Discussion:
trying to build Owl for Raspberry Pi: rpmquery --specfile
c***@openwall.com
2014-04-27 12:36:55 UTC
Permalink
Colleagues,

I'm trying to create an Owl distro version for Raspberry Pi. I decided to
begin with the Fedora for Raspberry (that thing which is named 'Pidora',
heh... sounds _very_, errrr, /funny/ for russian-speaking people). So I
created a SD-card with the Pidora, set it up so it boots, scp'ed the
image of Owl-current-20130408-i686.iso to that system just to have all the
sources handy (not much of other use for that file though, because
Raspberry is ARM)

I then created that group 'build' with gid=150, the user 'build' wigh
uid=151 and $home=/usr/src/world, copied the 'world' directory from the CD
there, and finally issued 'make buildworld' command. Here's what I got:

% ***@raspi ~ $ make buildworld
% test -d native/Owl/build || make checkout
% ln -s native/Owl/build/{*.conf,.rpm*} . 2> /dev/null || :
% BRANCH=Owl PACKAGE= native/Owl/build/buildworld.sh
% 01:56:38: Detecting version of RPM
% 01:56:39: It's RPM4, using different binaries for different tasks
% 01:56:39: Removing stale temporary files
% 01:56:39: Sanity check
% 01:56:39: #1: Scanning native
% rpmquery: --specfile: unknown option
% 01:56:39: #1: Skipping SimplePAMApps
% rpmquery: --specfile: unknown option
% [...]


% ***@raspi ~ $ rpmquery --version
% RPM version 4.9.1.3


Now I'm a bit confused where to go. The rpm in Owl seems to be older
(4.2), so it could be that the rpmquery's '--specfile' option is
implemented in the Owl's patches, but 'grep'ing them for the word
'specfile' didn't help, it's not there.

Anyway, could someone please provide any glue how do I build may be a
single package (rpm) from the Owl sources and what I do afterwards --
perhaps I'll be unable to replace it within my system without some tricks,
but, okay, it is still possible that replacing a single executable
(rpmquery) will help. Or may be it is possible to workaround this somehow,
patching the build scripts to not use the option?


Thanks!

--
Andrey "Croco" Stolyarov
Dmitry V. Levin
2014-04-27 12:58:49 UTC
Permalink
Post by c***@openwall.com
Colleagues,
I'm trying to create an Owl distro version for Raspberry Pi. I decided to
begin with the Fedora for Raspberry (that thing which is named 'Pidora',
heh... sounds _very_, errrr, /funny/ for russian-speaking people). So I
created a SD-card with the Pidora, set it up so it boots, scp'ed the
image of Owl-current-20130408-i686.iso to that system just to have all the
sources handy (not much of other use for that file though, because
Raspberry is ARM)
I then created that group 'build' with gid=150, the user 'build' wigh
uid=151 and $home=/usr/src/world, copied the 'world' directory from the CD
% test -d native/Owl/build || make checkout
% BRANCH=Owl PACKAGE= native/Owl/build/buildworld.sh
% 01:56:38: Detecting version of RPM
% 01:56:39: It's RPM4, using different binaries for different tasks
% 01:56:39: Removing stale temporary files
% 01:56:39: Sanity check
% 01:56:39: #1: Scanning native
% rpmquery: --specfile: unknown option
% 01:56:39: #1: Skipping SimplePAMApps
% rpmquery: --specfile: unknown option
% [...]
% RPM version 4.9.1.3
Now I'm a bit confused where to go. The rpm in Owl seems to be older
(4.2), so it could be that the rpmquery's '--specfile' option is
implemented in the Owl's patches, but 'grep'ing them for the word
'specfile' didn't help, it's not there.
rpmquery used to support --specfile option, but in rpm >= 4.9.0
this functionality was moved to a new tool called rpmspec.
"rpm --specfile" is aliased there to "rpmspec -q", but
there is no similar alias for "rpmquery --specfile":
$ grep -Fe --specfile /usr/lib/rpm/rpmpopt
rpm exec --specfile rpmspec -q

