LFS7.0rc2 - Section 2.2 space requirements calculation seems low
In section 2.2, we have the statement "A minimal system requires a partition of around 1.3 gigabytes (GB). This is enough to store all the source tarballs and compile the packages." I created a 2GB partition, and ran out of space during the make phase of 6.17 GCC The package details in that section show the disk space used as 1.7GB "6.17. GCC-4.6.1 The GCC package contains the GNU compiler collection, which includes the C and C++ compilers. Approximate build time: 47 SBU Required disk space: 1.7 GB" Perhaps the numbers need some tweaking? I'll resize the partition higher and start chapter 6 again, and see if i can rememeber to keep an eye on used sizes. -- -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
LFS7.0rc2 - Section 1.1 - LFS LiveCD still valid as a build host?
resending as previous is awaiting moderation as I forgot to re-join ML before posting --- The LFS LiveCD is listed in the 1.1 introduction as valid to use in building LFS as a host distribution, with a caveat about newer hardware not being detected. However, the kernel version of the LFS LiveCD does not meet the minimum host requirements anymore, and the configure step fails in 5.8 (Binutils pass 2) with a kernel too old message as a result. ISO version: lfslivecd-x86-6.3-r2145-min.iso configure:4217: result: a.out configure:4233: checking whether the C compiler works configure:4242: ./a.out FATAL: kernel too old configure:4246: $? = 1 configure:4253: error: in `/mnt/lfs/sources/binutils-build': configure:4257: error: cannot run C compiled programs. >From version-check.sh Linux version 2.6.22.5 (root@kerrek) (gcc version 4.1.2) #1 SMP Thu Nov 29 07:53:26 GMT 2007 >From Glibc Chapter 5.7.1 ../glibc-2.14.1/configure --prefix=/tools \ --host=$LFS_TGT --build=$(../glibc-2.14.1/scripts/config.guess) \ --disable-profile --enable-add-ons \ --enable-kernel=2.6.25 --with-headers=/tools/include \ libc_cv_forced_unwind=yes libc_cv_c_cleanup=yes So either a) instructions need to be modified to allow 2.6.22 kernel level to be specified in 5.7 and 6.9.Glibc (and host requirements modified accordingly) b) a note about LiveCD glibc configure parameter change needs to be added (either in 5.7/6.9, or in 1.1) c) LiveCD retired as a valid LFS build host d) an updated LiveCD build is made available. e) something else... If I have time I might look at updating the LiveCD to current levels, but it won't happen till the new year, which will be too late for LFS 7.0 release I suspect I'll modify the kernel level in this build and continue building with the current LFS LiveCD and see what else pops out. -- -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Ok, now I am having issues
On Tue, Nov 8, 2011 at 6:28 AM, Danny Vukobratovich wrote: > > From: "Bruce Dubbs" > To: "LFS Support List" > Sent: Monday, November 7, 2011 11:42:41 AM > Subject: Re: Ok, now I am having issues > > Danny Vukobratovich wrote: >> Ok, then what should I use to build my LFS if not the liveCD? Just >> trying to go with what is recommended...thanks, > > Don't top post. > > > From Chapter 1, first paragraph: > > "The LFS system will be built by using an already installed > Linux distribution (such as Debian, Mandriva, Red Hat, or SUSE). This > existing Linux system (the host) will be used as a starting point to > provide necessary programs, including a compiler, linker, and shell, > to build the new system." > > I will remove the 2nd paragraph. > > -- Bruce > -- > http://linuxfromscratch.org/mailman/listinfo/lfs-support > FAQ: http://www.linuxfromscratch.org/lfs/faq.html > Unsubscribe: See the above information page > > - > > Ok, I am downloading Fedora right now. Do