
> qualified student pricing.

If I remember correctly, the "professional" pricing of Janus/Ada matches
the price I had my dad pay for a Turbo Pascal compiler back in 1987
(assuming an inflation around 3.5%/p.a.), so you will not hear me
complain about the pricing of Janus/Ada.

... but I'm a member of the choir complaining about the absence of a
version of Janus/Ada for Linux.

Greetings,

Jacob
-- 
"The pictures on radio are always so much better than those on TV."
                                       -- Pratchet, Stewart & Cohen

Re: AdaCore Community 2018 Bugs

[Roger]

On Tuesday, August 21, 2018 at 4:06:21 PM UTC+10, Jacob Sparre Andersen wrote:
> Roger wrote:
> 
> > I enquired about Randy's Janus/Ada and have been advised that there is
> > no OSX version and that there is no home-use/hobbyist pricing. Only
> > qualified student pricing.
> 
> If I remember correctly, the "professional" pricing of Janus/Ada matches
> the price I had my dad pay for a Turbo Pascal compiler back in 1987
> (assuming an inflation around 3.5%/p.a.), so you will not hear me
> complain about the pricing of Janus/Ada.
> 
> ... but I'm a member of the choir complaining about the absence of a
> version of Janus/Ada for Linux.
> 

> Greetings,
> 
> Jacob


I just checked the prices.
Janus/Ada 95 Personal Edition for Microsoft Windows is certainly OK at $US195.
I'm surprised that Randy didn't associate "home-use" with "Personal Edition".
I wonder if it can be run under CrossOver? If so, I might try it.
Regards,Roger

Re: AdaCore Community 2018 Bugs

[Randy Brukardt]

"Roger" <rogermcm2@gmail.com> wrote in message 
news:343df459-0ac5-475a-9824-66d5b4863b31@googlegroups.com...
...
> I just checked the prices.
> Janus/Ada 95 Personal Edition for Microsoft Windows is certainly OK at 
> $US195.
> I'm surprised that Randy didn't associate "home-use" with "Personal 
> Edition".

There's no special price for "home use"; the personal edition is available 
to anyone for any purpose. (If IBM wants to purchase it, they're welcome.) 
And we sometimes get requests from people who think that just because it's 
home use, they should get a discount to near nothing -- which we don't do, 
since the price is reasonable and giving some people discounts wouldn't be 
fair to those who pay the full price.

I had assumed you'd looked at the price lists before asking, since they're 
on-line and I sent you a link in case you hadn't.

                                        Randy.

Re: AdaCore Community 2018 Bugs

[Roger]

On Wednesday, August 22, 2018 at 7:24:03 AM UTC+10, Randy Brukardt wrote:
> "Roger" <rogermcm2@gmail.com> wrote in message 
> news:343df459-0ac5-475a-9824-66d5b4863b31@googlegroups.com...
> ...
> > I just checked the prices.
> > Janus/Ada 95 Personal Edition for Microsoft Windows is certainly OK at 
> > $US195.
> > I'm surprised that Randy didn't associate "home-use" with "Personal 
> > Edition".
> 
> There's no special price for "home use"; the personal edition is available 
> to anyone for any purpose. (If IBM wants to purchase it, they're welcome.) 
> And we sometimes get requests from people who think that just because it's 
> home use, they should get a discount to near nothing -- which we don't do, 
> since the price is reasonable and giving some people discounts wouldn't be 
> fair to those who pay the full price.
> 
> I had assumed you'd looked at the price lists before asking, since they're 
> on-line and I sent you a link in case you hadn't.
> 
>                                         Randy.

Home Use is a fairly common synonym for  Personal Edition.

For example, I have Home Use licenses for Mathematica and Microsoft Office.

Re: AdaCore Community 2018 Bugs

[Dennis Lee Bieber]

On Tue, 21 Aug 2018 17:51:13 -0700 (PDT), Roger <rogermcm2@gmail.com>
declaimed the following:

>For example, I have Home Use licenses for Mathematica and Microsoft Office.

	In the case of Mathematica, it might be cheaper to just by a
Raspberry-Pi -- Mathematica/Wolfram have been included in the "Raspbian"
distribution. http://www.wolfram.com/raspberry-pi/

	As for M$ Office... Well... The "home use" license I have was
predicated upon having the same version used at my employer (of the time --
I am now in violation of license as I'm no longer employed <G>).


-- 
	Wulfraed                 Dennis Lee Bieber         AF6VN
	wlfraed@ix.netcom.com    HTTP://wlfraed.home.netcom.com/ 

Re: AdaCore Community 2018 Bugs

[Per Sandberg]

Roger

The code base I am usually working with is not intended for the internet.