You can add one and see where the build stops next.
--
ldv
c***@openwall.com
2014-04-27 21:24:41 UTC
Permalink
Post by Dmitry V. Levin
rpmquery used to support --specfile option, but in rpm >= 4.9.0
this functionality was moved to a new tool called rpmspec.
"rpm --specfile" is aliased there to "rpmspec -q", but
$ grep -Fe --specfile /usr/lib/rpm/rpmpopt
rpm exec --specfile rpmspec -q
You can add one and see where the build stops next.
Well, I'm not that familiar with rpm, and as of now I'm totally confused
with the file /usr/lib/rpm/rpmpopt-4.9.1.3 (is this the one I need?)
WHAT (I mean, _which_ "one") should I add, and where? BTW, the file in
question is commented as "shoud not be modified", and there's a note that
"local changes should go to /etc/popd, not here"; well, there's an empty
directory /etc/popd.d, no other files there and no idea how to deal with
all these mess.

Anyway, in such a situation, may be it could be wiser to patch Owl build
scripts instead of trying to reconfigure rpm? Depending on a feature
which is known to disappear doesn't look a good idea to me.


Thanks!


--
Croco
Dmitry V. Levin
2014-04-27 21:46:46 UTC
Permalink
Post by c***@openwall.com
Post by Dmitry V. Levin
rpmquery used to support --specfile option, but in rpm >= 4.9.0
this functionality was moved to a new tool called rpmspec.
"rpm --specfile" is aliased there to "rpmspec -q", but
$ grep -Fe --specfile /usr/lib/rpm/rpmpopt
rpm exec --specfile rpmspec -q
You can add one and see where the build stops next.
Well, I'm not that familiar with rpm, and as of now I'm totally confused
with the file /usr/lib/rpm/rpmpopt-4.9.1.3 (is this the one I need?)
WHAT (I mean, _which_ "one") should I add, and where? BTW, the file in
question is commented as "shoud not be modified", and there's a note that
"local changes should go to /etc/popd, not here"; well, there's an empty
directory /etc/popd.d, no other files there and no idea how to deal with
all these mess.
Anyway, in such a situation, may be it could be wiser to patch Owl build
scripts instead of trying to reconfigure rpm? Depending on a feature
which is known to disappear doesn't look a good idea to me.
buildworld.sh already has some logic around rpm/rpmquery: it defines RPMQ
variable to "rpm" by default, and changes it to "rpmquery" for rpm 4.x.
You can comment out that "RPMQ=rpmquery" expression in setup_rpm().
--
ldv
c***@openwall.com
2014-05-11 17:03:53 UTC
Permalink
Post by Dmitry V. Levin
buildworld.sh already has some logic around rpm/rpmquery: it defines RPMQ
variable to "rpm" by default, and changes it to "rpmquery" for rpm 4.x.
You can comment out that "RPMQ=rpmquery" expression in setup_rpm().
Thanks!

Just to document the process, after I did that, I had:

# rpm: /usr/bin/rpmspec: No such file or directory

Investigating the problem I noticed that it is 'rpm --specfile' who says
this, and the problem is solved with 'yum install rpm-build.armv6hl' which
provided me with that rpmspec binary (this is Pidora issue, not related to
Owl, and it may be a feature rather than a bug -- but I think it is not bad
to have a memo within the list's archive).

Anyway, now it looks like this:

# ***@raspi ~$ make buildworld
# test -d native/Owl/build || make checkout
# ln -s native/Owl/build/{*.conf,.rpm*} . 2> /dev/null || :
# BRANCH=Owl PACKAGE= native/Owl/build/buildworld.sh
# 00:06:54: Detecting version of RPM
# 00:06:54: It's RPM4, using different binaries for different tasks
# 00:06:54: Removing stale temporary files
# 00:06:54: Sanity check
# 00:06:54: #1: Scanning native
# warning: line 25: prereq is deprecated: PreReq: owl-control >= 0.4,
# owl-control < 2.0
# 00:06:55: #1: Building SimplePAMApps
# 00:06:55: #1: Failed SimplePAMApps
# 00:06:56: #1: Building SysVinit
# 00:07:25: #1: Failed SysVinit
# warning: line 22: prereq is deprecated: PreReq: /sbin/install-info, grep,
# coreutils >= 5.3.0, sed >= 4.0.9
# 00:07:25: #1: Building acct
# 00:07:26: #1: Failed acct
# warning: line 17: prereq is deprecated: PreReq: /sbin/install-info
# warning: line 17: prereq is deprecated: PreReq: /sbin/install-info
# 00:07:26: #1: Building autoconf
# 00:07:27: #1: Failed autoconf

