qemu fails to build with texinfo 5.0
as i don't need the docs, i just applied the following sed and have been
able to build qemu-kvm 1.2.0
sed -e "s@DOCS=qemu-doc.html qemu-tech.html @DOCS=@" \
-e "/_DATA) qemu-doc.html/d" \
-i Makefile
tobias
--
http://linuxfromscratch.org/mailman
Am 23.02.2013 00:47, schrieb Tobias Gasser:
>
> i'll report if i run into problems.
>
i finished the build without any errors. this includes xfce, firefox
18.0.2, thunderbird 17.0.2, libreoffice 3.6.5.2, gimp 2.8.2.
i did not check about warnings, but none of my scripts aborted
Am 23.02.2013 00:13, schrieb Bruce Dubbs:
>
> We've worked around the gcc problems by avoidign building the .info
> files (does anyone really use them?).
i've seen the workaround.
i don't use .info-files, i even delete them!
> I just did a test build of wget-1.14 with lfs-7.3-rc1 and got no er
is it a wise decision to use upgrade to 5.0 for the 7.3 release?
as far i can understand there are no security issues with 4.13a but as i
can see from google there are quite some compatibilty issues in
different packages with 5.0 (coreutils, gcc, wget, qemu, docbook...)
i prefer to keep the 4.
> ERROR: g++.dg/abi/mangle33.C -std=c++98: error executing dg-final:
> couldn't compile regular expression pattern: out of memory
> ERROR: g++.dg/abi/mangle33.C -std=c++11: error executing dg-final:
> couldn't compile regular expression pattern: out of memory
no errors here
> in addition to your
Am 21.11.2012 18:21, schrieb Bruce Dubbs:
> I've been testing udev from systemd-196. I have been able to build and
> install it with some changes to the LFS Makefile. One new capability is
> that it has it's own hw database instead of lspci and lsusb.
udev/systemd gets uglier with each release...
Ragnar Thomsen schrieb:
> to the kde pam file. Now I can authenticate as root in KDE, but this only
> works when booting using KDM. When booting into console and then starting KDE
> by the startx command, logind doesn't get started.
> Is there a way to start logind in the .xinitrc file, similar
Am 31.08.2012 19:35, schrieb Nathan Coulson:
>> i'll write about the results as soon as the build is done.
i now have (almost) finished my system.
xfce is up with firefox and thunderbird. only libreoffice and vlc are
missing - but that's another story ;)
everything is fine, no errors occurred
Tobias Gasser schrieb:
>
> until now i used --strip-all on binaries. a little googling convinced me
> to change my script as --strip-unneeded seems to to have intelligence ;)
>
there is NO difference in size!
at least not with what i've built so far.
now i'll go ahead to
Nathan Coulson schrieb:
> Back when that text was added to the book, --strip-all removed
> something from the static libraries that is needed to link to them
> (basically making them useless). Symbol table perhaps?
i don't strip any static libs.
see my script:
file xx.a | grep ELF
will be empt
Baho Utot schrieb:
>
> #-- Options to be used when stripping binaries. See `man strip' for details.
> STRIP_BINARIES="--strip-all"
> #-- Options to be used when stripping shared libraries. See `man strip'
> for details.
> STRIP_SHARED="--strip-unneeded"
> #-- Options to be used when stripping sta
Ragnar Thomsen schrieb:
> It is stated in LFS that --strip-unneeded should not be used on
> libraries, as the static ones will be destroyed.
>
> I found this page:
> http://www.technovelty.org/linux/
>
> Which states that --strip-unneeded is safe to use on both shared and
> static libraries, whil
Am 29.08.2012 17:56, schrieb Bruce Dubbs:
>
> Yes, we have that fixed in svn and that will make it to the release.
>
to save some time and space, i don't build the posix files but use links:
for tz in etcetera southamerica northamerica europe africa antarctica \
asia australasia bac
Bruce Dubbs schrieb:
>
> I don't see glibc building timezones in my log.
i can confirm there is no timezone data installed in this run.
no idea why the last build installed the zoneinfo files.
during make, in the build-directory a subdir 'timezone' the zic, zdump
and tzselect are built. while t
Bruce Dubbs schrieb:
> Tobias Gasser wrote:
>> iso3166.tab should be copied to!
