GNAT GPL 2010 now available, update portfile to use it?

[sjs@…]

GNAT GPL 2010 has now been released and supports Snow Leopard. This should make possible a working build on SL. At present the portfile uses GNAT GPL 2009. Request an update to use GNAT GPL 2010, or provide an option for it.

[jmr@…]

Please remember to cc the maintainer.

[krischik@…]

Yes I am currently updating the Portfile to create 4.5.1 with GPL 2010. Note that gnat-gcc is a very time consuming port - any help is welcome. Drop me a mail an I send you the current Portfile and test scripts.

[larryv@…]

Cc Me!

[mikewolz@…]

Cc Me!

[jarimatti@…]

Cc Me!

[terry.droeger@…]

Cc Me!

[krischik@…]

Portfile for gcc 4.5.2

[krischik@…]

Hello,

since there are now two users which want to help me fixing GNAT/GCC I have regained hope that it could be done between the now three of us.

I have update the make files to 4.5.2. If you just call port install you get:

configure: error: in `/opt/local/var/macports/build/_Volumes_Daten_Developer_MacPorts_dports_lang_gnat-gcc/work/build/gcc':
configure: error: C++ preprocessor "/lib/cpp" fails sanity check

This sanity checks are one of the more frustrating parts us using configure. The checks are often fail even everything is ok. Or /lib/cpp" is used even when configure was told to use "/opt/gnat-2010-x86_64-apple-darwin9.6.0-bin/bin/cpp".

Anyway, this is where attached test.command comes to play. After the failed port install you use test.command to compile again. Without Macports in my back I have better control over the options send to configure and as result the compile runs further.

How far I don't yet know as the compile is still running.

Regards

Martin

PS: I attach the files after submitting this comment.

[krischik@…]

Test compile script useable for bug hunting.

[krischik@…]

Now I know how far it gets (faster then expected):

/opt/gnat-2010-x86_64-apple-darwin9.6.0-bin/bin/gcc -c -g -fkeep-inline-functions  -gnatpg -gnata -gnatwns -nostdinc -I- -I. -Iada -I../../gcc-4.5.2/gcc/ada -I../../gcc-4.5.2/gcc/ada/gcc-interface ../../gcc-4.5.2/gcc/ada/uintp.adb -o ada/uintp.o
uintp.adb:242:22: expect type "Hnum" defined at line 91
make[3]: *** [ada/uintp.o] Error 1
make[2]: *** [all-stage1-gcc] Error 2
make[1]: *** [stage1-bubble] Error 2
make: *** [all] Error 2

Looks like an incompatible change to the system. We have to analyse this now and create patch.

Martin

[jarimatti@…]

Replying to krischik@…:

Now I know how far it gets (faster then expected):

/opt/gnat-2010-x86_64-apple-darwin9.6.0-bin/bin/gcc -c -g -fkeep-inline-functions  -gnatpg -gnata -gnatwns -nostdinc -I- -I. -Iada -I../../gcc-4.5.2/gcc/ada -I../../gcc-4.5.2/gcc/ada/gcc-interface ../../gcc-4.5.2/gcc/ada/uintp.adb -o ada/uintp.o
uintp.adb:242:22: expect type "Hnum" defined at line 91
make[3]: *** [ada/uintp.o] Error 1
make[2]: *** [all-stage1-gcc] Error 2
make[1]: *** [stage1-bubble] Error 2
make: *** [all] Error 2

Looks like an incompatible change to the system. We have to analyse this now and create patch.

I noticed that last weekend. The error is in the GNAT Ada sources

/opt/local/var/macports/sources/rsync.macports.org/release/ports/lang/gnat-gcc/work/gcc-4.5.1/gcc/ada/uintp.adb

Here's a fix, it's at line 242 or so (sorry, don't know how to make a proper patch yet):

   --------------
   -- Hash_Num --
   --------------

   function Hash_Num (F : Int) return Hnum is
   begin
      --  return Standard."mod" (F, Hnum'Range_Length);
      return Hnum (Standard."mod" (F, Hnum'Range_Length));
   end Hash_Num;

After this the compiler gets built but the output is different between stages and this causes it to fail.