So what I do when I get a dump like this in the closed code-base is:
Copy  all required files in one folder then i put 60 minutes or so to 
reduce the code to a bare minimum that reproduces the problem, that way 
i could rewrite the offending code by hand on a computer with internet 
access and publish the problem.

And when working on code intended for the internet i got no problems 
what so ever to send the complete set of files (however, some small 
effort usually reduces the amount quite a bit) and i find it decent to 
do that initial work before i send a report.

And in my opinion to report problems on "open software" with good 
quality reports is for certain within the price-range of a hobbyist and 
everyone benefits from it.

/P


On 08/11/18 03:06, Roger wrote:
> I'm often getting this:
> +===========================GNAT BUG DETECTED==============================+
> | Community 2018 (20180523-73) (x86_64-apple-darwin16.7.0) GCC error:      |
> | in gnat_to_gnu_entity, at ada/gcc-interface/decl.c:2985                  |
> | Error detected at gl.ads:27:53                                           |
> | Please submit a bug report by email to report@adacore.com.               |
> 
> I usually get around it by rearranging code or finding just where in my code the bug is getting generated and working on something else for a while.
> The bug request asks me to provide a large number of files which I'm not happy to do as in the past of had little success in submitting such bug reports.
> Any recommendations.
> Perhaps use a different Ada compiler?
> 

Re: AdaCore Community 2018 Bugs

[olivermkellogg]

On Saturday, August 11, 2018 at 7:15:12 AM UTC+2, Per Sandberg wrote:
> Roger
> 
> The code base I am usually working with is not intended for the internet.
> [...]

Similar for me, and for cases where narrowing down is difficult or even impossible I've been using an Ada code obfuscator [1].

- Oliver

[1] (http://okellogg.de/obfuscada.pl

Re: AdaCore Community 2018 Bugs

[Jeffrey R. Carter]

On 08/11/2018 03:06 AM, Roger wrote:
> Perhaps use a different Ada compiler?

Have you tried the FSF GNAT 8.1.0 compiler?

https://sourceforge.net/projects/gnuada/files/GNAT_GCC%20Mac%20OS%20X/8.1.0/native-2017/

-- 
Jeff Carter
"I would never want to belong to any club that
would have someone like me for a member."
Annie Hall
41

Re: AdaCore Community 2018 Bugs

[Roger]

On Saturday, August 11, 2018 at 6:42:27 PM UTC+10, Jeffrey R. Carter wrote:
> On 08/11/2018 03:06 AM, Roger wrote:
> > Perhaps use a different Ada compiler?
> 
> Have you tried the FSF GNAT 8.1.0 compiler?
> 
> https://sourceforge.net/projects/gnuada/files/GNAT_GCC%20Mac%20OS%20X/8.1.0/native-2017/
> 
> -- 
> Jeff Carter
> "I would never want to belong to any club that
> would have someone like me for a member."
> Annie Hall
> 41

Yes, I tried it and obtained the same result:
8.1.0 (x86_64-apple-darwin15) Constraint_Error erroneous memory access 
 so I've submitted a bug report to  https://gcc.gnu.org/bugs/ .
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86916

Re: AdaCore Community 2018 Bugs

[Roger]

On Saturday, August 11, 2018 at 6:42:27 PM UTC+10, Jeffrey R. Carter wrote:
> On 08/11/2018 03:06 AM, Roger wrote:
> > Perhaps use a different Ada compiler?
> 
> Have you tried the FSF GNAT 8.1.0 compiler?
> 
> https://sourceforge.net/projects/gnuada/files/GNAT_GCC%20Mac%20OS%20X/8.1.0/native-2017/
> 

I am also getting this problem with gcc 6.1.0
+===========================GNAT BUG DETECTED==============================+
| 6.1.0 (x86_64-apple-darwin15) GCC error:                                 |
| in gnat_to_gnu_entity, at ada/gcc-interface/decl.c:3011                  |
| Error detected at assimp_texture.ads:65:34   

As I have used gcc 6.1.0 without getting this error in the past I'm wondering if it has something to do with OSX.
Mac mini (Mid 2011)
OSX 10.13.6
Xcode 9.4.1

Re: AdaCore Community 2018 Bugs

[Simon Wright]

Roger <rogermcm2@gmail.com> writes:

> On Saturday, August 11, 2018 at 6:42:27 PM UTC+10, Jeffrey R. Carter wrote:
>> On 08/11/2018 03:06 AM, Roger wrote:
>> > Perhaps use a different Ada compiler?
>> 
>> Have you tried the FSF GNAT 8.1.0 compiler?
>> 
>> https://sourceforge.net/projects/gnuada/files/GNAT_GCC%20Mac%20OS%20X/8.1.0/native-2017/
>> 
>
> I am also getting this problem with gcc 6.1.0
> +===========================GNAT BUG DETECTED==============================+
> | 6.1.0 (x86_64-apple-darwin15) GCC error:                                 |
> | in gnat_to_gnu_entity, at ada/gcc-interface/decl.c:3011                  |
> | Error detected at assimp_texture.ads:65:34   
>
> As I have used gcc 6.1.0 without getting this error in the past I'm
> wondering if it has something to do with OSX.
> Mac mini (Mid 2011)
> OSX 10.13.6
> Xcode 9.4.1

I see your first error in this case.

Similar error on Debian stretch (GCC 6.3.0, GNAT GPL 2017).

Since the error seems to be related to AI_Mesh_Morph_Anim, I tried
various changes such as replacing your

   type Thing_List is new Thing_Package.List with null record;

by

   subtype Thing_List is Thing_Package.List;

but no joy.

Re: AdaCore Community 2018 Bugs

[Roger]

On Sunday, August 12, 2018 at 3:20:28 AM UTC+10, Simon Wright wrote:
> Roger <rogermcm2@gmail.com> writes:
> 
> > On Saturday, August 11, 2018 at 6:42:27 PM UTC+10, Jeffrey R. Carter wrote:
> >> On 08/11/2018 03:06 AM, Roger wrote:
> >> > Perhaps use a different Ada compiler?
> >> 
> >> Have you tried the FSF GNAT 8.1.0 compiler?
> >> 
> >> https://sourceforge.net/projects/gnuada/files/GNAT_GCC%20Mac%20OS%20X/8.1.0/native-2017/
> >> 
> >
> > I am also getting this problem with gcc 6.1.0
> > +===========================GNAT BUG DETECTED==============================+
> > | 6.1.0 (x86_64-apple-darwin15) GCC error:                                 |
> > | in gnat_to_gnu_entity, at ada/gcc-interface/decl.c:3011                  |
> > | Error detected at assimp_texture.ads:65:34   
> >
> > As I have used gcc 6.1.0 without getting this error in the past I'm
> > wondering if it has something to do with OSX.
> > Mac mini (Mid 2011)
> > OSX 10.13.6
> > Xcode 9.4.1
> 
> I see your first error in this case.
> 
> Similar error on Debian stretch (GCC 6.3.0, GNAT GPL 2017).
> 
> Since the error seems to be related to AI_Mesh_Morph_Anim, I tried
> various changes such as replacing your
> 
>    type Thing_List is new Thing_Package.List with null record;
> 
> by
> 
>    subtype Thing_List is Thing_Package.List;
> 
> but no joy.

Thanks for trying.
This type of problem first occurred a few months ago with a program that was much simpler than the one I am currently working on.
I've been writing Ada code for a few years and had not previously had this problem.
As it now occurs even with GCC 6.1.0 I'm thinking it most likely has to do with recent OSX changes.

Re: AdaCore Community 2018 Bugs

[Roger]

On Sunday, August 12, 2018 at 3:20:28 AM UTC+10, Simon Wright wrote:
> Roger <rogermcm2@gmail.com> writes:
> 
> > On Saturday, August 11, 2018 at 6:42:27 PM UTC+10, Jeffrey R. Carter wrote:
> >> On 08/11/2018 03:06 AM, Roger wrote:
> >> > Perhaps use a different Ada compiler?
> >> 
> >> Have you tried the FSF GNAT 8.1.0 compiler?
> >> 
> >> https://sourceforge.net/projects/gnuada/files/GNAT_GCC%20Mac%20OS%20X/8.1.0/native-2017/
> >> 
> >
> > I am also getting this problem with gcc 6.1.0
> > +===========================GNAT BUG DETECTED==============================+
> > | 6.1.0 (x86_64-apple-darwin15) GCC error:                                 |
> > | in gnat_to_gnu_entity, at ada/gcc-interface/decl.c:3011                  |
> > | Error detected at assimp_texture.ads:65:34   
> >
> > As I have used gcc 6.1.0 without getting this error in the past I'm
> > wondering if it has something to do with OSX.
> > Mac mini (Mid 2011)
> > OSX 10.13.6
> > Xcode 9.4.1
> 
> I see your first error in this case.
> 
> Similar error on Debian stretch (GCC 6.3.0, GNAT GPL 2017).
> 
> Since the error seems to be related to AI_Mesh_Morph_Anim, I tried
> various changes such as replacing your
> 
>    type Thing_List is new Thing_Package.List with null record;
> 
> by
> 
>    subtype Thing_List is Thing_Package.List;
> 
> but no joy.