I boot to it via a LiveCD or do I > perform a full installation? The wording appears to be a full > installation. Thanks > > > > Danny Vukobratovich > System Administrator > > > > > -- > http://linuxfromscratch.org/mailman/listinfo/lfs-support > FAQ: http://www.linuxfromscratch.org/lfs/faq.html > Unsubscribe: See the above information page > > You can either run it live or installed to your hard disk - if you run it live, you will lose any changes you make to the host system (such as installing required packages, etc) Since Fedora\Ubuntu\et al are intended to be desktop systems, they sometimes come without components needed to build LFS, or use different versions of the packages (mawk instead of gawk, etc) so you will need to use the hosts package manager to install any missing components - use the Host Requirements script in section 1 to determine what is installed in your host system, Note you can use the SLAX distribution, which is based on the original LFS Live CD, and contains an updated toolchain that should be able to build LFS without issue (untested, but on my list to do). You can also search the LFS mailing lists - from the July LiveCD mailing list, someone has apparently already updated the LiveCD 2011/7/12 : > 2011/5/23 : >> 2011/5/22 Jeremy Huntwork : >>> On 5/21/11 8:53 PM, parttor at gmail.com wrote: >>>> I've created the livecd few weeks before. >>>> http://yesit.tk/livecd.html > I've updated it to 2011-0709 > kernel 2.6.39.3 > gcc 4.6.1 > glibc 2.14 now 2011-0723 released with 3.0 kernel >>> >>> Nice. Did you document how you built it? >> >> I did it follow these hints. >> http://www.linuxfromscratch.org/hints/downloads/files/lfscd-remastering-howto.txt >> http://www.linuxfromscratch.org/hints/downloads/files/boot-cd_easy.txt >>> >>> JH >>> -- >>> http://linuxfromscratch.org/mailman/listinfo/livecd >>> FAQ: http://www.linuxfromscratch.org/faq/ >>> Unsubscribe: See the above information page >>> >> >> >> >> -- >> code farmer from lanxiangjixiao >> > -- -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: [lfs-support] Use VM to do LFS
On Mon, Jan 16, 2012 at 1:03 PM, Li, David wrote: > > I am new to LFS and am very interested in using a virtual machine (e.g > virtual box vm). Is this something documented in the book already. Does this > mean it's doable? > > Thanks. > Works fine. Building current SVN using VMWare 7 workstation on Windows (using SLAX Live-CD as a host) as I type this (mainly to test\experiment with instructions\hints for initramfs and systemd) Only caveats to be aware of are making sure your eventual kernel configuration (and selected host distribution) support the VM drivers used for disk and networking by your VM system. -- -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
[lfs-support] LFS Museum - 6.7 directory permissions issue
http://archive.linuxfromscratch.org/lfs-museum/6.7/ all files report Access Denied when attempting to download. -- -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: [lfs-support] LFS Museum - 6.7 directory permissions issue
On Tue, Jan 17, 2012 at 2:00 PM, Kevin Wise wrote: > On Tue, 2012-01-17 at 13:54 +1300, Steve Crosby wrote: >> http://archive.linuxfromscratch.org/lfs-museum/6.7/ >> >> all files report Access Denied when attempting to download. > > Hello Steve, Not quite sure what could be the problem there. I tried the > link here and had no problem. Server could have had a temporary glitch > so you may want to try again now. > > Kevin > Could be the work proxy server (although I can access all other versions files okay) - will try from home later. -- -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: [lfs-support] Q: best choice of Live CD to host/rescue LFS?