and so on. Well, as far as I can tell, all this is described in BUILD.txt
with a short phrase <<Building on another GNU/*/Linux system is sometimes
possible, but tricky and requires that you slowly produce an Owl system
with multiple iterations of "make buildworld" and "make installworld".>>
So, any help with further tricks of that 'tricky building'? I'd love more
documentation to read, but it seems that none exists.


Thanks!

--
Croco
Vasily Kulikov
2014-05-11 17:14:08 UTC
Permalink
Post by c***@openwall.com
# test -d native/Owl/build || make checkout
# BRANCH=Owl PACKAGE= native/Owl/build/buildworld.sh
# 00:06:54: Detecting version of RPM
# 00:06:54: It's RPM4, using different binaries for different tasks
# 00:06:54: Removing stale temporary files
# 00:06:54: Sanity check
# 00:06:54: #1: Scanning native
# warning: line 25: prereq is deprecated: PreReq: owl-control >= 0.4,
# owl-control < 2.0
# 00:06:55: #1: Building SimplePAMApps
# 00:06:55: #1: Failed SimplePAMApps
# 00:06:56: #1: Building SysVinit
# 00:07:25: #1: Failed SysVinit
# warning: line 22: prereq is deprecated: PreReq: /sbin/install-info, grep,
# coreutils >= 5.3.0, sed >= 4.0.9
# 00:07:25: #1: Building acct
# 00:07:26: #1: Failed acct
# warning: line 17: prereq is deprecated: PreReq: /sbin/install-info
# warning: line 17: prereq is deprecated: PreReq: /sbin/install-info
# 00:07:26: #1: Building autoconf
# 00:07:27: #1: Failed autoconf
and so on. Well, as far as I can tell, all this is described in BUILD.txt
with a short phrase <<Building on another GNU/*/Linux system is sometimes
possible, but tricky and requires that you slowly produce an Owl system
with multiple iterations of "make buildworld" and "make installworld".>>
So, any help with further tricks of that 'tricky building'? I'd love more
documentation to read, but it seems that none exists.
Please post several log files from /usr/src/world/logs/$package (the
last lines from it with 'failed' or 'error' strings of compiler/building
scripts output). It is not obvious what's missing/broken for your arch.

Also several packages have scripts for specific architectures only.
In this case you should investigate whether the package or building
scripts have arch-specific things and try to fix it. Probably simple
addition of your arch into *Arch might fix it.

$ grep ExclusiveArch native/Owl/packages/*/*spec
native/Owl/packages/cdrkit/cdrkit.spec:ExclusiveArch: %ix86 x86_64
native/Owl/packages/dev86/dev86.spec:ExclusiveArch: %ix86 x86_64
native/Owl/packages/dmidecode/dmidecode.spec:ExclusiveArch: %ix86 x86_64
native/Owl/packages/elftoaout/elftoaout.spec:ExclusiveArch: sparc sparcv9 sparc64
native/Owl/packages/kernel/kernel.spec:ExclusiveArch: i686 x86_64
native/Owl/packages/lilo/lilo.spec:ExclusiveArch: %ix86 x86_64
native/Owl/packages/ltrace/ltrace.spec:ExclusiveArch: %ix86 x86_64 sparc sparcv9
native/Owl/packages/owl-cdrom/owl-cdrom.spec:ExclusiveArch: %ix86 x86_64
native/Owl/packages/prtconf/prtconf.spec:ExclusiveArch: sparc sparcv9 sparc64
native/Owl/packages/setarch/setarch.spec:ExclusiveArch: %ix86 x86_64 sparc sparcv9 sparc64 ppc ppc64 mips mips64 ia64 s390 s390x
native/Owl/packages/silo/silo.spec:ExclusiveArch: sparc sparcv9 sparc64
native/Owl/packages/syslinux/syslinux.spec:ExclusiveArch: %ix86 x86_64
--
Vasily
c***@openwall.com
2014-05-25 08:07:48 UTC
Permalink
Vasily, All,
Post by Vasily Kulikov
Please post several log files from /usr/src/world/logs/$package (the
last lines from it with 'failed' or 'error' strings of compiler/building
scripts output). It is not obvious what's missing/broken for your arch.
Oh, actually speaking, I simply didn't know about these files. Errr... I
strongly believe that this worth to be mentioned in the BUILD.txt file.

