GNAT v3.09 says GNAT bug- but 2.07 didn't. Please HELP!

GNAT v3.09 says GNAT bug- but 2.07 didn't. Please HELP!

[William Paul Berriss]

Hi

I have recently upgraded to GNAT v3.09 from v2.07 on 
Solaris 2.3(2.4 GNAT) for a Sun SPARC Classic workstation.
It has installed fine and works but I get the following
(major) problem on compiling all  (well nearly all, as I have just got
one or two to compile somehow?) of my programs that use a specific 
generic package called raster_io.

Here is what happens:

strberis@stssrita% gcc -c multiplier.adb
+===========================GNAT BUG DETECTED==============================+
| Error detected at multiplier.adb:64:46                                   |
| Please submit bug report by email to report@gnat.com                     |
| Use a subject line meaningful to you and us to track the bug             |
| Include full sources in ASCII in a format compatible with gnatchop       |
| First line of sources must be marked by an Ada -- comment line           |
| Last line of sources must be last line of email message (no signature!)  |
| See gnatinfo.txt file for more info on procedure for submitting bugs     |
| 3.09 (970121) (sparc-sun-solaris2.4) Gigi abort, Code=999                |
+==========================================================================+
compilation abandoned
strberis@stssrita% 

strberis@stssrita% gnatmake multiplier.adb
gcc -c multiplier.adb
+===========================GNAT BUG DETECTED==============================+
| Error detected at multiplier.adb:64:46                                   |
| Please submit bug report by email to report@gnat.com                     |
| Use a subject line meaningful to you and us to track the bug             |
| Include full sources in ASCII in a format compatible with gnatchop       |
| First line of sources must be marked by an Ada -- comment line           |
| Last line of sources must be last line of email message (no signature!)  |
| See gnatinfo.txt file for more info on procedure for submitting bugs     |
| 3.09 (970121) (sparc-sun-solaris2.4) Gigi abort, Code=999                |
+==========================================================================+
compilation abandoned
gnatmake: "multiplier.adb" compilation error
[2]  + Done                 textedit binary_image_erosion.adb
strberis@stssrita%

However, the line 64 it refers to is:
    My_Raster_IO.Read_Raster_File(FILENAME1, I); 
and I (column 46)  is of type Image,
    type Image is array(RowCoord, ColCoord) of Colour_Pixel;   -- RGB colour
where Colour_Pixel is defined in another with'd package and is
  type COLOUR_PIXEL is
    record
      R, G, B : PIXEL;
    end record;
  for COLOUR_PIXEL'SIZE use 24;         

and I have with'd my generic package:

  with EEG.WPB.RASTER_IO;

and instantiated it with the line :

    package My_Raster_IO is new EEG.WPB.Raster_IO(RowCoord, ColCoord, Colour_Pixel, Image);

This generic raster_io package spec is

generic

  type RowCoordinate is range <>;
  type ColumnCoordinate is range <>;
  type Pixel is private;                         -- Monochrome or Colour, 8-bit or 24-bit
  type Image is array(RowCoordinate, ColumnCoordinate) of Pixel;

package EEG.WPB.Raster_IO is

  procedure Read_Raster_File(FILENAME: in String; I: out Image);  -- You get given an image
  procedure Write_Raster_File(FILENAME: in String; I: in Image);


end EEG.WPB.Raster_IO;


I hope enough information is given above.
This (and all my) programs worked fine under v2,07 last week, but wont under v3.09 as you can see.
Has something been left over from v2.07 that I should delete?
Can you explain to me what has happened?  I do not think there is a bug, as two of my 
other programs that use raster_io now work.

Can you tell me how to fix it? or what is happening?

I am really confused, especially now that some of my programs that use
raster_io will compile


Ah ha!  I think I can narrow it down to this.
Everthing is fine if I instantiate raster_io with plain old Pixel which is defined as:

  type PIXEL is range 0 .. 255;
  for  PIXEL'SIZE use 8;

But if I instantiate with Colour_Pixel  (a record) then if give the gnat bug error message
when I try to compile it.
What do I do now?  My research is COLOUR image processing, I use colour images a lot.
Can you tell me if it is MY syntax error or not?  If so please e-mail me.

Sorry this is so long but I wanted to be explicit.

Yours sincerely

Will Berriss
============


-- 

W P Berriss                 E-mail: W.P.Berriss@reading.ac.uk
Department of Engineering
The University of Reading
Whiteknights
Reading                     Tel:  0118 987 5123 
Berkshire                     (+44 118 987 5123 outside UK)
RG6 6AY    
England                     Fax:  0118 931 3327    

World Wide Web Home Page: 

http://www.elec.rdg.ac.uk/people/postgrads/will.html

Re: GNAT v3.09 says GNAT bug- but 2.07 didn't. Please HELP!

[Robert Dewar]

William Berris posted a raw unanalyzed GNAT bug report to CLA. I trust
he read the contents of the bug boxes, particularly the line that says

"| Please submit bug report by email to report@gnat.com                    "

So far, we have not seen the bug report! Generally the only folks likely
to be able to interprete the contents of GNAT bug boxes live at
report@gnat.com. We do NOT take bug reports from CLA, and do not
encourage posting them here. It is sometimes useful to post an analyzed
bug -- "don't do xyz when using compiler bla-bla, because mutter-mutter",
but it is generally not useful to post raw bug reports, and in any
case it is critical to send them to the proper support location, no
matter what compiler you are using!

Robert Dewar
Ada Core Technologies