head 1.19;
access;
symbols
RELEASE_7_4_0:1.11
RELEASE_8_2_0:1.11
RELEASE_6_EOL:1.10
RELEASE_8_1_0:1.9
RELEASE_7_3_0:1.8
RELEASE_8_0_0:1.5
RELEASE_7_2_0:1.4
RELEASE_7_1_0:1.4
RELEASE_6_4_0:1.4
RELEASE_5_EOL:1.2
RELEASE_7_0_0:1.2
RELEASE_6_3_0:1.2;
locks; strict;
comment @# @;
1.19
date 2011.04.24.16.37.51; author flz; state dead;
branches;
next 1.18;
1.18
date 2011.03.08.22.03.18; author linimon; state Exp;
branches;
next 1.17;
1.17
date 2011.03.04.06.11.51; author linimon; state Exp;
branches;
next 1.16;
1.16
date 2011.02.27.06.55.38; author linimon; state Exp;
branches;
next 1.15;
1.15
date 2011.02.27.06.16.01; author linimon; state Exp;
branches;
next 1.14;
1.14
date 2011.02.27.06.01.20; author linimon; state Exp;
branches;
next 1.13;
1.13
date 2011.02.24.22.28.22; author linimon; state Exp;
branches;
next 1.12;
1.12
date 2011.02.22.03.07.11; author linimon; state Exp;
branches;
next 1.11;
1.11
date 2011.01.24.00.44.53; author linimon; state Exp;
branches;
next 1.10;
1.10
date 2010.10.07.13.13.35; author linimon; state Exp;
branches;
next 1.9;
1.9
date 2010.03.30.21.14.16; author linimon; state Exp;
branches;
next 1.8;
1.8
date 2010.02.19.16.27.24; author linimon; state Exp;
branches;
next 1.7;
1.7
date 2010.02.05.18.09.52; author linimon; state Exp;
branches;
next 1.6;
1.6
date 2009.12.19.14.05.51; author linimon; state Exp;
branches;
next 1.5;
1.5
date 2009.08.24.02.32.52; author linimon; state Exp;
branches;
next 1.4;
1.4
date 2008.09.07.05.34.13; author linimon; state Exp;
branches;
next 1.3;
1.3
date 2008.07.02.08.44.20; author linimon; state Exp;
branches;
next 1.2;
1.2
date 2007.10.15.00.40.38; author linimon; state Exp;
branches;
next 1.1;
1.1
date 2007.10.15.00.20.29; author linimon; state Exp;
branches;
next ;
desc
@@
1.19
log
@Remove portbuild scripts from pcvs, as they now live in svn/projects.
@
text
@
FreeBSD Package building logs and errors
FreeBSD Package building logs and errors
FreeBSD maintains a build farm (the "pointyhat cluster") that is used
to build all packages across all supported architectures and major releases.
This page contains the build logs and errors for all the ports built by
the cluster. See the notes for additional information,
or the types of package errors detected.
Port cross-reference summaries
These live on
portsmon.FreeBSD.org.
Build errors
by portname;
by portname vs. build environment;
by error type vs. build environment- Problem Reports
for existing ports;
for new ports;
for the ports framework;
for repocopies requested;
for unknown
- Build Errors and
Problem Reports
by portname;
for one maintainer;
for broken ports;
for deprecated ports;
for forbidden ports
- Everything about
one port
- The
complete list of all reports
New build failures
Check here to find the most recent error log from your port.