>>
>> tzselect fails if the file is missing.
>
> Yes, we have that fixed in svn and that will make it to the release.
>
fine
>>
>> to avoid building the timezone
iso3166.tab should be copied to!
tzselect fails if the file is missing.
and the one provided in tzdata2012e is newer then the one provided by
glibc where SS (South Sudan) is missing.
to avoid building the timezones while compiling glibc, one can apply the
following sed:
sed -e "s/localedata tim
in the current development the documentation for gawk was added by
'mkdir' and 'cp'. most other packages use 'install' when copying
additional files.
i already added the install statements for gawk 4.0 in my own scripts.
so i'd like to know why you use copy instead of install here. install
has th
zlib 1.2.6 seems to have some issues.
building partimage 0.6.9 fails with several errors in imagefile.cpp
error cannot convert 'gzFile_s**' to 'gzFile' for argument...
using zlib 1.2.5 fixes this error.
google was no help, possibly i'm the first to hit this issue as zlib
1.2.6 was released jus
Tobias Gasser schrieb:
> ext4magic
> error: invalid use of undefined type 'struct opaque_ext2_group_desc'
>
for ext4magic there is a patch available at
http://developer.berlios.de/patch/?group_id=11787
tobias
--
http://linuxfromscratch.org/mailman/listinf
ext4magic and extundelete won't compile with 1.42
with 1.41.14 both compile fine.
ext4magic
error: invalid use of undefined type 'struct opaque_ext2_group_desc'
extundelete
error: invalid use of incomplete type 'struct opaque_ext2_group_desc'
there seems to be a major change in the api
photorec
Gilles Espinasse schrieb:
>>> t-get_d fails here
>>>
>>>
>>> i use 32bit, thus ABI=32 is set
>>>
>>>
it's a 32bit kernel on 64bit hw (intel i3)
> I use
> # --build=i486-linux force 32-bits ABI on an AMD 64-bits CPU
> # and disable sse2 intructions usage on i686
> # on sparc64 and powerpc
t-get_d fails here
the first 9 test run fine
then 1 of 30 fail
no more tests are run, maybe more tests are broken
mpn_get_d wrong on 2^n
want = 1
got = 5.18...e-318
can anybody confirm the failure?
or any hint what might be wrong with my build?
google was no help
i use 32bit, thus ABI=32 is se
Ken Moffat schrieb:
> On Wed, Jan 12, 2011 at 12:42:59AM +0100, Tobias Gasser wrote:
> In my case, attempting to try out xorg 7.6 was one of the reasons
> for me to build a new system, and doing that made me update my LFS
> packages and look at what other changes were in prospect. I
Ken Moffat schrieb:
> On Tue, Jan 11, 2011 at 02:18:24PM +0100, Tobias Gasser wrote:
>> next occurence is libXaw, widgets.ps in specs
>>
> I had no problems with either libX11-1.4.0 or libXaw-1.0.8 (one
> complete build with xorg-7.6 and gnome-2.30, one partial build
next occurence is libXaw, widgets.ps in specs
--
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page
compiling libX11-1.3.4 fails with groff 1.21, with groff 1.20.1
everything's fine.
the problem occurs in specs/libX11 and specs/XIM (but not with
specs/i18n) when building the .ps files.
the error is "table wider than line width"
as the error does not occur when removing .ps from the makefile or
Bruce Dubbs schrieb:
> Tobias Gasser wrote:
> I'd like to help...
>
> Can you give me the exact url you were trying?
www.linuxfromscratch.org
left-menu: Wiki (reaching wiki.linuxfromscratch.org)
now either LFS or BLFS
(currently i get the error right now, as i seem to have t
sorry if i'm writing in the wrong list, but book or support seem not to
fit better...
i try to login into the wiki. i already have an account on lfs and blfs,
but didn't login for quite a while. now trying to login, i end up in a
redirection loop. i tried with both firefox and opera without succes
Bruce Dubbs wrote:
> Can you paste the output from version-check.sh and uname -a please.
>
> Did you install xubuntu or just run from the CD?
