Re: Recent -CURRENT unable to build numerous ports

[Dimitry Andric]

On 22 Oct 2016, at 06:47, Aryeh Friedman  wrote:
> 
>> This was a side-effect of r307676, which added transformation rules for
>> .bco and .llo files (LLVM bytecode in binary and text representation).
>> 
>> Because .SUFFIXES was not updated to match, bmake was actually trying to
>> build a ".c.bco" file in the above case...
>> 
>> I committed a fix in r307754.
>> 
>> 
> Still doesn't work and still corrupts the base system (see output after
> portmaster failure) on a fresh from CD install of 11-STABLE (post svlite of
> /usr/src and make DESTDIR=/ world kernel):
> 
> root@lilith:/usr/src # uname -a
> FreeBSD lilith 10.3-STABLE FreeBSD 10.3-STABLE #0 r307759: Sat Oct 22
> 00:14:55 EDT 2016 aryehl@lilith:/usr/obj/usr/src/sys/GENERIC  amd64

You said that you have a fresh install of 11-STABLE, but this shows
10.3-STABLE?  What is it?


> ===>   Registering installation for gettext-runtime-0.19.8.1
> pkg-static: Unable to access file
> /usr/ports/devel/gettext-runtime/work/stage/usr/local/include/autosprintf.h:
> No such file or directory
> pkg-static: Unable to access file
> /usr/ports/devel/gettext-runtime/work/stage/usr/local/lib/libasprintf.a: No
> such file or directory
> pkg-static: Unable to access file
> /usr/ports/devel/gettext-runtime/work/stage/usr/local/lib/libasprintf.so:
> No such file or directory
> pkg-static: Unable to access file
> /usr/ports/devel/gettext-runtime/work/stage/usr/local/lib/libasprintf.so.0:
> No such file or directory
> pkg-static: Unable to access file
> /usr/ports/devel/gettext-runtime/work/stage/usr/local/lib/libasprintf.so.0.0.0:
> No such file or directory
> *** Error code 74

This has nothing to do with the fix I checked in.  It seems more like
packaging problem in the gettext port, or some problem on your system.

Maybe you can try cleaning the port, making sure the work directory is
gone, then build it again.



> root@lilith:/usr/ports # cd /usr/src
> root@lilith:/usr/src # make DESTDIR=/ world kernel;

Don't do this.  IMHO the "world" target should be removed, as it is too
easy to shoot yourself in the foot with it.

The correct way to do a source-based upgrade is:
* make buildworld
* make buildkernel
* make installkernel
* reboot in single user
* etcupdate -p
* make installworld
* etcupdate
* reboot


> --
 stage 1.1: legacy release compatibility shims
> --
> cd /usr/src; MAKEOBJDIRPREFIX=/usr/obj/usr/src/tmp  INSTALL="sh
> /usr/src/tools/install.sh"
> PATH=/usr/obj/usr/src/tmp/legacy/usr/sbin:/usr/obj/usr/src/tmp/legacy/usr/bin:/usr/obj/usr/src/tmp/legacy/usr/games:/usr/obj/usr/src/tmp/legacy/bin:/sbin:/bin:/usr/sbin:/usr/bin
> WORLDTMP=/usr/obj/usr/src/tmp  VERSION="FreeBSD 10.3-STABLE amd64 1003509"

Here is the discrepancy again, if this is 11-STABLE, why does it show
10.3-STABLE here?


> c++  -O2 -pipe
> -I/usr/src/lib/clang/libllvmsupport/../../../contrib/llvm/include
> -I/usr/src/lib/clang/libllvmsupport/../../../contrib/llvm/tools/clang/include
> -I/usr/src/lib/clang/libllvmsupport/../../../contrib/llvm/lib/Support -I.
> -I/usr/src/lib/clang/libllvmsupport/../../../contrib/llvm/../../lib/clang/include
> -DLLVM_ON_UNIX -DLLVM_ON_FREEBSD -D__STDC_LIMIT_MACROS
> -D__STDC_CONSTANT_MACROS -DNDEBUG -fno-strict-aliasing
> -DLLVM_DEFAULT_TARGET_TRIPLE=\"x86_64-unknown-freebsd10.3\"
> -DLLVM_HOST_TRIPLE=\"x86_64-unknown-freebsd10.3\" -DDEFAULT_SYSROOT=\"\"

And here the target and host are both 10.3.  What gives?


> /usr/src/lib/clang/libllvmsupport/../../../contrib/llvm/include/llvm/Support/AlignOf.h:19:
> /usr/include/c++/v1/cstddef:58:30: error: redefinition of 'nullptr_t'
> struct _LIBCPP_TYPE_VIS_ONLY nullptr_t
> ^
> /usr/include/c++/v1/__nullptr:24:30: note: previous definition is here
> struct _LIBCPP_TYPE_VIS_ONLY nullptr_t
> ^

It looks like you might have partially installed an 11-STABLE userland,
and now are trying to build a 10-STABLE world?  I am not sure if that
is supposed to work.