On Mon, Jul 9, 2012 at 5:18 AM, Jeremy Henty wrote: > > Yasser Zamani wrote: > >> I recommend Slax (http://www.slax.org/get_slax.php/); I gracefully >> finished LFS with that as host. > > I ran the verson check script in Slax and it could not find bison, > yacc or makeinfo. Did you customise slax with extra packages? > I've sucessfully customised it in the past, however the author announced a few weeks ago that he has recently found some interest in funding futher development, so version 7 with updated kernel and tools is due for alpha release soonish. -- -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: [lfs-support] LFS 7.2 GCC pass 1
On Tue, Sep 4, 2012 at 4:23 AM, Richard Melville wrote: >> I extracted all of these packages from within the GCC-4.7.1 folder > > I'd still be interested to know why we build GMP, MPC, and MPFR inside GCC > except on the final build where they are built separately. > > Richard > During pass 1, GCC requires the GMP, MPC and MPFR libraries, but we don't want GCC to get these libraries from the host. GCC searches for the libraries either via the regular search paths *or* inside it's own source tree. We install a temporary copy inside the GCC source tree to take advantage of this, and thus allow GCC to not be contaminated with host libraries During pass 2, we are in a protected chroot environment, so we are no longer concerned about the host. So other programs in pass 2 can take advantage of the GMP, MPC and MPFR libraries later in the build, we install them before GCC instead of in the source tree. -- -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
[lfs-support] Disk Usage in Stable LFS 7.3
There is a comment in section 5.33 Stripping that reads "At this point, you should have at least 850 MB of free space in $LFS that can be used to build and install Glibc in the next phase. If you can build and install Glibc, you can build and install the rest too." However the disk usage requirements for GCC-4.7.2 in section 6.17 are stated as 2GB? "Approximate build time:53.5 SBU" "Required disk space:2.0 GB" Either the comment should be change to reflect GCC as the larger package, or the required disk space for GCC updated\corrected? -- -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: [lfs-support] Disk Usage in Stable LFS 7.3
I was more concerned about the statement that if you can build glibc, you can build all the rest... On Tue, Apr 30, 2013 at 12:19 PM, Bruce Dubbs wrote: > Steve Crosby wrote: >> There is a comment in section 5.33 Stripping that reads >> >> "At this point, you should have at least 850 MB of free space in $LFS >> that can be used to build and install Glibc in the next phase. If you >> can build and install Glibc, you can build and install the rest too." >> >> However the disk usage requirements for GCC-4.7.2 in section 6.17 are >> stated as 2GB? >> "Approximate build time:53.5 SBU" >> "Required disk space:2.0 GB" >> >> Either the comment should be change to reflect GCC as the larger >> package, or the required disk space for GCC updated\corrected? > > Yes, both statements were true at one time. The build I finihed a few > minutes ago used 921M for glibc an dgcc took 2226.622 M. > > Both numbers need to be updated. I'll update in a few minutes. > >-- Bruce > > -- > http://linuxfromscratch.org/mailman/listinfo/lfs-support > FAQ: http://www.linuxfromscratch.org/lfs/faq.html > Unsubscribe: See the above information page -- -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
[lfs-support] Coreutils 8.21 test failure - many-dir-entries-vs-oom
Just an FYI - This test fails for me (and looking at the archives at least one other recently), and looking at the logs, it's because it's attempting to create 200,000 small files - that exceeds the inode count on the seperate 2GB ext3 filesystem I created for sources. e.g. + expensive_ + test yes '!=' yes + mkdir d + cd d + seq 20 + xargs touch touch: cannot touch '125627': No space left on device touch: cannot touch '125628': No space left on device root@slax:~# tune2fs -l /dev/sdb1 | grep Inode Inode count: 131072 Inodes per group: 8192 Inode blocks per group:512 Inode size: 256 Perhaps a note to check the inode count before running this test might be in order, or a note that it might fail? -- -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
[lfs-support] Errata for LFS 7.4
Minor issue with LFS 7.4 Host Requirements Changelog lists host system requirements changes for kernel to 2.6.34, however host requirement section only says 2.6.32 -- -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
[lfs-support] Minor cosmetic book issue
Just noticed that the instructions for groff in section 6.46 is missing a verbose flag for the mkdir command in the installation section. Noticed in 7.4, still present in current development. e.g. Install the package: mkdir -p /usr/share/doc/groff-1.22/pdf make install --- Also FYI, browsing to http:\\linuxfromscratch.org presents a webpage menu, with links that go to an LFS 404 page. -- -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: nscd
"John Lane" <[EMAIL PROTECTED]> wrote in news:[EMAIL PROTECTED]: > > Hello, > > I want to use nscd but I can't find anything in the lfs document that > explains how to set it up. > > The binary is installed as /usr/sbin/nscd but it is not running. I > presume I need a boot script (couldn't see one in the lfs bootscripts) > and a conf file (again, couldn't find one). > > I would appreciate being pointed in the right direction. Google has > not been much help in this instance. > > Thanks! > start with "info nscd" and go from there... -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: packages needed?