Thanks for trying. 
This type of problem first occurred a few months ago with a program that was much simpler than the one I am currently working on. 
I've been writing Ada code for a few years and had not previously had this problem. 
As it now occurs even with GCC 6.1.0 I'm thinking it most likely has to do with recent OSX changes. 

This problem doesn't occur for compile only.
It occurs when I Build All from GPS

Re: AdaCore Community 2018 Bugs

[Simon Wright]

Roger <rogermcm2@gmail.com> writes:

> As it now occurs even with GCC 6.1.0 I'm thinking it most likely has
> to do with recent OSX changes.

Since it occurs on Debian Stretch, GCC 6.3.0, it can't be just macOS.

The actual error reported can change with successive compilations
(sometimes it says something like 'undeclared variable ""', seems to
settle on the CE), a surefire symptom of bad memory handling.

> This problem doesn't occur for compile only.
> It occurs when I Build All from GPS

The error is a crash in the compiler, doesn't matter how you get there.

Re: AdaCore Community 2018 Bugs

[Roger]

On Sunday, August 12, 2018 at 5:33:35 PM UTC+10, Simon Wright wrote:
> Roger <rogermcm2@gmail.com> writes:
> 
> > As it now occurs even with GCC 6.1.0 I'm thinking it most likely has
> > to do with recent OSX changes.
> 
> Since it occurs on Debian Stretch, GCC 6.3.0, it can't be just macOS.
> 
> The actual error reported can change with successive compilations
> (sometimes it says something like 'undeclared variable ""', seems to
> settle on the CE), a surefire symptom of bad memory handling.
> 
> > This problem doesn't occur for compile only.
> > It occurs when I Build All from GPS
> 
> The error is a crash in the compiler, doesn't matter how you get there.

Actually, my main program crashes in the same place every time, even when I change things.
I have some C code in my program, could that be the cause of f bad memory handling?
Solving this sort of problem is somewhat outside my scope of expertise so any advice is welcome.
The first time I had this problem was with a much shorter test program that I was experimenting with C interfaces suggesting that my C code might be the problem.
I'll see if I can investigate that.

Re: AdaCore Community 2018 Bugs

[Roger]

On Sunday, August 12, 2018 at 5:33:35 PM UTC+10, Simon Wright wrote:
> Roger <rogermcm2@gmail.com> writes:
> 
> > As it now occurs even with GCC 6.1.0 I'm thinking it most likely has
> > to do with recent OSX changes.
> 
> Since it occurs on Debian Stretch, GCC 6.3.0, it can't be just macOS.
> 
> The actual error reported can change with successive compilations
> (sometimes it says something like 'undeclared variable ""', seems to
> settle on the CE), a surefire symptom of bad memory handling.
> 
> > This problem doesn't occur for compile only.
> > It occurs when I Build All from GPS
> 
> The error is a crash in the compiler, doesn't matter how you get there.

I Have cut most of the code out of the test program but the crash still occurs.
However, I am making progress on tracking back to the minimal code where the crash first occurs.

Re: AdaCore Community 2018 Bugs

[Simon Wright]

Roger <rogermcm2@gmail.com> writes:

> On Sunday, August 12, 2018 at 5:33:35 PM UTC+10, Simon Wright wrote:
>> Roger <rogermcm2@gmail.com> writes:
>> 
>> > As it now occurs even with GCC 6.1.0 I'm thinking it most likely has
>> > to do with recent OSX changes.
>> 
>> Since it occurs on Debian Stretch, GCC 6.3.0, it can't be just macOS.
>> 
>> The actual error reported can change with successive compilations
>> (sometimes it says something like 'undeclared variable ""', seems to
>> settle on the CE), a surefire symptom of bad memory handling.
>> 
>> > This problem doesn't occur for compile only.
>> > It occurs when I Build All from GPS
>> 
>> The error is a crash in the compiler, doesn't matter how you get there.
>
> I Have cut most of the code out of the test program but the crash
> still occurs.  However, I am making progress on tracking back to the
> minimal code where the crash first occurs.

The problem is in the packages Morph.Mesh_Morph_Value_Pointers,
Mesh_Morph_Weight_Pointers.