>
both.
sorry for my late response, but i'm currently running some more test
with different setups and i will put the logfiles on this page:
http://s37
Chris Staub schrieb:
> No, it is not necessary. All that's needed to avoid host tools is
> following the book's instructions. That is it. If anything else is
> "necessary", it means you've goofed up and need to go back to the beginning.
i restarted more than once. not only using my scripts which
Tobias Gasser schrieb:
> for quite some time i used an older lfs to build a new one.
>
> now i'm building a new one from a xubuntu 9.10 lifecd.
thanks for all your replies. i try to summarise.
i use the same scripts i always use to build. as mentionned, i used to
start from w
Andrew Benton schrieb:
> Support questions should be sent to lfs-supp...@linuxfromscratch.org
sorry, as i posted another message in this list (which i guess is
correct) i forgot to switch to the correct list... but as i solved the
problem some minutes ago, i reply here again and don't move.
> Did
Guy Dalziel schrieb:
>> + This package contains math libraries and has useful functions for
>> arbitrary precision arithmetic. It is required to build Gcc.
>
> I disagree which the correction, 'has' feels more singular than 'have'
> and therefore I feel that 'have' is correct. However, it would so
after chapter 5 i run
readelf -f /tools/bin/* | grep "Request"
everything is fine except perl
perl still points to the hosts /lib/ld-linux.so !!
in chapter 6.7 (linux-headers) the job stops as perl can't run.
i guess to symlink ld-linux.so from /tools/lib to /lib is not really a
solution - wi
for quite some time i used an older lfs to build a new one.
now i'm building a new one from a xubuntu 9.10 lifecd.
to build glibc in chapter 5.7, i had to add the same construction with
CC/AR/RANLIB as with binutils/gcc pass 2 (ch 5.9/5.10) to use the newly
built toolchain.
without, the hosts t
Bryan Kadzban schrieb:
>
> Noting which programs are needed wouldn't be too bad. Moving a lot of
> them into LFS, I wouldn't like. Most of the stuff you mentioned is
> pretty rarely needed in reality -- or at least, on my system, which is
> mostly all I care about. Obviously this is a bit of a
Nathan Coulson schrieb:
>
> I have been experimenting with a multilib LFS System (where /lib,
> /usr/lib are used for 64bit, and /lib/32, and /usr/lib/32 are used for
> 32bit). I wanted something as close to LFS as possible, primarily
> 64bit, but just enough 32bit so if I wanted to compile some
> i didn't build the static zlib as i remove all static libs whenever a
> dynamic one is installed too. i didn't have a zlib.a just the zlib.so
> (in /lib to have it available during boot).
i've got the problem fixed.
haveing read my last post again, i got a little pale ... then more than
just a
i currently have 2 approches to get the perl zlib up and running. both
don't please me.
first:
as mentionned, removeing the sed-stuff everything runs fine.
second:
i didn't build the static zlib as i remove all static libs whenever a
dynamic one is installed too. i didn't have a zlib.a just
Matthew Burgess schrieb:
> On Mon, 3 Aug 2009 14:25:35 -0600, Matthew Burgess
> wrote:
>
>> What does 'strings /usr/lib/zlib.so | grep -i zlibVersion' output for you?
>
> 'nm -g /usr/lib/libz.so | grep zlibVersion' is probably the more proper
> command to use,
> although they should both outpu
Matthew Burgess schrieb:
> On Mon, 3 Aug 2009 14:25:35 -0600, Matthew Burgess
> wrote:
>
>> What does 'strings /usr/lib/zlib.so | grep -i zlibVersion' output for you?
>
> 'nm -g /usr/lib/libz.so | grep zlibVersion' is probably the more proper
> command to use,
> although they should both outpu
Tobias Gasser schrieb:
> i propose to remove the sed from the perl page.
> just keep the build in zlib.
>
i withdraw my proposal.
as matthew explaind, i must have a bug in my script(s).
i'll try it tomorrow
thanks for your help
tobias
--
http://linuxfromscratch.org/mailman/li
Matthew Burgess schrieb:
> What does 'strings /usr/lib/zlib.so | grep -i zlibVersion' output for you?
>
> On mine I get 'zlibVersion', as expected.
same here.