tom <[EMAIL PROTECTED]> wrote in news:[EMAIL PROTECTED]: > maybe a stupid question but what packages are needed by host in order > to do a succesful build of LFS? Each package has it's dependecies listed. For example, LFS 6.0 Chapter 5 binutils has: Binutils installation depends on: Bash, Coreutils, Diffutils, GCC, Gettext, Glibc, Grep, Make, Perl, Sed, and Texinfo these are not always 100% accurate, and may be missing some packages, but are a good general guide. Doing a quick look through the list, your host should have the following packages: (note you should have the -devel versions of these packages installed, as that installes the necessary headers as well as the binaries). Bash Bison Coreutils Diffutils Findutils Gawk Gettext Glibc Grep Make Ncurses Perl Sed Texinfo Zlib Note that only Chapter 5 has host system requirements - once we start Chapter 6, we have already installed everything we need. In addition, the host system needs to have the Glibc headers installed, and (for LFS 6.x) a kernel version 2.6.2 or higher, compiled with gcc 3.0 or higher. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: m4?
tom <[EMAIL PROTECTED]> wrote in news:[EMAIL PROTECTED]: > theres no m4-1.4.2 that I see? theres 1.4.o or 1.4ppre2 so which one? The freshmeat page is out of date. The URL below has the version you are looking for. ftp://ftp.gnu.org/gnu/m4/ -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: gettext error?
tom <[EMAIL PROTECTED]> wrote in news:[EMAIL PROTECTED]: > On Friday 18 February 2005 01:04 am, tom wrote: >> Im on chapter 6 of the SVN version I get this error: >> >> make[2]: Leaving directory >> `/usr/sources/gettext-0.14/gettext-runtime/intl-csharp' >> make[1]: *** [install-recursive] Error 1 >> make[1]: Leaving directory >> `/usr/sources/gettext-0.14/gettext-runtime' make: *** >> [install-recursive] Error 1 > > I did disable c-sharp in chapter 5 why this error? > use --without-csharp, instead of --disable-csharp -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: gettext error?
Matthew Burgess <[EMAIL PROTECTED]> wrote in
news:[EMAIL PROTECTED]:
> Steve Crosby wrote:
>> use --without-csharp, instead of --disable-csharp
>
> Tom, could you let me know if this works please? According to
> gettext's configure script (lines 1911-1915 and 2038-2042)
> --disable-csharp should work fine.
>
> Regards,
>
> Matt.
>
actually, I was just regurgitating previous statements I've seen - it
*is* --disable-csharp (which is the default option btw)
looking at configure, it appears at first glance there may be a bug in
gettext-runtime/configure - see below. (not looked at other configure
scripts)
Here, CSHARP_CHOICE is set to "any"
===
if test "${enable_csharp+set}" = set; then
enableval="$enable_csharp"
CSHARP_CHOICE="$enableval"
else
CSHARP_CHOICE=any
fi;
===
Here, a check is made if CSHARP_CHOICE is set to "no"
===
for impl in "$CSHARP_CHOICE" pnet mono sscli no; do
case "$impl" in
no)
HAVE_CSHARPCOMP=
ac_result="no"
break
;;
esac
done
echo "$as_me:$LINENO: result: $ac_result" >&5
echo "${ECHO_T}$ac_result" >&6
===
And here, is where the decision to build CHSARP is decided, and where a
problem will exist if you have a C# compiler installed on your host - the
wrong decision will be made, since the CSHARP_CHOICE should have been set
to "no", not "any"
===
if test -n "$HAVE_CSHARPCOMP" && test "$CSHARP_CHOICE" != no; then
BUILDCSHARP=yes
else
BUILDCSHARP=no
fi
===
It would appear at first glance therefore, that the initial set of
CSHARP_CHOICE to "any" is buggy. Can't test at the moment as build is in
progress, but specifying --enable-csharp=no should work okay, and be
fairly trivial to test.