You say

   package Mesh_Morph_Value_Pointers is new Interfaces.C.Pointers
     (Interfaces.C.unsigned, API_Morph_Value, API_Morph_Values_Array,
      API_Morph_Value'(others => <>));

but

   type API_Morph_Value is new Interfaces.C.unsigned;

so others => <> isn't a legal expression for an API_Morph_Value; you
should say

      API_Morph_Value'(0));


There is still a bug, which ought to be fixed, but it's that GNAT has
failed to recognise the actual problem and has got fatally confused.

Re: AdaCore Community 2018 Bugs

[Roger]

On Monday, August 13, 2018 at 2:37:52 AM UTC+10, Simon Wright wrote:
> Roger <rogermcm2@gmail.com> writes:
> 
> > On Sunday, August 12, 2018 at 5:33:35 PM UTC+10, Simon Wright wrote:
> >> Roger <rogermcm2@gmail.com> writes:
> >> 
> >> > As it now occurs even with GCC 6.1.0 I'm thinking it most likely has
> >> > to do with recent OSX changes.
> >> 
> >> Since it occurs on Debian Stretch, GCC 6.3.0, it can't be just macOS.
> >> 
> >> The actual error reported can change with successive compilations
> >> (sometimes it says something like 'undeclared variable ""', seems to
> >> settle on the CE), a surefire symptom of bad memory handling.
> >> 
> >> > This problem doesn't occur for compile only.
> >> > It occurs when I Build All from GPS
> >> 
> >> The error is a crash in the compiler, doesn't matter how you get there.
> >
> > I Have cut most of the code out of the test program but the crash
> > still occurs.  However, I am making progress on tracking back to the
> > minimal code where the crash first occurs.
> 
> The problem is in the packages Morph.Mesh_Morph_Value_Pointers,
> Mesh_Morph_Weight_Pointers.
> 
> You say
> 
>    package Mesh_Morph_Value_Pointers is new Interfaces.C.Pointers
>      (Interfaces.C.unsigned, API_Morph_Value, API_Morph_Values_Array,
>       API_Morph_Value'(others => <>));
> 
> but
> 
>    type API_Morph_Value is new Interfaces.C.unsigned;
> 
> so others => <> isn't a legal expression for an API_Morph_Value; you
> should say
> 
>       API_Morph_Value'(0));
> 
> 
> There is still a bug, which ought to be fixed, but it's that GNAT has
> failed to recognise the actual problem and has got fatally confused.


Thanks, very much appreciated.
I have made the change but still get the problem in my main program.
I'll check all my other new Interfaces.C.Pointers packages for similar errors.
I'll also update my bug report with this information.

Re: AdaCore Community 2018 Bugs

[Roger]

On Monday, August 13, 2018 at 2:37:52 AM UTC+10, Simon Wright wrote:
> Roger <rogermcm2@gmail.com> writes:
> 
> > On Sunday, August 12, 2018 at 5:33:35 PM UTC+10, Simon Wright wrote:
> >> Roger <rogermcm2@gmail.com> writes:
> >> 
> >> > As it now occurs even with GCC 6.1.0 I'm thinking it most likely has
> >> > to do with recent OSX changes.
> >> 
> >> Since it occurs on Debian Stretch, GCC 6.3.0, it can't be just macOS.
> >> 
> >> The actual error reported can change with successive compilations
> >> (sometimes it says something like 'undeclared variable ""', seems to
> >> settle on the CE), a surefire symptom of bad memory handling.
> >> 
> >> > This problem doesn't occur for compile only.
> >> > It occurs when I Build All from GPS
> >> 
> >> The error is a crash in the compiler, doesn't matter how you get there.
> >
> > I Have cut most of the code out of the test program but the crash
> > still occurs.  However, I am making progress on tracking back to the
> > minimal code where the crash first occurs.
> 
> The problem is in the packages Morph.Mesh_Morph_Value_Pointers,
> Mesh_Morph_Weight_Pointers.
> 
> You say
> 
>    package Mesh_Morph_Value_Pointers is new Interfaces.C.Pointers
>      (Interfaces.C.unsigned, API_Morph_Value, API_Morph_Values_Array,
>       API_Morph_Value'(others => <>));
> 
> but
> 
>    type API_Morph_Value is new Interfaces.C.unsigned;
> 
> so others => <> isn't a legal expression for an API_Morph_Value; you
> should say
> 
>       API_Morph_Value'(0));
> 
> 
> There is still a bug, which ought to be fixed, but it's that GNAT has
> failed to recognise the actual problem and has got fatally confused.

Thanks, very much appreciated. 
I now have my main program building without crashing GNAT.
I'll also update my bug report with this information.