With the new knowledge, I was able to install some packages (such as
texinfo and libtool, that appear to be build-deps for some of the Owl
packages) into my Raspberry, and this allowed me to proceed a little
further. I'm still playing with installing more previously-failed build
dependencies; Raspberry Pi is *S*L*O*W* so I'm still unable to give a
result summary :)

However, another problem appeared which perhaps I'm unable to fix myself.
Some Owl patches failed to apply:

# ***@raspi /usr/src/world/logs# grep ^Hunk * | grep FAILED
# acct:Hunk #1 FAILED at 5.
# bash:Hunk #2 FAILED at 357.
# ed:Hunk #1 FAILED at 8.
# flex:Hunk #1 FAILED at 454.
# gdb:Hunk #1 FAILED at 1205.
# glibc:Hunk #1 FAILED at 1075.
# iputils:Hunk #1 FAILED at 31.
# man-pages:Hunk #3 FAILED at 169.
# man-pages:Hunk #4 FAILED at 330.
# man-pages:Hunk #2 FAILED at 146.
# procps:Hunk #3 FAILED at 854.
# tcp_wrappers:Hunk #1 FAILED at 28.
# time:Hunk #1 FAILED at 1.
# vixie-cron:Hunk #1 FAILED at 1.
# ***@raspi /usr/src/world/logs#

# ***@raspi /usr/src/world/logs# patch --version
# patch 2.6.1
# Copyright (C) 1988 Larry Wall
# Copyright (C) 2003, 2009 Free Software Foundation, Inc.
[...]

I'm attaching a file with more verbose log excerpts, that include the patch
numbers. Could someone please look into this? This is a bit surprising
for me, as I'm used to consider the diff/patch couple as ascii text
manipulation utilities that should not depend on the architecture. May be
the version of patch is the problem? However, if this is the case, I'd say
it could be better to provide patches that are good for more versions of
the utility; is there a chance for such a solution?


Also, building any particular package with 'make PACKAGE=...', I get the
following:

# ***@raspi ~$ make PACKAGE=owl-setup
[...]
# Usage: fgrep [OPTION]... PATTERN [FILE]...
# Try 'fgrep --help' for more information.

# ***@raspi ~$ fgrep --version
# fgrep (GNU grep) 2.13
# Copyright (C) 2012 Free Software Foundation, Inc.
[...]

Not yet understood what particular fgrep option triggers this. The funny
thing is that the build actually continues after these messages, and in
some cases it successfully completes.
Post by Vasily Kulikov
Also several packages have scripts for specific architectures only.
In this case you should investigate whether the package or building
scripts have arch-specific things and try to fix it. Probably simple
addition of your arch into *Arch might fix it.
$ grep ExclusiveArch native/Owl/packages/*/*spec
native/Owl/packages/cdrkit/cdrkit.spec:ExclusiveArch: %ix86 x86_64
native/Owl/packages/dev86/dev86.spec:ExclusiveArch: %ix86 x86_64
native/Owl/packages/dmidecode/dmidecode.spec:ExclusiveArch: %ix86 x86_64
native/Owl/packages/elftoaout/elftoaout.spec:ExclusiveArch: sparc sparcv9 sparc64
native/Owl/packages/kernel/kernel.spec:ExclusiveArch: i686 x86_64
native/Owl/packages/lilo/lilo.spec:ExclusiveArch: %ix86 x86_64
native/Owl/packages/ltrace/ltrace.spec:ExclusiveArch: %ix86 x86_64 sparc sparcv9
native/Owl/packages/owl-cdrom/owl-cdrom.spec:ExclusiveArch: %ix86 x86_64
native/Owl/packages/prtconf/prtconf.spec:ExclusiveArch: sparc sparcv9 sparc64
native/Owl/packages/setarch/setarch.spec:ExclusiveArch: %ix86 x86_64 sparc sparcv9 sparc64 ppc ppc64 mips mips64 ia64 s390 s390x
native/Owl/packages/silo/silo.spec:ExclusiveArch: sparc sparcv9 sparc64
native/Owl/packages/syslinux/syslinux.spec:ExclusiveArch: %ix86 x86_64
Some of these (e.g. lilo (?)) are simply unneeded on Raspberry; as of now,
I'm going to try without them all.