-
New build
failures on 7.x-stable:
amd64
i386
ia64
sparc64
New build
failures on 8.x-stable:
amd64
i386
ia64
powerpc
sparc64
New build
failures on 9.x-current:
amd64
i386
ia64
powerpc
sparc64
New build failures on 7.x-stable with experimental port patches:
amd64
i386
New build failures on 8.x-stable with experimental port patches:
amd64
i386
New build failures on 9.x-current with experimental port patches:
amd64
i386
Error logs
-
Previous run
on 7.x-stable:
amd64
i386
ia64
sparc64
Current run
on 7.x-stable:
amd64
i386
ia64
sparc64
Previous run
on 8.x-stable:
amd64
i386
ia64
powerpc
sparc64
Current run
on 8.x-stable:
amd64
i386
ia64
powerpc
sparc64
Previous run
on 9.x-current:
amd64
i386
ia64
powerpc
sparc64
Current run
on 9.x-current:
amd64
i386
ia64
powerpc
sparc64
Previous run
on 7.x-stable with experimental port patches:
amd64
i386
Current run
on 7.x-stable with experimental port patches:
amd64
i386
Previous run
on 8.x-stable with experimental port patches:
amd64
i386
Current run
on 8.x-stable with experimental port patches:
amd64
i386
Previous run
on 9.x-current with experimental port patches:
amd64
i386
Current run
on 9.x-current with experimental port patches:
amd64
i386
Build logs (errors and otherwise)
-
Previous run
on 7.x-stable:
amd64
i386
ia64
sparc64
Current run on
7.x-stable:
amd64
i386
ia64
sparc64
Previous run
on 8.x-stable:
amd64
i386
ia64
powerpc
sparc64
Current run on
8.x-stable:
amd64
i386
ia64
powerpc
sparc64
Previous run
on 9.x-current:
amd64
i386
ia64
powerpc
sparc64
Current run on
9.x-current:
amd64
i386
ia64
powerpc
sparc64
Previous run
on 7.x-stable with experimental port patches:
amd64
i386
Current run on
7.x-stable with experimental port patches:
amd64
i386
Previous run
on 8.x-stable with experimental port patches:
amd64
i386
Current run on
8.x-stable with experimental port patches:
amd64
i386
Previous run
on 9.x-current with experimental port patches:
amd64
i386
Current run on
9.x-current with experimental port patches:
amd64
i386
Packages
-
Packages from
latest run on 6-stable:
amd64
i386
sparc64
Packages from
latest run on 7-stable:
amd64
i386
ia64
sparc64
Packages from
latest run on 8-stable:
amd64
i386
ia64
powerpc
sparc64
Packages from
latest run on 9-current:
amd64
i386
ia64
powerpc
sparc64
Package
building statistics (current state of all package builds)
Archive
All error
logs:
amd64
i386
ia64
powerpc
sparc64
All portbuild
logs:
amd64
i386
ia64
powerpc
sparc64
Additional information
All of the "Current run" links are possibly in progress and may be
partial, so keep that in mind if there appear to be some missing.
None of the ports marked IS_INTERACTIVE or
NO_PACKAGE are built any more -- if you have ports that
fall into those categories, assume their packages or distfiles will
never show up in ftp sites or CDROMs. RESTRICTED packages
are built and deleted (using "make
clean-restricted-list"). NO_CDROM packages are built
but deleted (using "make clean-cdrom-list") before being
put on a CDROM.
See also the types of errors detected.
Notes on the building process:
- Every port is built in its own chroot environment
, starting with an
empty /usr/local and /usr/X11R6. The dependencies are installed as packages just before
the build. You can see the list of dependencies on the third line
of the log -- the "foo.tgz bar.tgz" stuff are the dependencies. To
make sure that these actually work, DEPENDS_TARGET is set
to "/usr/bin/true"; if you see "/usr/bin/true is up to date" or
some such, that means there is something wrong with the dependency
lines or the packages this port is depending on.
- The build is done on a shared (read-only) /usr/ports
with
WRKDIRPREFIX set to /work. If your
WRKSRC looks funny, that's probably why.
- MASTER_SITE_OVERRIDE is pointing to pointyhat's distfile
dir, and MASTER_SITE_BACKUP is pointing to
ftp-master.freebsd.org or a local mirror.
Please pay attention to
Emanuel Haupt's distfile checker
for that.
- The ports are built on machines that are mostly running
9-CURRENT, with some 8.x-STABLE machines. If your port depends on
the result of uname(3) or sysctl to determine the running version of
FreeBSD, change it to use uname(1) instead (the builds use a dummy
uname(1) that reports the target version of FreeBSD), or change it to
use the value of the OSVERSION variable that can be passed in from
the port makefile.
Types of package errors detected
Here is the alphabetical list of current errors
detected by the AI script. Note that this is all just a rough guess --
it is merely for your aid.
Key:
- The most common
errors.
- Some less common
errors.
Transient
errors. These may not be your fault.
- arch
- The port does not build on a particular architecture, due to
assembler or linker errors. In some easy cases this is due to
not picking up the various ARCH configuration variables
in the Makefile; you'll see this via, e.g., a Sparc make
failing while looking for an i386 subdirectory. For the 64-bit
architectures, a common problem is the assumption many programmers
make that pointers may be cast to and from 32-bit ints. In other cases
the problems run much deeper, in which case ONLY_FOR_ARCHS
may be needed.
- autoconf
- Your port depends on autoconf, but the Makefile
either doesn't have USE_AUTOCONF, or does not use
USE_AUTOCONF_VER correctly.
- autoheader
- Your port depends on autoheader, but the Makefile
cannot find it; set USE_AUTOHEADER.
- automake
- Your port depends on automake, but the Makefile
either doesn't have USE_AUTOMAKE, or does not use
USE_AUTOMAKE_VER correctly.
- bad C++ code
- There is a compiler error which is caused by something specific
to C++.
