public inbox for gcc@gcc.gnu.org
 help / color / mirror / Atom feed
* [Ada] Partial bootstrap success on Debian/x86
@ 2001-10-10  4:57 Florian Weimer
  2001-10-10  7:12 ` Florian Weimer
  2001-10-11 11:45 ` Gordon Sadler
  0 siblings, 2 replies; 5+ messages in thread
From: Florian Weimer @ 2001-10-10  4:57 UTC (permalink / raw)
  To: gcc

I'm glad to report that current CVS is close to building
out-of-the-box on Debian (x86 architecture, not-too-current sid),
using the Debian gnat package (which includes a driver called
"gnatgcc") and the C compiler of GCC 2.95.4 20010721 (Debian
prerelease).  The stage3 binaries seems to be able to compile
correctly most of my code, at least my test cases work. ;-)

Only the xnmake problem reported earlier and the debug information
comparison failures remain.

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [Ada] Partial bootstrap success on Debian/x86
  2001-10-10  4:57 [Ada] Partial bootstrap success on Debian/x86 Florian Weimer
@ 2001-10-10  7:12 ` Florian Weimer
  2001-10-10  7:16   ` Lutz Donnerhacke
  2001-10-11 11:45 ` Gordon Sadler
  1 sibling, 1 reply; 5+ messages in thread
From: Florian Weimer @ 2001-10-10  7:12 UTC (permalink / raw)
  To: gcc

Florian Weimer <fw@deneb.enyo.de> writes:

> Only the xnmake problem reported earlier and the debug information
> comparison failures remain.

Lutz Donnerhacke mentioned another problem which didn't occur to me
because it also exists in the public GNAT releases (and I've got a
proper build environment for them):

The makefile tries to link with libaddr2line.a.  This library does not
exist on most systems, and it is not part of binutils (there's only a
addr2line command, which seems to do roughly the same).

I've never seen source code for this library, so I can't tell if it's
feasible to build it during the GCC build.

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [Ada] Partial bootstrap success on Debian/x86
  2001-10-10  7:12 ` Florian Weimer
@ 2001-10-10  7:16   ` Lutz Donnerhacke
  0 siblings, 0 replies; 5+ messages in thread
From: Lutz Donnerhacke @ 2001-10-10  7:16 UTC (permalink / raw)
  To: gcc

* Florian Weimer wrote:
>The makefile tries to link with libaddr2line.a.  This library does not
>exist on most systems, and it is not part of binutils (there's only a
>addr2line command, which seems to do roughly the same).
>
>I've never seen source code for this library, so I can't tell if it's
>feasible to build it during the GCC build.

I'm on the way to provide a addr2line.c for the gcc tree. Should not take
longer than some weeks.

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [Ada] Partial bootstrap success on Debian/x86
  2001-10-10  4:57 [Ada] Partial bootstrap success on Debian/x86 Florian Weimer
  2001-10-10  7:12 ` Florian Weimer
@ 2001-10-11 11:45 ` Gordon Sadler
  2001-10-11 13:18   ` Florian Weimer
  1 sibling, 1 reply; 5+ messages in thread
From: Gordon Sadler @ 2001-10-11 11:45 UTC (permalink / raw)
  To: gcc

On Wed, Oct 10, 2001 at 02:14:05PM +0200, Florian Weimer wrote:
> I'm glad to report that current CVS is close to building
> out-of-the-box on Debian (x86 architecture, not-too-current sid),
> using the Debian gnat package (which includes a driver called
> "gnatgcc") and the C compiler of GCC 2.95.4 20010721 (Debian
> prerelease).  The stage3 binaries seems to be able to compile
> correctly most of my code, at least my test cases work. ;-)
> 
> Only the xnmake problem reported earlier and the debug information
> comparison failures remain.
> 
After reading most of the various threads relating to ada being added to
cvs head, I've tried various ways to bootstrap, nothing succeeds as of
now.

I use Debian unstable(updated daily) and use CC=gnatgcc in my build
scripts. I used gnatmake to make xtreeprs xsinfo xnmake xeinfo and place
them in /usr/local/bin. I try to bootstrap daily and have seen various
different errors every day.

Today's error:
gnatgcc -c  -DIN_GCC    -g -O2 -march=i686 -W -Wall -Wwrite-strings
-Wstrict-prototypes -Wmissing-prototypes -Wtraditional -pedantic
-Wno-long-long  -DHAVE_CONFIG_H -DGENERATOR_FILE    -I. -I.
-I/usr/src/cvs/gcc/gcc -I/usr/src/cvs/gcc/gcc/.
-I/usr/src/cvs/gcc/gcc/config -I/usr/src/cvs/gcc/gcc/../include
/usr/src/cvs/gcc/gcc/gengenrtl.c
cc1: Invalid option `-Wno-long-long'
In file included from /usr/include/unistd.h:500,
                 from /usr/src/cvs/gcc/gcc/system.h:157,
		 from /usr/src/cvs/gcc/gcc/gengenrtl.c:23:
				 
/usr/include/bits/confname.h:565: warning: comma at end of enumerator list
make[2]: *** [gengenrtl.o] Error 1
make[2]: Leaving directory `/usr/local/src/build/gcc/gcc'
make[1]: *** [stage2_build] Error 2
make[1]: Leaving directory `/usr/local/src/build/gcc/gcc'
make: *** [bootstrap] Error 2

Any pointers on your method of success?

-- 
Gordon Sadler

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [Ada] Partial bootstrap success on Debian/x86
  2001-10-11 11:45 ` Gordon Sadler
@ 2001-10-11 13:18   ` Florian Weimer
  0 siblings, 0 replies; 5+ messages in thread
From: Florian Weimer @ 2001-10-11 13:18 UTC (permalink / raw)
  To: gcc

Gordon Sadler <gbsadler1@lcisp.com> writes:

> I use Debian unstable(updated daily) and use CC=gnatgcc in my build
> scripts. I used gnatmake to make xtreeprs xsinfo xnmake xeinfo and place
> them in /usr/local/bin. 

Neither should be necessary with current CVS.  

Could you give the instructions in < 873d4pdist.fsf@deneb.enyo.de > a
try?  Especially on Debian, you shouldn't need to see any environment
variables or compile any helper applications.

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2001-10-11 13:18 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2001-10-10  4:57 [Ada] Partial bootstrap success on Debian/x86 Florian Weimer
2001-10-10  7:12 ` Florian Weimer
2001-10-10  7:16   ` Lutz Donnerhacke
2001-10-11 11:45 ` Gordon Sadler
2001-10-11 13:18   ` Florian Weimer

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).