If your system is rather hosed, as I think it is, it might be easier to
just reinstall directly from the 11.0-RELEASE ISOs.  Then you can try
building 11-STABLE again.

-Dimitry



signature.asc
Description: Message signed with OpenPGP using GPGMail

Re: Recent -CURRENT unable to build numerous ports

[Marcus von Appen]

On, Fri Oct 21, 2016, Dimitry Andric wrote:

> On 21 Oct 2016, at 20:47, Marcus von Appen  wrote:
> >
> > -CURRENT as of r307731 seems to have some serious flaw with its build
> > tool environment. Many ports fail to build with some error similar to
> > the following (devel/apr1):
> >
> > cc -emit-llvm -fno-strict-aliasing -pipe -march=native -DLIBICONV_PLUG
> >  -fstack-protector -c -o .c.bco
> > cc: error: no input files
> > *** [.c.bco] Error code 1
> > [...]
> >
> > Since not all ports are affected, I assume some auto[conf|tool|make]
> > issue to be triggered. Hints for getting this one solved are highly
> > appreciated.
>
> This was a side-effect of r307676, which added transformation rules for
> .bco and .llo files (LLVM bytecode in binary and text representation).
>
> Because .SUFFIXES was not updated to match, bmake was actually trying to
> build a ".c.bco" file in the above case...
>
> I committed a fix in r307754.

Thanks, everything seems to be back to normal for me now!

Cheers
Marcus


signature.asc
Description: PGP signature

Re: Recent -CURRENT unable to build numerous ports

[Aryeh Friedman]

> This was a side-effect of r307676, which added transformation rules for
> .bco and .llo files (LLVM bytecode in binary and text representation).
>
> Because .SUFFIXES was not updated to match, bmake was actually trying to
> build a ".c.bco" file in the above case...
>
> I committed a fix in r307754.
>
>
Still doesn't work and still corrupts the base system (see output after
portmaster failure) on a fresh from CD install of 11-STABLE (post svlite of
/usr/src and make DESTDIR=/ world kernel):

root@lilith:/usr/src # uname -a
FreeBSD lilith 10.3-STABLE FreeBSD 10.3-STABLE #0 r307759: Sat Oct 22
00:14:55 EDT 2016 aryehl@lilith:/usr/obj/usr/src/sys/GENERIC  amd64


===>   Registering installation for gettext-runtime-0.19.8.1
pkg-static: Unable to access file
/usr/ports/devel/gettext-runtime/work/stage/usr/local/include/autosprintf.h:
No such file or directory
pkg-static: Unable to access file
/usr/ports/devel/gettext-runtime/work/stage/usr/local/lib/libasprintf.a: No
such file or directory
pkg-static: Unable to access file
/usr/ports/devel/gettext-runtime/work/stage/usr/local/lib/libasprintf.so:
No such file or directory
pkg-static: Unable to access file
/usr/ports/devel/gettext-runtime/work/stage/usr/local/lib/libasprintf.so.0:
No such file or directory
pkg-static: Unable to access file
/usr/ports/devel/gettext-runtime/work/stage/usr/local/lib/libasprintf.so.0.0.0:
No such file or directory
*** Error code 74

Stop.
make: stopped in /usr/ports/devel/gettext-runtime

===>>> Installation of gettext-runtime-0.19.8.1 (devel/gettext-runtime)
failed
===>>> Aborting update

===>>> Installation of print/indexinfo (indexinfo-0.2.5) complete


===>>> You can restart from the point of failure with this command line:
   portmaster  devel/gettext-runtime

root@lilith:/usr/ports # cd /usr/src
root@lilith:/usr/src # make DESTDIR=/ world kernel;
--
>>> make world started on Sat Oct 22 00:41:33 EDT 2016
--
--
>>> World build started on Sat Oct 22 00:41:33 EDT 2016
--

--
>>> Rebuilding the temporary build tree
--
rm -rf /usr/obj/usr/src/tmp/legacy/usr/include
rm -f /usr/obj/usr/src/usr.bin/kdump/ioctl.c
rm -f /usr/obj/usr/src/usr.bin/kdump/kdump_subr.c
rm -f /usr/obj/usr/src/usr.bin/truss/ioctl.c
mkdir -p /usr/obj/usr/src/tmp/lib
mkdir -p /usr/obj/usr/src/tmp/usr
mkdir -p /usr/obj/usr/src/tmp/legacy/bin
mkdir -p /usr/obj/usr/src/tmp/legacy/usr
mtree -deU -f /usr/src/etc/mtree/BSD.usr.dist  -p
/usr/obj/usr/src/tmp/legacy/usr >/dev/null
mtree -deU -f /usr/src/etc/mtree/BSD.groff.dist  -p
/usr/obj/usr/src/tmp/legacy/usr >/dev/null
mtree -deU -f /usr/src/etc/mtree/BSD.usr.dist  -p /usr/obj/usr/src/tmp/usr
>/dev/null
mtree -deU -f /usr/src/etc/mtree/BSD.include.dist  -p
/usr/obj/usr/src/tmp/usr/include >/dev/null
ln -sf /usr/src/sys /usr/obj/usr/src/tmp