-- -
Steve Crosby
--
http://linuxfromscratch.org/mailman/listinfo/lfs-support
FAQ: http://www.linuxfromscratch.org/lfs/faq.html
Unsubscribe: See the above information page
Re: kernel panic :No init found.try passing INIT=option to kernel
[EMAIL PROTECTED] wrote in news:[EMAIL PROTECTED]: > hello, > i am new bie to linux ,i made LFS-5.1 successfully.now i want to make > LIVE CD of lfs , so in that direction i follow algorithm and try to > make cd but when it boot from CD ram0 is mounted as root file system > but after it shows message : kernel panic : No init found .try > passing INIT=option to kernel and refuses to procede further. > I searched in FAQ's but not able to find hint. > My kernel is comipled with INITAL Ram disk support also, and int grub > menu.lst > my enteries are > --- > kernel --no-mem-option /boot/kernelimage root=/dev/ram0 init=/linuxrc > initrd /boot/initrd.gz > > plz suggest me where i go wrong or what to do in order to overcome > the problem. > Thanx in advance. > Regards > Narendra > In addition to Initial RAM Disk support, you need to ensure you have CONFIG_BLK_DEV_RAM=y CONFIG_BLK_DEV_RAM_COUNT=2 CONFIG_BLK_DEV_RAM_SIZE=16384 where: RAM_COUNT is *at least* 2 (ram0 and ram1) RAM_SIZE is sufficient to fit your uncompressed initrd image There are several hints available at http://hints.linuxfromscratch.org that deal with creating a "live cd" image from LFS - look to them for further assistance. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: File System Issues during bootup
Gerard Beekmans <[EMAIL PROTECTED]> wrote in news:[EMAIL PROTECTED]: > On February 21, 2005 12:53 pm, Phoe6 wrote: >> I guess the lfs online should be updated to use e2fsprogs out, >> version 1.36 instead of 1.35 if people are using Fedora Core 3 as the >> host system. > > The Development version of LFS already uses e2fsprogs-1.36 > (http://www.linuxfromscratch.org/lfs/view/development/) > > Fedora Core 3 presented us with a number of issues lately. Either an > errata page is in order or simply an updated FAQ. > The actual problem is that FC3 provides an e2fsprogs that is patched to include additional capbilities, and uses those capabilities by default when creating partitions. The stock e2fs distribution we use does not understand those options, and therefore barfs. If I get some time today I'll scribble up a "fix" process for this issue - it's not hard to do. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: about LFS license problem
Gerard Beekmans <[EMAIL PROTECTED]> wrote in news:[EMAIL PROTECTED]: > On February 21, 2005 03:02 am, Bluesheet wrote: >> 1/ If I using 100% LFS to be my linux system of my product, do I need >> LFS license problem to sell my product? > > If you use LFS to create your own product the least you should do is > give this project credit for it. > >> 3/ If I following the LFS book to build but didn't use the LFS >> packages, that I download all the source from GNU and I make some >> scripts by myself, do I need LFS license to sell my product? > > The packages we use are almost all GNU packages. Most all packages are > GPL licensed and not licensed by LFS itself. You will need to check > every individual package's license to see what they require of you. > >> 4/ If I using 100% LFS packages to build my linux system, but I erase >> all the name of "LFS" and "linuxfromscratch.org" inside the code and >> source, do I need LFS license to sell my product? > > LFS' BSD license only applies to the book and the methodology we > employ to build a system. If you use our book to build your system, > you should mention this somewhere. It is the right and honorable thing > to do, after all. > > The actual software is from their respective owners. You can't take > Glibc and remove its GPL license for example, unless you modify it as > your own version of Glibc. Again check the licenses to see how this > work. Not every package is the same. > Note there are some LFS specific packages (bootscripts is one that comes to mind) and also the LFS created patches. These may need consideration as to the license. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: File System Issues during bootup