- compiler error
- There is a C compiler error which is caused by something other
than e.g. "new compiler error".
- CATEGORIES
- The CATEGORIES line in Makefile includes an
invalid category.
- checksum
- The checksum of one or more of the files is incorrect.
- clang
- Your code does not run with the the experimental clang compiler. See
(TBA)
for further information.
- clang bug
- You have tickled a bug in clang itself. See
(TBA)
for further information.
- cluster
- There was some kind of transient error on the build cluster. It is not your
fault.
- compat6x
- This port needs to depend on a port misc/compat6x.
- configure error
- The port's configure script produced some kind of
error.
(Note: using clang as the ports compiler can also trigger this message.)
- coredump
- Some process in the build chain dropped core. While your port may indeed
be faulty, the process that dropped core should also be fixed.
- cpusetsize
- This port needs to catch up with the cpusetsize sysctl change in 9-CURRENT.
- depend object
- The port is trying to reinstall a dependency that already
exists. This is usually caused by the first field of a
*_DEPENDS line (the obj of
obj:dir[:target]) indicating a file that is not installed
by the dependency, causing it to be rebuilt even though it has
already been added from a package.
- depend package
- There was an error during adding dependencies
from packages. It is the fault of the package being added, not
this port.
- dirent
- The port has not caught up with the change to
scandir(3)
as committed in FreeBSD version 800501.
- disk full
- The disk filled up on the build system. It is not your
fault.
- DISPLAY
- This port requires an X display to build. There is nothing you
can do about it unless you can somehow make it not require an X
connection.
- distinfo update
- The contents of distinfo does not match the list of
distfiles or patchfiles.
- fetch
- One or more of the files could not be fetched.
- fetch timeout
- Your fetch process was killed because it took too long. (More
accurately, it did not produce any output for a long time.) Please
put sites with better connectivity near the beginning of
MASTER_SITES.
- forbidden
- Someone has marked this port as "forbidden", almost always due
to security concerns. See the logfile for more information.
- gcc bug
- You have tickled a bug in gcc itself. See the
GNU bug report documentation
for further information.
- gcc4
- Your code does not run with the latest gcc version
See the wiki page
for further information.
- gmake
- Your code does not run with the latest, incompatible, gmake version
(3.82.)
- install error
- There was an error during installation.
- LIB_DEPENDS
- The LIB_DEPENDS line specifies a library name
incorrectly. This often happens when a port is upgraded and the
shared library version number changes.
- linker error
- There is a linker error which is caused by something other than
those flagged elsewhere.
(Note: using clang as the ports compiler can also trigger this message.)
- makefile
- There is an error in the Makefile, possibly in the default
targets.
- manpage
- There is a manpage listed in a MAN? macro that does not
exist or is not installed in the right place.
- missing header
- There is a missing header file. This is usually caused by
either (1) a missing dependency, or (2) specifying an incorrect
location with -I in the compiler command line.
- mtree
- The port leaves ${PREFIX} in a state that is not
consistent with the mtree definition after pkg_delete. This
usually means some files are missing from PLIST. It could
also mean that your installation scripts create files or
directories not properly deleted by the deinstallation scripts.
Another possibility is that your port is deleting some directories
it is not supposed to, or incorrectly modifying some directory's
permission.
- nested_declaration
- There is a nested declaration in the source code.
- new compiler error
- The newest version of gcc in the base does not like the source code. This is
usually due to stricter C++ type checking or changes in register
allocation policy.
- NFS
- There was either a temporary NFS error on the build system
(which is not your fault), or the WRKSRC is invalid
(which is your fault).
- patch
- One or more of the patches failed.
- perl
- perl is no longer included by default in the base
system, but your port's configuration process depends on it. While
this change helps avoid having a stale version of perl
in the base system, it also means that many ports now need to include
USE_PERL5.
- PLIST
- There is a missing item in the PLIST. Note that this is
often caused by an earlier error that went undetected. In this case,
you should fix the error and also the build process so it will fail
upon an error instead of continuing, since that makes debugging
that much harder.
- process failed
- The make process terminated unexpectedly, due to
something like a signal 11 or bus error.
- runaway process
- Your make package process was killed because it took
too long. (More accurately, it did not produce any output for a long
time.) It is probably because there is a process spinning in an infinite
loop. Please check the log to determine the exact cause of the
problem.
- segfault
- Some process in the build chain got a segmentation fault.
- sem_wait
- This port needs to catch up with semaphore changes in 9-CURRENT.
- termios
- This port needs to catch up with the termios.h changes in src.
- threads
- There is a linker error which is caused by failing to find one of
the thread libraries.