Anyway, during all the builds I see '-march=armv6' on the gcc's command
line, and 'armv6hl' suffix in Pidora's package names. What should I add to
ExclusiveArch? armv6? armv6hl? Both?



Thanks!


--
Croco
c***@openwall.com
2014-05-25 08:17:10 UTC
Permalink
Post by c***@openwall.com
I'm attaching a file with more verbose log excerpts, that include the patch
Err sorry forgot the file. Here is it:

acct-+ echo 'Patch #3 (acct-6.4pre1-owl-info.diff):'
acct-Patch #3 (acct-6.4pre1-owl-info.diff):
acct-+ /usr/bin/patch -p1 --fuzz=0
acct-+ /usr/bin/cat /usr/src/world/rpm-work-1/SOURCES/acct-6.4pre1-owl-info.diff
acct-patching file accounting.texi
acct:Hunk #1 FAILED at 5.
--
bash-+ echo 'Patch #0 (bash-3.1-up-patchlevel.diff):'
bash-Patch #0 (bash-3.1-up-patchlevel.diff):
bash-+ /usr/bin/patch -p0 --fuzz=0
bash-+ /usr/bin/cat /usr/src/world/rpm-work-1/SOURCES/bash-3.1-up-patchlevel.diff
bash-patching file parse.y
bash-patching file patchlevel.h
bash-patching file doc/bash.1
bash-patching file doc/bashref.texi
bash-patching file doc/version.texi
bash-patching file patchlevel.h
bash-patching file variables.c
bash-patching file patchlevel.h
bash-patching file subst.c
bash-patching file patchlevel.h
bash-patching file subst.c
bash-patching file patchlevel.h
bash-patching file lib/readline/terminal.c
bash-patching file patchlevel.h
bash-patching file jobs.c
bash-patching file patchlevel.h
bash-patching file jobs.c
bash-Hunk #1 succeeded at 2234 (offset 35 lines).
bash-patching file patchlevel.h
bash-patching file jobs.c
bash-Hunk #2 succeeded at 1694 (offset 35 lines).
bash-patching file patchlevel.h
bash-patching file parse.y
bash-patching file patchlevel.h
bash-patching file parse.y
bash-Hunk #1 succeeded at 2919 (offset 11 lines).
bash-patching file patchlevel.h
bash-patching file parse.y
bash-Hunk #1 succeeded at 3590 (offset 11 lines).
bash-patching file patchlevel.h
bash-patching file lib/readline/readline.c
bash-patching file patchlevel.h
bash-patching file lib/readline/display.c
bash-patching file patchlevel.h
bash-patching file lib/glob/sm_loop.c
bash-patching file patchlevel.h
bash-patching file lib/glob/glob.c
bash-patching file patchlevel.h
bash-patching file subst.c
bash-patching file subst.h
bash-patching file arrayfunc.c
bash-patching file patchlevel.h
bash-+ echo 'Patch #10 (bash-3.1-owl-warnings.diff):'
bash-Patch #10 (bash-3.1-owl-warnings.diff):
bash-+ /usr/bin/patch -p1 --fuzz=0
bash-+ /usr/bin/cat /usr/src/world/rpm-work-1/SOURCES/bash-3.1-owl-warnings.diff
bash-patching file bashhist.c
bash-patching file bashline.c
bash-patching file braces.c
bash-patching file builtins/bind.def
bash-patching file builtins/caller.def
bash-patching file builtins/cd.def
bash-patching file builtins/common.c
bash-patching file builtins/declare.def
bash-patching file builtins/echo.def
bash-patching file builtins/enable.def
bash-patching file builtins/evalstring.c
bash-patching file builtins/fg_bg.def
bash-patching file builtins/getopt.c
bash-patching file builtins/getopts.def