> A rebuild of zlib might get things working for you again.
i already rebuild it, nevertheless perl tests fail.
i just restart once
> no i'll remove perl again and rebuild it with out the zlib-sed thus
> using the included zlib. i'll post the result as soon as available.
that's it
without the sed everything is fine:
All tests successful.
please try to update cpan by
perl -MCPAN -e shell
install CPAN
i propose to
Tobias Gasser schrieb:
> the following sefaults (i've rebuild the book on a second machine with
> the same result!)
meanwhile my second machine has rebuild from scratch and reached the
perl script.
the last builds i omitted the test suites. this time i run the test for
perl and get
Tobias Gasser schrieb:
> the following sefaults (i've rebuild the book on a second machine with
> the same result!)
next try (1):
i removed perl, and built it again without the zlib-sed.
fine. everyting ok. except i don't like perl using it's own lib not the
dynamic from
the following sefaults (i've rebuild the book on a second machine with
the same result!)
perl -MCPAN -e shell
install CPAN
the author file is downloaded to
~/.cpan/sources/authors/01mailrec.txt.tmp123456
after reaching 100% i just get a segfault
same with "cpan" instead of "perl -MCPAN
> Has anyone tried to use this yet?
yep!!
i just applied the patch from the book for 4.4.0 in chapter 5:
hunk #1 succeeded at 6467 (offset 97 lines).
no real problems during compiling the book.
system booted fine.
i'll now continue with some more packages from blfs, but lfs seems to be
fine
Randy McMurchy schrieb:
> The easiest and best way to keep track of changes is to subscribe to
> LFS-Book and read the commit messages. If it ain't there, then it didn't
> happen.
depents on how you define "easy".
i agree the log is complete. absolute. each and every change is logged.
but easy
Bruce Dubbs schrieb:
> You can find this yourself by browsing the source xml and looking at the
> revision log.
>
> http://wiki.linuxfromscratch.org/lfs/log/trunk/BOOK/chapter06/chapter06.xml
>
h.
searching for "inetutils" in the search-box i couldn't see it.
checking the mentionned 8836
just a short question where i could not find the answer in the changelog:
when did you move inetutils from after gzip to after libtools??
thanks
tobias
--
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page
Tobias Gasser schrieb:
> i'll post my results as soon as i have them...
i just attached an additional disk to the esata port. the /dev/disk/by-*
structure was updated, the new partitions showed up. mounting (even by
-U uuid) works fine. even "tune2fs -U " works fine despite t
Tobias Gasser schrieb:
> i'll now run a new build with removing libblkid.so.1.0.0 after
> installing e2fsprogs but i guess there might some package(s) requireing
> it too bevore util-linux-ng will install it's version.
>
despite i didn't really expect it, but it reall
Matthew Burgess schrieb:
> Hi all,
>
> So, I'm trying to get Util-Linux-NG-2.15.1 to play nicely with
> e2fsprogs-1.41.6.
>
> I'm taking the following approach:
>
> 1. Build pkg-config in chapter 5 (required for point 3)
> 2. Build util-linux-ng with --with-fsprobe=builtin and
>'make -C lib
Bruce Dubbs schrieb:
> I don't think saving the space being
> discussed is worth the multiple changes and risks required.
my concern is not space.
i just like to have dynamic libs for security reasons. whenever i update
a package i know to have covered all dependent packages. now libcrypt in
sys
Gilles Espinasse schrieb:
>> i don't link anything static. i move all .a files out of the library
>> path if there is a corresponding .so available.
>>
>> excptions are the modutils package (which require libc and libz to be
>> linked static), and sysvinit (libcrypt) where i copy the static
>> libr
Bruce Dubbs schrieb:
> Static linking for this is OK with me, but do we know all the packages that
> use
> libuuid?
i don't link anything static. i move all .a files out of the library
path if there is a corresponding .so available.