- truncated_distfile
- A package node encountered an error during pkg_add. It is not your
fault. Linimon is trying to figure out this problem.
- utmp_x
- This port needs to catch up with the utmp_x.h changes in src.
- WRKDIR
- The port is attempting to change something outside
${WRKDIR}. See handbook
for details.
- ??? (unknown)
- The automated script cannot even guess what is wrong with your
port. portmgr tries to keep the processonelog script
reasonably efficient while covering as many errors as possible, but many
errors are not common enough to try to catch.
Here is an alphabetical list of obsolete errors
that used to be detected by the AI script, but are now uncommon enough to
be skipped:
- alignment
- You've managed to confuse the assembler with a misaligned
structure.
- apxs
- Your port depends on Apache (in particular, the apxs
binary) but the Makefile doesn't have Apache in
BUILD_DEPENDS and/or LIB_DEPENDS.
- assert
- Compilation failed due to an assert. This is often a variation
on arch or missing header.
- awk
- awk is complaining about some kind of bogus string
expression.
- bison
- Your port requires bison, which does not exist in
4.x-stable or newer anymore. Either patch it to use byacc
instead, or define USE_BISON.
- cgi-bin
- Your port assumes that a directory (usually
/usr/local/www/cgi-bin) already exists,
but by default it doesn't.
- chown
- POSIX has deprecated the usage
"chown user.group filename" in favor of
"chown user:group filename". This happened quite some time
ago, actually, but it is only now being enforced. (The change was
made to allow '.' in usernames).
- ELF
- The port does not properly work in the new ELF world. It is
probably looking for an a.out object (e.g.,
crt0.o).
- f77
- gcc in base no longer includes the Fortran compiler
by default.
- ffs conflict
- Both /usr/include/machine/cpufunc.h and
/usr/include/strings.h are attempting to define int ffs().
- getopt.h
- <getopt.h> is conflicting with unistd.h.
-
getopt
- Your port may need to set the new port variable
USE_GETOPT_LONG.
- libc_r not found
- This library has not yet been ported to e.g. the Sparc.
- malloc.h
- Including <malloc.h> is now deprecated in favor of
<stdlib.h>.
- MOTIF
- This port requires Motif but does not define
REQUIRES_MOTIF. See the handbook
for details.
- MOTIFLIB
- This port requires Motif but does not refer to the libraries
using ${MOTIFLIB}. See handbook
for details.
- munmap
- POSIX specifies that munmap cannot be called on a zero-length file.
Because of this, during 4.X builds, if cp tries to copy a zero-length file, it
may fail saying, "cp: ...: Invalid argument". This is a problem with the
bindist image on pointyhat, and not the fault of the porter.
- perl5
- There is a problem in processing a perl5 module.
- pod2man
- perl is no longer included by default in the base
system, but your port's documentation process depends on it. While
this change helps avoid having a stale version of perl
in the base system, it also means that many ports now need to include
USE_PERL5.
- python
- The Makefile needs to define USE_PYTHON.
- sed
- sed is complaining about some kind of bogus regular
expression, probably as a side-effect of its being invoked by
${REINPLACE_COMMAND}. This is often a result of having
replaced usages of perl in the Makefile with usages
of ${REINPLACE_COMMAND} but having left
perl-specific regexps in place.
- soundcard.h
- machine/soundcard.h has been moved.
- stdio
- You need to bring your port up to date with the current
<stdio.h>.
- stl
- Your port requires the STL library but cannot find it.
- struct changes
- Your port is trying to refer to structure elements that are not
really there. This is often due to changes in the underlying
include files.
- sysvipc
- Your port is interacting badly with the System V InterProcess
Communication code.
- uname
- For a short period of time, gcc was not handling uname properly.
- values.h
- values.h has been moved.
- texinfo
- The new makeinfo cannot process a texinfo source file.
You can probably add a "--no-validate" option to force it
through if you are sure it's correct regardless of what
makeinfo says.
- union wait
- The compiler could not calculate the storage size of an object,
often due to misuse of a union.
- USE_XLIB
- You should specify USE_XLIB for this port since it
appears to use X.
- varargs
- varargs.h is obsolete with newer versions of gcc.
- X manpage
- This port does not install a manpage but imake rules are
generating commands to convert manpages to HTML format. This is
most likely fixed by changing ComplexProgramTarget() in
Imakefile to ComplexProgramTargetNoMan(). Note that
defining NO_INSTALL_MANPAGES in the Makefile is no
longer sufficient in XFree86-4.
The ports page
| Maintained by portmgr@@FreeBSD.org
| Last modified
$FreeBSD: ports/Tools/portbuild/errorlogs/index.shtml,v 1.18 2011/03/08 22:03:18 linimon Exp $