Gerard Beekmans <[EMAIL PROTECTED]> wrote in news:[EMAIL PROTECTED]: > On February 21, 2005 04:38 pm, Steve Crosby wrote: >> If I get some time today I'll scribble up a "fix" process for this >> issue - it's not hard to do. > > That would be greatly appreciated too. It might even be preferable to > put such a fix right into the LFS book. It's kind of a serious show > stopper. > Took less time than I thought. = When creating the LFS partition, you may need to make some slight adjustments if your host system uses certain filesystem features that won't be supported by LFS later on in the build process. To check if this is the case, run the following command: tune2fs -l /dev/ | grep features you should see output similar to the following: Filesystem features: filetype sparse_super The only LFS supported Filesystem features are (from the e2fsprogs 1.36 tune2fs man page): dir_index filetype has_journal sparse_super If you have any other features listed, you'll need to remove them from your new LFS partition before you proceed. This is done using the following command: tune2fs -O ^,^ /dev/ You may also need to run fsck on the partition you just modified to make the changes effective - tune2fs will let you know if thats the case. Some common features added by host systems that are not supported are: ext_attr resize_inode large_file To remove these, assuming your LFS partition is on /dev/hdb1 tune2fs -O ^ext_attr,^resize_inode,^large_file /dev/hdb1 == Feel free to mangle that blurb into something useful - probably best placed in the creating your LFS partition section if it goes directly into the book. Note that this is *untested* by me, I don't have a borked e2fsprogs to create such partitions with - someone should confirm the documented command syntax for removing the features works as advertised by the man pages. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: File System Issues during bootup
Vincent <[EMAIL PROTECTED]> wrote in
news:[EMAIL PROTECTED]:
> Steve Crosby wrote:
>
>> When creating the LFS partition, you may need to make some slight
>> adjustments if your host system uses certain filesystem features that
>> won't be supported by LFS later on in the build process. To check if
>> this is the case, run the following command:
>>
>> tune2fs -l /dev/ | grep features
>
> [snip]
>
>> If you have any other features listed, you'll need to remove them
>> from your new LFS partition before you proceed. This is done using
>> the following command:
>>
>> tune2fs -O ^,^ /dev/
>
> From a standard ISO FC3 install, my ext2 partition had additional
> [resize_inode] and [large_file] enabled. Unfortunately, the installed
> tune2fs from FC3 refused to remove those features.
>
> Anybody else managed to do so? Other features could be removed by
> tune2fs indicating for me the command syntax and executable were ok.
>
>
> Only solution I had was chroot into LFS, create a new partition and
> copy my LFS directories to this new partition. Edit /etc/fstab,
> /boot/grub etc etc by hand and reboot into this new partition. Worked
> like a charm, but might be a bit steep for first users.
>
> Groetjes,
> Vincent
Downloaded and checked e2fsprogs source from Fedora 3
yuck! The FC3 patches to add resize_inode don't add the capability to
remove the feature, and enable it by default. sigh. Haven't looked for
the large_file capability, since the fix below solves both.
Double yuck! Not only is resize_inode set as the default, it is ignoring
the -O flag to mke2fs, so you can't even disable it if you want to...
This is very ugly, but does work (tested here with FC3 e2fsprogs)
echo "features -" | debugfs -w -f- /dev/{xxx}
followed by an e2fsck to remove any resize_inode remnants from the actual
FS.
for example, to remove large_file and resize_inode from /dev/sdc1
echo "features -resize_inode" | debugfs -w -f- /dev/sdc1
echo "features -large_file" | debugfs -w -f- /dev/sdc1
e2fsck -p /dev/sdc1
-- -
Steve Crosby
--
http://linuxfromscratch.org/mailman/listinfo/lfs-support
FAQ: http://www.linuxfromscratch.org/lfs/faq.html
Unsubscribe: See the above information page
RE: File System Issues during bootup
"Sparhawk" <[EMAIL PROTECTED]> wrote in news:[EMAIL PROTECTED]: > I can confirm that Redhat Enterprise Linux 4.0 suffers from this same > problem as well. After completing the steps outlined below you > receive the following msg upon rebooting to LFS. > > Filesystem contains large files, but lacks LARGE_FILE flag in > superblock. > > You are then instructed to run e2fsck manually without -a or -p > options. When you do this from redhat it fixes the filesystem by > re-adding the large_file option back to the filesystem. > > Should this version of redhat be added to the bug as well? > > -Original Message- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On Behalf Of Gerard > Beekmans > Sent: Tuesday, February 22, 2005 11:20 AM > To: LFS Support List > Subject: Re: File System Issues during bootup > > On February 21, 2005 08:43 pm, Steve Crosby wrote: >> for example, to remove large_file and resize_inode from /dev/sdc1 >> >> echo "features -resize_inode" | debugfs -w -f- /dev/sdc1 >> echo "features -large_file" | debugfs -w -f- /dev/sdc1 >> e2fsck -p /dev/sdc1 > > Bad FC3, bad. Wow that's ugly. Rumor has it that e2fsprogs-1.36 may > have these > new features built in as well. Maybe not. > > I'll add your latest notes to the bugzilla bug and we'll have to do a > more thorough investigation and consider our options. > although it's not documented that I could see quickly, the large_file capability appears to already be in e2fsprogs upstream - so you only need to remove the resize_inode capability, which the e2fsprogs from LFS 6.0 and earlier can't understand. An alternative is to use version 1.36 of e2fsprogs in LFS, which seems to have implemented the resize_inode capability. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: esfsck error on boot
Kevin Lyles <[EMAIL PROTECTED]> wrote in news:4222869E.30604 @comcast.net: > -BEGIN PGP SIGNED MESSAGE- > Hash: SHA1 > > Hello all, > > I get an error from e2fsck every time I boot into my (brand new) LFS > partition: > > Mounting root file system in read-only mode [ OK ] > Checking file systems... > fsck.ext3: Filesystem has unsupported feature(s) (/dev/hda7) > e2fsck: Get a newer version of e2fsck! > File system errors were encountered that couldn't be > fixed automatically. This system cannot continue to boot > > As far as I can tell, the version of e2fsck on the LFS partition and on > my FC3 partition (which was the host for the LFS build) are identical; > 1.35. I can mount, unmount, and run e2fsck on /dev/hda7 from FC3 and > the rescued CD I have without problems or errors detected. I used nALFS Although FC3 reports the version as 1.35, they have actually applied some patches that add new functionality not supported by the stock LFS e2fsprogs. The *safest* way to fix this is to update your LFS parition to e2fsprogs 1.36, which has the FC3 extensions included. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: esfsck error on boot
Kevin Lyles <[EMAIL PROTECTED]> wrote in news:[EMAIL PROTECTED]: > Steve Crosby wrote: >| Kevin Lyles <[EMAIL PROTECTED]> wrote in news:4222869E.30604 >| @comcast.net: >| >| >|>-BEGIN PGP SIGNED MESSAGE- >|>Hash: SHA1 >|> >|>Hello all, >|> >|>I get an error from e2fsck every time I boot into my (brand new) LFS >|>partition: >|> >|> Mounting root file system in read-only mode [ OK ] >|> Checking file systems... >|> fsck.ext3: Filesystem has unsupported feature(s) (/dev/hda7) >|> e2fsck: Get a newer version of e2fsck! >|> File system errors were encountered that couldn't be >|> fixed automatically. This system cannot continue to boot >| >| >| >| >|>As far as I can tell, the version of e2fsck on the LFS partition and >|>on my FC3 partition (which was the host for the LFS build) are >|>identical; 1.35. I can mount, unmount, and run e2fsck on /dev/hda7 >|>from FC3 and the rescued CD I have without problems or errors >|>detected. I used nALFS >| >| >| Although FC3 reports the version as 1.35, they have actually applied >| some patches that add new functionality not supported by the stock >| LFS e2fsprogs. The *safest* way to fix this is to update your LFS >| parition to e2fsprogs 1.36, which has the FC3 extensions included. >| > I figured it'd be something like that. So, should I compile that > while booting from my rescue disk, and just chroot into my LFS > partition? > > Kevin > either a rescue disk or your host - chroot and build e2fsprogs, and you should be good to go. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Text stops before end of screen
> > On Wed, Jun 13, 2007 at 05:27:23PM -0400, Colin Kemp wrote: > card correctly or something, because the screen will go blank after > about 10 minutes of nonactivity and stay blank, but the comp itself By default the linux kernel implements powersave in text-mode consoles - use setterm -blank 0 -powersave off to remove that "feature" (although that doesn't explain why you can't get it back after a keypress. -- -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: 18 hours and counting