--
>>> stage 1.1: legacy release compatibility shims
--
cd /usr/src; MAKEOBJDIRPREFIX=/usr/obj/usr/src/tmp  INSTALL="sh
/usr/src/tools/install.sh"
PATH=/usr/obj/usr/src/tmp/legacy/usr/sbin:/usr/obj/usr/src/tmp/legacy/usr/bin:/usr/obj/usr/src/tmp/legacy/usr/games:/usr/obj/usr/src/tmp/legacy/bin:/sbin:/bin:/usr/sbin:/usr/bin
WORLDTMP=/usr/obj/usr/src/tmp  VERSION="FreeBSD 10.3-STABLE amd64 1003509"
MAKEFLAGS="-m /usr/src/tools/build/mk  -m /usr/src/share/mk"
COMPILER_TYPE=clang make  -f Makefile.inc1  DESTDIR=
BOOTSTRAPPING=1003509  SSP_CFLAGS=  -DWITHOUT_HTML -DWITHOUT_INFO -DNO_LINT
-DWITHOUT_MAN  -DNO_PIC -DNO_PROFILE -DNO_SHARED  _BOOTSTRAP_MAKEINFO=yes
-DNO_CPU_CFLAGS -DNO_WARNS -DNO_CTF -DEARLY_BUILD -DNO_TESTS legacy
===> tools/build (obj,includes,depend,all,install)
set -e; cd /usr/src/tools/build; make buildincludes; make installincludes
cc  -O2 -pipe  -I/usr/src/tools/build/../../contrib/libc-pwcache
-I/usr/src/tools/build/../../lib/libc/include -std=gnu99
-I/usr/obj/usr/src/tmp/legacy/usr/include -c
/usr/src/tools/build/../../contrib/libc-pwcache/pwcache.c -o pwcache.o
building static egacy library
ranlib -D libegacy.a
sh /usr/src/tools/install.sh -C -o root -g wheel -m 444   libegacy.a
/usr/obj/usr/src/tmp/legacy/usr/lib

--
>>> stage 1.2: bootstrap tools
--
cd /usr/src; MAKEOBJDIRPREFIX=/usr/obj/usr/src/tmp  INSTALL="sh
/usr/src/tools/install.sh"
PATH=/usr/obj/usr/src/tmp/legacy/usr/sbin:/usr/obj/usr/src/tmp/legacy/usr/bin:/usr/obj/usr/src/tmp/legacy/usr/games:/usr/obj/usr/src/tmp/legacy/bin:/sbin:/bin:/usr/sbin:/usr/bin
WORLDTMP=/usr/obj/usr/src/tmp  VERSION="FreeBSD 

Re: Recent -CURRENT unable to build numerous ports

[Dimitry Andric]

On 21 Oct 2016, at 20:47, Marcus von Appen  wrote:
> 
> -CURRENT as of r307731 seems to have some serious flaw with its build
> tool environment. Many ports fail to build with some error similar to
> the following (devel/apr1):
> 
> cc -emit-llvm -fno-strict-aliasing -pipe -march=native -DLIBICONV_PLUG
>  -fstack-protector -c -o .c.bco
> cc: error: no input files
> *** [.c.bco] Error code 1
> [...]
> 
> Since not all ports are affected, I assume some auto[conf|tool|make]
> issue to be triggered. Hints for getting this one solved are highly
> appreciated.

This was a side-effect of r307676, which added transformation rules for
.bco and .llo files (LLVM bytecode in binary and text representation).

Because .SUFFIXES was not updated to match, bmake was actually trying to
build a ".c.bco" file in the above case...

I committed a fix in r307754.

-Dimitry



signature.asc
Description: Message signed with OpenPGP using GPGMail

Re: Recent -CURRENT unable to build numerous ports

[Marcus von Appen]

On, Fri Oct 21, 2016, Marcus von Appen wrote:

> Hi,
>
> -CURRENT as of r307731 seems to have some serious flaw with its build
> tool environment. Many ports fail to build with some error similar to
> the following (devel/apr1):
>
> cc -emit-llvm -fno-strict-aliasing -pipe -march=native -DLIBICONV_PLUG
>   -fstack-protector -c -o .c.bco
> cc: error: no input files
> *** [.c.bco] Error code 1
> [...]
>
> Since not all ports are affected, I assume some auto[conf|tool|make]
> issue to be triggered. Hints for getting this one solved are highly
> appreciated.
>

After some more testing, it seems to relate to our default make.
Switching MAKE to gmake seems to resolve the issue.

A simple test case:
1. create an empty Makefile
2. run make

Cheers
Marcus

signature.asc
Description: PGP signature