Re: gcc10 with Ada support added to wip

[Chavdar Ivanov <ci4ic4%gmail.com@localhost>]

On Sat, 6 Nov 2021 at 14:33, Chavdar Ivanov <ci4ic4%gmail.com@localhost> wrote:
>
> On Sat, 6 Nov 2021 at 13:39, Fernando Oleo Blanco <irvise_ml%irvise.xyz@localhost> wrote:
> >
> > Am Sat, 6 Nov 2021 10:37:33 +0000
> > schrieb "Chavdar Ivanov" <ci4ic4%gmail.com@localhost>:
> > > The difference between 9.2 and -current has grown up quite a bit, I'd
> > > say. I've had a few other occasions when stuff was reluctant to work
> > > under -current and had to be frogmarched somehow. If one is running
> > > rolling pkgsrc (pkg_rolling-replace every ten days or so) on a
> > > -current system (updated daily), one is expecting things to break from
> > > time to time. All in all, my NetBSD systems have been remarkably
> > > well-behaved.
> > >
> > > >
> > > > P.S: did you 'make clean' between option changes? I have had 'make
> > > > patch' fail if it had already patched the system.
> > >
> > > Yes, every attempt started with 'make clean'. The 'make patch'
> > > failures were always due to not being able to locate the target file
> > > to patch.
> > >
> > > >
> > > > Regards,
> > > > Fer
> > > >
> > >
> > > Cheers,
> > >
> > > Chavdar
> > >
> > > --
> > > ----
> >
> > I come with good news and bad news... Good news is that it works in
> > NetBSD-current. It was built successfully by racoon on 9.99.92 and me
> > just today.
> >
> > I suppose racoon built it in a sandbox. I built it in a clean install
> > that I just did, so equivalent to a sandbox.
> >
> > So... I do not know what is wrong with your setup. gcc6-aux should not
> > require ISL, since it is not being configured for it and it is not
> > pulling it as a dependency. And your gcc10-aux was failing in libgcc...
> > Which is not something that I have touched at all...
> >
> > If you could send me the make logs, I could take a closer look. But
> > with the issues that you have been having, I would say that you have
> > some flag, environment, system or variable configured in a way that
> > makes gcc6-aux/gcc10-aux break. I am not saying it is intentional. If
> > you have been rocking -current for a while, I suppose that after some
> > updates and changes, something is bound to change in an unexpected way.
> >
> > I am currently running 'make test' to see if there are any differences
> > with NetBSD-9.2.
>
> I think I located the problem - what fails for me is actually 'make
> patch'. Here is an excerpt from the output of 'make -d all patch' :
>
> ....
>            test -n "" -a -d /wip/gcc10-aux; then  echo 1>&2 "ERROR:"
> "==========================================================================";
>  echo 1>&2 "ERROR:";
>  echo 1>&2 "ERROR:" "Some of the selected build options and/or local
> patches may be incompatible.";  echo 1>&2 "ERROR:" "Please try
> building with fewer options or patches.";  echo 1>&2 "ERROR:";
>                                               echo 1>&2 "ERROR:"
> "==========================================================================";
>  fi; exit 1;                                   fi
> No such line 2320 in input file, ignoring
> ....
>
> For whatever reason it doesn't prepend /usr/pkgsrc , the patch fails
> and it then  ignores ... Obviously if none of the patches have taken
> place the build should fail.
>
> I'll trim my /etc/mk.conf and try again.

I am getting the same with almost empty /etc/mk.conf:
...
ar  rc libgcov.a $objects
ranlib libgcov.a
/usr/pkgsrc/wip/gcc10-aux/work/build/./gcc/xgcc
-B/usr/pkgsrc/wip/gcc10-aux/work/build/./gcc/
-B/usr/pkg/gcc10-aux/x86_64--netbsd/bin/
-B/usr/pkg/gcc10-aux/x86_64--netbsd/lib/ -isystem
/usr/pkg/gcc10-aux/x86_64--netbsd/include -isystem
/usr/pkg/gcc10-aux/x86_64--netbsd/sys-include   -fno-checking -g -O2
-I/usr/include -I/usr/pkg/include -O2  -g -O2 -I/usr/include
-I/usr/pkg/include -DIN_GCC    -W -Wall -Wno-narrowing -Wwrite-strings
-Wcast-qual -Wno-format -Wstrict-prototypes -Wmissing-prototypes
-Wold-style-definition  -isystem ./include  -fpic -g -DIN_LIBGCC2
-fbuilding-libgcc -fno-stack-protector  -fpic -I. -I. -I../.././gcc
-I../../../gcc-10.3.0/libgcc -I../../../gcc-10.3.0/libgcc/.
-I../../../gcc-10.3.0/libgcc/../gcc
-I../../../gcc-10.3.0/libgcc/../include  -DHAVE_CC_TLS  -o
unwind-dw2.o -MT unwind-dw2.o -MD -MP -MF unwind-dw2.dep -fexceptions
-c ../../../gcc-10.3.0/libgcc/unwind-dw2.c -fvisibility=hidden
-DHIDE_EXPORTS
In file included from ../../../gcc-10.3.0/libgcc/unwind-dw2.c:35:
../../../gcc-10.3.0/libgcc/unwind-pe.h:100:8: error: unknown type name
'_Unwind_Ptr'
  100 | static _Unwind_Ptr
      |        ^~~~~~~~~~~
../../../gcc-10.3.0/libgcc/unwind-pe.h:131:39: error: unknown type
name '_uleb128_t'
  131 | read_uleb128 (const unsigned char *p, _uleb128_t *val)
....


FYI my cut-down /etc/mk.conf is as follows:

PKGSRC_USE_SSP=NO
SKIP_LICENSE_CHECK=             yes
PKGSRC_USE_FORTIFY=NO
ALLOW_VULNERABLE_PACKAGES=True
CURSES_DEFAULT=ncursesw
GO_VERSION_DEFAULT=117
MKARGON2=       YES
PKG_OPTIONS.gcc10-aux+= -gcc-c++ -gcc-fortran
PYTHON_VERSION_DEFAULT=39


>
>
> >
> > Regards,
> > Fer
> >
>
>
> --
> ----



-- 
----