excptions are the modutils package (which require libc and li
Bryan Kadzban schrieb:
> Tobias Gasser wrote:
> Out of curiosity: did your util-linux-ng version of mkswap build with, or
> without, UUID support? (See the output of configure.) That's the part
> that I believe is still missing in 2.15.1. :-)
i can't see anything i
Bruce Dubbs schrieb:
> Tobias Gasser wrote:
>> kernel 2.26.30 allows compression with gz, bz2 and lzma
>
>> my current kernel-sizes:
>> gzip 1913696
>> bzip21820944
>> lzma 1605712
>
> When the smallest disk dive you
kernel 2.26.30 allows compression with gz, bz2 and lzma
gzip and bzip2 are fine with the installed tools.
lzma is missing in the book.
first i downloaded the "original" lzma sdk mentionned on www.7-zip.org
but had no success.
next i tried http://tukaani.org/lzma
i took the current source lzma-
jp schrieb:
> Le vendredi 19 juin 2009 18:19:51, Tobias Gasser a écrit :
> I found my way with those packages, using e2fsprogs-1.41.6 and util-linux-
> ng-2.15.1.
> The trick was to build util-linux-ng _after_ e2fsprogs. I also had to add the
>
> option --with-fsprobe=builtin
Bryan Kadzban schrieb:
> Tobias Gasser wrote:
>> chapter 6.56 / svn-20090611
>>
>> the vol_id mentionned in the short descriptions does not exist any more.
>> the libvolume_id is missing too.
>
> See bug #2391.
>
> In short: the current book is broken.
meanwhile i found a thread on this list started by nathan coulson about
the same issue, thus sorry for the noise...
..but:
the blkid is NOT from util-linux-ng as i guessed but from e2fsprogs!
i tried with the newest versions of e2fsprogs (1.41.6), util-linux-ng
(2.15.1) and udev (142).
i tr
chapter 6.56 / svn-20090611
the vol_id mentionned in the short descriptions does not exist any more.
the libvolume_id is missing too.
i guess this is why dev/disk/by-* is incomplete, only by-id and by-path
are created but no by-label or by-uuid.
i usded the by-label to identify devices and now a
>> * * PART2
>> the testsuite complains about 'getline'
>> same error, thus same fix as with binutils:
>> sed -i -e 's/getline/get_line/' libiberty/testsuite/test-demangle.c
>
> There's a gcc-bugs list, and a Bugzilla instance running. I suspect
> they've already been notified, but it wouldn't h
** PART 1
the book actualy applies
sed -i 's/^XCFLAGS =$/& -fomit-frame-pointer/' \
gcc/Makefile.in ;;
but:
i can't find any XCFLAGS in the gcc/Makefile.in
the actual sed just does nothing, and the system still seems to be stable.
but the book explains clearly why we should have the f
Bruce Dubbs schrieb:
>> error: conflicting types for 'getline'
>>
>
> This is already in the -dev version of the book and I have reported it
> upstream.
>
i just checked the changelog, not the binutils chapter.
this is not the first time i overlooked something when i try to update
my system.
libiberty/testsuite/test-demangle.c
error: conflicting types for 'getline'
stdio.h now defines getline().
i didn't check wether older versions did this define too, or another
include is responsible for setting __USE_XOPEN2K8 requireing stdio.h to
define this function. i've no clue what "Define
Chris Staub schrieb:
> What mountpoint is this exactly? I've never had any issues with removing
> the Coreutils source dir, and have never seen anything mounted there.
the root partition is mounted in
[...]/coreutils-7.1/tests/cu-one-file-system.KqeMg6JYEx/a/b
options rw,errors=continue,data=ord
Chris Staub schrieb:
> What mountpoint is this exactly? I've never had any issues with removing
> the Coreutils source dir, and have never seen anything mounted there.
> I've just run the Coreutils testsuite, exactly as stated in the book,
> and nothing was mounted in the Coreutils source dir a
running the tests from coreutils, the sources can't be removed as a
mountpoint within the test is not unmounted.
solution:
MNT=$( mount | grep coreutils | awk '{print 3}' )
for i in $MNT ; do
umount $i >/dev/null
done
i added this lines after "make install" quite some time ago to both
cha
glibc 2.9 is NOT available at ftp.lfs-matrix.net as mentionned in ch03 /
packages.html (dev. branch)
i found it at
http://ftp.gnu.org/gnu/glibc
either make it available at the lfs-site, or update the packages.html
thanks
tobias
--
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: htt
Matthew Burgess schrieb:
> That also probably explains why the chapter 6 builds of GMP & MPFR configure
> correctly, as they're configured by hand, not by some Makefile hocus pocus.
oops!
reading again, my problem seems not to be the same as yours.
i've no problem in chapter 5 where gmp is me
Matthew Burgess schrieb:
> On Mon, 24 Nov 2008 00:35:16 +0100, "Gilles Espinasse" <[EMAIL PROTECTED]>
> wrote:
>
>> Just look in configure.log, you will see at wich line of configure the
>> test
>> is made.