bash-patching file builtins/help.def
bash-patching file builtins/history.def
bash-patching file builtins/printf.def
bash-patching file builtins/set.def
bash-patching file builtins/setattr.def
bash-patching file builtins/trap.def
bash-patching file builtins/type.def
bash-patching file builtins/ulimit.def
bash-patching file error.c
bash-patching file eval.c
bash-patching file execute_cmd.c
bash-patching file expr.c
bash-patching file findcmd.c
bash-patching file general.c
bash-patching file hashlib.c
bash-patching file jobs.c
bash-Hunk #1 succeeded at 746 (offset 3 lines).
bash-Hunk #2 succeeded at 934 (offset 38 lines).
bash-Hunk #3 succeeded at 995 (offset 38 lines).
bash-patching file jobs.h
bash-patching file lib/glob/glob.c
bash:Hunk #2 FAILED at 357.
--
ed-+ echo 'Patch #0 (ed-1.5-owl-alt-progname.diff):'
ed-Patch #0 (ed-1.5-owl-alt-progname.diff):
ed-+ /usr/bin/patch -p1 --fuzz=0
ed-+ /usr/bin/cat /usr/src/world/rpm-work-1/SOURCES/ed-1.5-owl-alt-progname.diff
ed-patching file main.c
ed-+ echo 'Patch #1 (ed-1.4-alt-owl-info.diff):'
ed-Patch #1 (ed-1.4-alt-owl-info.diff):
ed-+ /usr/bin/patch -p1 --fuzz=0
ed-+ /usr/bin/cat /usr/src/world/rpm-work-1/SOURCES/ed-1.4-alt-owl-info.diff
ed-patching file doc/ed.texinfo
ed:Hunk #1 FAILED at 8.
--
flex-+ echo 'Patch #0 (flex-2.5.4a-rh-skel.diff):'
flex-Patch #0 (flex-2.5.4a-rh-skel.diff):
flex-+ /usr/bin/patch -p1 --fuzz=0
flex-+ /usr/bin/cat /usr/src/world/rpm-work-1/SOURCES/flex-2.5.4a-rh-skel.diff
flex-patching file skel.c
flex:Hunk #1 FAILED at 454.
--
gdb-+ echo 'Patch #11 (gdb-6.3-owl-desc_bounds.diff):'
gdb-Patch #11 (gdb-6.3-owl-desc_bounds.diff):
gdb-+ /usr/bin/patch -p1 --fuzz=0
gdb-+ /usr/bin/cat /usr/src/world/rpm-work-1/SOURCES/gdb-6.3-owl-desc_bounds.diff
gdb-patching file gdb/ada-lang.c
gdb:Hunk #1 FAILED at 1205.
--
glibc-Patch #2 (glibc-2.3.6-cvs-20060103-ctermid.diff):
glibc-+ /usr/bin/patch -p0 --fuzz=0
glibc-+ /usr/bin/cat /usr/src/world/rpm-work-1/SOURCES/glibc-2.3.6-cvs-20060103-ctermid.diff
glibc-patching file posix/unistd.h
glibc:Hunk #1 FAILED at 1075.
--
iputils-+ echo 'Patch #4 (iputils-20001007-rh-bug23844.diff):'
iputils-Patch #4 (iputils-20001007-rh-bug23844.diff):
iputils-+ /usr/bin/patch -p1 --fuzz=0
iputils-+ /usr/bin/cat /usr/src/world/rpm-work-1/SOURCES/iputils-20001007-rh-bug23844.diff
iputils-patching file rarpd.c
iputils-patching file Modules/pg3.c
iputils-Hunk #1 succeeded at 89 (offset 12 lines).
iputils-patching file rdisc.c
iputils:Hunk #1 FAILED at 31.
--
man-pages-+ echo 'Patch #0 (man-pages-3.30-deb-owl-misc.diff):'
man-pages-Patch #0 (man-pages-3.30-deb-owl-misc.diff):
man-pages-+ /usr/bin/patch -p1 --fuzz=0
man-pages-+ /usr/bin/cat /usr/src/world/rpm-work-1/SOURCES/man-pages-3.30-deb-owl-misc.diff
man-pages-patching file man3/dbopen.3
man-pages-patching file man3/fgetpwent.3
man-pages-patching file man3/getpw.3
man-pages-patching file man3/getpwent.3
man-pages:Hunk #3 FAILED at 169.
man-pages-1 out of 3 hunks FAILED -- saving rejects to file man3/getpwent.3.rej