-Jarimatti

[jarimatti@…]

Hi,

Compiling using test.command and GCC 4.5.2 fails at:

/opt/local/var/macports/sources/rsync.macports.org/release/ports/lang/gnat-gcc/work/build/./gcc/xgcc -B/opt/local/var/macports/sources/rsync.macports.org/release/ports/lang/gnat-gcc/work/build/./gcc/ -B/opt/local/x86_64-apple-darwin10/bin/ -B/opt/local/x86_64-apple-darwin10/lib/ -isystem /opt/local/x86_64-apple-darwin10/include -isystem /opt/local/x86_64-apple-darwin10/sys-include    -c -g -O2 -m32 -fPIC -pipe  -W -Wall -gnatpg -m32  a-direct.adb -o a-direct.o
a-direct.adb:676:13: warning: types for unchecked conversion have different sizes
make[8]: *** [a-direct.o] Error 1
make[7]: *** [gnatlib] Error 2
make[6]: *** [gnatlib-shared-darwin] Error 2
make[5]: *** [gnatlib-shared] Error 2
make[4]: *** [gnatlib-shared] Error 2
make[3]: *** [multi-do] Error 1
make[2]: *** [all] Error 2
make[1]: *** [all-target-libada] Error 2
make: *** [all] Error 2

The error report comes from a-direct.adb, procedure Fetch_Next_Entry. Reading the build output I see that ADAFLAGS and CFLAGS both contain "-m32", which could explain the error (I'm compiling on iMac i3, SL 10.6.6). The error is in conversion between System.Address and access value, if I read it correctly.

Any idea where to fix those '-m32' flags?

[krischik@…]

creating a patch file:

cd work
sudo gcp --archive gcc-4.5.2 gcc-4.5.2.orig
sudo gvim gcc-4.5.2/gcc/ada/uintp.adb
sudo zsh -c 'diff gcc-4.5.2.orig/gcc/ada/uintp.adb gcc-4.5.2/gcc/ada/uintp.adb >uintp.adb.diff'

And yes, the fact that the compile is done by root is a bloody pain in the arse. The next MacPorts is supposed to fix that.

[krischik@…]

The current error message is:

/Volumes/Daten/Developer/MacPorts/dports/lang/gnat-gcc/work/build/./gcc/xgcc -B/Volumes/Daten/Developer/MacPorts/dports/lang/gnat-gcc/work/build/./gcc/ -B/opt/local/x86_64-apple-darwin10/bin/ -B/opt/local/x86_64-apple-darwin10/lib/ -isystem /opt/local/x86_64-apple-darwin10/include -isystem /opt/local/x86_64-apple-darwin10/sys-include    -c -g -O2 -m32 -fPIC -pipe  -W -Wall -gnatpg -m32  a-direct.adb -o a-direct.o
a-direct.adb:676:13: warning: types for unchecked conversion have different sizes

The code in question is:

            function Address_To_Access is new
              Ada.Unchecked_Conversion
                (Source => Address,
                 Target => Path_String_Access);

Strange. Anybody got an idea what went wrong here?

[jarimatti@…]

Replying to krischik@…:

The current error message is:

/Volumes/Daten/Developer/MacPorts/dports/lang/gnat-gcc/work/build/./gcc/xgcc -B/Volumes/Daten/Developer/MacPorts/dports/lang/gnat-gcc/work/build/./gcc/ -B/opt/local/x86_64-apple-darwin10/bin/ -B/opt/local/x86_64-apple-darwin10/lib/ -isystem /opt/local/x86_64-apple-darwin10/include -isystem /opt/local/x86_64-apple-darwin10/sys-include    -c -g -O2 -m32 -fPIC -pipe  -W -Wall -gnatpg -m32  a-direct.adb -o a-direct.o
a-direct.adb:676:13: warning: types for unchecked conversion have different sizes

The code in question is:

            function Address_To_Access is new
              Ada.Unchecked_Conversion
                (Source => Address,
                 Target => Path_String_Access);

Strange. Anybody got an idea what went wrong here?

Could the issue be 32 bit vs. 64 bit? (see #comment:10). Is there any way to force -m64 instead of -m32 for the xgcc arguments?

[egall@…]

Cc Me!