>> configure:1654: checking build system type
>> configure:1672: result: athlon-pc-linux-g
Gilles Espinasse schrieb:
> You have written
>> i really have no clue what happens. possibly missing a library, but i
>> can't see which and/or why. i'm shure to have followed the book except
>> CFLAGS="-O3 -march=i486"
>
> perl understand your arch as i686.
> What is the result if you remove your
Ken Moffat schrieb:
>
> Do you have /mnt/lfs/usr/include/crypt.h ?
>
last night i run a complete rebuild, the 1st perl run was a failure, the
second successfull.
crypt.h is installed, and ldd perl shows:
linux-gate.so.1 => (0xb7fd2000)
libnsl.so.1 => /lib/libnsl.so.1 (0xb7fb7
Ken Moffat schrieb:
> On Fri, Nov 14, 2008 at 04:14:33PM +0100, Tobias Gasser wrote:
>
> [ not commenting on the main problem, or your seds - I find perl's
> configuration painful enough at the best of times, so I'm chickening
> out for the moment ]
[...]
&
Gilles Espinasse schrieb:
> You should post an url where the log file could be loaded.
> Mail message size is probably limited on the list to reject spam and
save our
> boxes.
http://s37a.e-b-p.net/daten/perl-problem.tar.bz2
compressed just 24k. maybe that would even be small enough for the
second try. the first with attached logfiles (194kb) seems to hang
arround somewhere since 3 hours now. if anybody wants to help me, please
send a mail and i'll send you the logfiles.
* * *
i've finished chapter 6. had to adjust some of my scripts, but no
problems except the GMP issue with A
Gilles Espinasse schrieb:
> Selon Tobias Gasser <[EMAIL PROTECTED]>:
> You could look at http://www.gossamer-threads.com/lists/perl/porters/205466
> That's because your machine is 64b, has /lib64 and configure script fail to
> handle properly your case.
i have a 32bit sys
im sticking with perl 5.10.0 (chapter 6.26)
** log stdout:
cc -o miniperl \
`echo gv.o toke.o perly.o op.o pad.o regcomp.o dump.o util.o mg.o
reentr.o mro.o hv.o av.o perl.o run.o pp_hot.o sv.o pp.o scope.o
pp_ctl.o pp_sys.o doop.o doio.o regexec.o utf8.o taint.o deb.o
universal
Ken Moffat schrieb:
>>> This implies Tobias has both a compiler and libc in chroot that are
>>> able to handle -m64.
>> No. See above.
base system is lfs, compiled in february 08 with kernel 2.6.24
no 64bit compiler or library
the cpu is 64bit capable
> You appear to have set $CFLAGS, perhaps
Bruce Dubbs schrieb:
>> What is the answer of uname -m on this machine?
i686
> or /proc/cpuinfo ?
processor : 0
vendor_id : GenuineIntel
cpu family : 15
model : 4
model name : Intel(R) Celeron(R) CPU 2.93GHz
stepping: 1
cpu MHz : 2927.865
cache si
i had to add ABI=32 as my system was identified ad 64bit.
./configure ABI=32 --prefix=/usr --enable-cxx --enable-mpbsd
i'm using CFLAGS="-O3 -march=i486" as a global setting, overwritten for
some special cases mentionned in the book.
any idea why i have to add ABI=32 ???
thanks
tobias
--
htt
the current rc1 and SVN-20081106 both require the following patch
expect-5.43.0-tcl_8.5.4_fix-1.patch
both versions use tcl 8.5.5.
as i did not compile the book but just am downloading the packages i
can't tell you wether the fix is still required or not.
but imho either rename the patch or r
85 matches
Mail list logo