man-pages-patching file man3/getpwent_r.3
man-pages-patching file man3/getpwnam.3
man-pages:Hunk #4 FAILED at 330.
man-pages-1 out of 4 hunks FAILED -- saving rejects to file man3/getpwnam.3.rej
man-pages-patching file man3/nl_langinfo.3
man-pages-patching file man3/putpwent.3
man-pages-patching file man3/qsort.3
man-pages-patching file man3/tsearch.3
man-pages-patching file man4/console_codes.4
man-pages-patching file man5/dir_colors.5
man-pages-patching file man5/nsswitch.conf.5
man-pages-patching file man5/resolv.conf.5
man-pages-patching file man5/tzfile.5
man-pages:Hunk #2 FAILED at 146.
--
procps-+ echo 'Patch #24 (procps-3.2.5-alt-proc-sbuf.diff):'
procps-Patch #24 (procps-3.2.5-alt-proc-sbuf.diff):
procps-+ /usr/bin/patch -p1 --fuzz=0
procps-+ /usr/bin/cat /usr/src/world/rpm-work-1/SOURCES/procps-3.2.5-alt-proc-sbuf.diff
procps-patching file proc/readproc.c
procps-Hunk #1 succeeded at 491 (offset -1 lines).
procps-Hunk #2 succeeded at 568 (offset -1 lines).
procps:Hunk #3 FAILED at 854.
--
tcp_wrappers-Patch #5 (tcp_wrappers_7.6-alt-fix_options.diff):
tcp_wrappers-+ /usr/bin/patch -p1 --fuzz=0
tcp_wrappers-+ /usr/bin/cat /usr/src/world/rpm-work-1/SOURCES/tcp_wrappers_7.6-alt-fix_options.diff
tcp_wrappers-patching file fix_options.c
tcp_wrappers:Hunk #1 FAILED at 28.
--
time-+ echo 'Patch #2 (time-1.7-deb-info_quiet.diff):'
time-Patch #2 (time-1.7-deb-info_quiet.diff):
time-+ /usr/bin/patch -p1 --fuzz=0
time-+ /usr/bin/cat /usr/src/world/rpm-work-1/SOURCES/time-1.7-deb-info_quiet.diff
time-patching file version.texi
time-patching file time.texi
time-patching file time.info
time:Hunk #1 FAILED at 1.
--
vixie-cron-Patch #5 (vixie-cron-4.1.20040916-alt-owl-Makefile.diff):
vixie-cron-+ /usr/bin/patch -p1 --fuzz=0
vixie-cron-+ /usr/bin/cat /usr/src/world/rpm-work-1/SOURCES/vixie-cron-4.1.20040916-alt-owl-Makefile.diff
vixie-cron-patching file linux.mk
vixie-cron-patching file usr.bin/at/Makefile
vixie-cron-patching file usr.bin/crontab/Makefile
vixie-cron:Hunk #1 FAILED at 1.
Pavel Kankovsky
2014-06-08 10:17:44 UTC
Permalink
Post by c***@openwall.com
Anyway, during all the builds I see '-march=armv6' on the gcc's command
line, and 'armv6hl' suffix in Pidora's package names. What should I add to
ExclusiveArch? armv6? armv6hl? Both?
As far I know, "armv6" stands for ARMv6 architecture, "h" stands for
hard-float (hardware FPU; it is optional for v<7), and "l" probably
stands for little-endian (hardware supports both kinds of endianness).

Machine code built for one variant can run on some but not all other
variants (this includes both hardware and runtime environment). In this
sense, "armv6" is not a single architecture but a family of similar but
different and often incompatible architectures.

I think you should call the architecture "armv6hl" (assuming you want to
stick to tools & nomenclature of Fedora et al.) and disregard the
existence of other variants, at least for the time being.
--
Pavel Kankovsky aka Peak "Que sais-je?"
Loading...