On 9/2/07, J. Greenlees <[EMAIL PROTECTED]> wrote: > > and right now, make bootstrap of gcc4.1.2 pass1 of chapter 5 is at 35 > hours and 17 minutes. > I may have to restart the build, with a larger swap partition to try to > get this time down. just about to burn the lfslivecd-6.3 r2052. I'll Way back when I was building LFS4.0 on my Pentium 100, I recall glibc took over three and a half days to compile...that was with 64Mb RAM though, and 128Mb swap. Since that event (took three weeks in total to build LFS) I have used the cross compiling hint and my main desktop running a VMWare instance of LFS CD to build my system and burn to CD, then boot Pentium from that - works fine, build time is only a few hours for current LFS + a few extras, resulting CD image is 34Mb -- -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: LFS 6.3 6.59 Stripping Again
On Feb 4, 2008 12:47 PM, Baho Utot <[EMAIL PROTECTED]> wrote: > I am scripting a build of LFS 6.3 and I would like to be able to skip 6.59 > Stripping Again by adding CFLAGS=-s to the bashrc environment, then building > all the packages. > Stripping is done by the linker, not the compiler, so you want to add the -s to the LDFLAGS instead. And no, it shouldn't cause any issues, although from memory not all programs respect LDFLAGS set from the environment, and hardcode values in makefiles. Also, the amount of disk space you will be saving by doing this is miniscule compared to the size of some of the packages - so there is no real advantage to doing it for each package as opposed to at the end. -- -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: VmWare: Giving back control after halt
On Mon, Mar 3, 2008 at 5:20 AM, lanas <[EMAIL PROTECTED]> wrote: > Hi all, > > Is completing the halt cycle something to do with ACPI options of the > kernel ? > in brief, yes - the VMWare virtual machine does not receive a "power-off" ACPI command, and therefore sits waiting. If you compile various elements of ACPI into your kernal, then a halt will also close the VMWare virtual machine - I'm not near my VMWare at the moment, so can't advise on the actual kernel parameters involved. -- -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Confusion re: separate gcc-build directory
On Wed, Aug 27, 2008 at 1:36 PM, <[EMAIL PROTECTED]> wrote: > I am using the html version of Linux From Scratch Version 6.3. > > I have completed all steps up to and including application of the specs > patch in section "5.11. GCC-4.1.2 - Pass 2, 5.11.1. Re-installation of > GCC". > > HOWEVER, the following instructions have me confused: > >>> >>> Create a separate build directory again: >>> >>> mkdir -v ../gcc-build >>> cd ../gcc-build >>> > > What is the meaning of create "AGAIN"? The separate "gcc-build" directory > I created in section "5.4. GCC-4.1.2 - Pass 1" still exists and is still > populated with files and directories. Am I supposed to delete the > contents and start over with an empty "gcc-build" directory in the same > location? > Yes. Please see section 5.1 Introduction, and refer to the second "Important" note, which tells you to remove *every* package source directory after completing the compilation\installation, unless you're explicitly instructed to leave it in place for later. -- -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: useradd now creates groups too
> On Mon, Sep 29, 2008 at 09:07:28AM +0100, Jeremy Henty wrote: >> >> I've noticed a change in the behaviour of useradd when upgrading LFS >> from 6.1.1 to 6.3: "useradd foo" now creates a default group "foo" >> and makes it the primary group of user "foo". It this meant to >> happen or have I messed something up? > It's a policy change introduced into the shadow package by the maintainers, to follow the "redhat" standard of defaulting to creating a group with the same name as the user. You can change this behaviour by modifying login.defs, but I don't have an installation handy to identify the option needed to change. -- -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Kernel compilation error
Ken Moffat <[EMAIL PROTECTED]> wrote in news:[EMAIL PROTECTED]: > On Thu, 3 Mar 2005 [EMAIL PROTECTED] wrote: > >> New information: >> >> The error message appears when I deactivate the "Networking support" >> from the options. >> >> Any help? >> >> Carlos >> >> On Thu, 03 Mar 2005 11:11:34 -0600, [EMAIL PROTECTED] said: >> > Hi. I'm trying to compile a kernel for my fresh LFS 6.0 (installed >> > usin > g >> > the ALFS profile). >> > >> > I'm getting an Error Message: >> > >> > LD .tmp_vmlinux1 >> > drivers/built-in.o(.text+0x823cd): In function 'hpsb_alloc_packet': >> > : undefined reference to 'alloc_skb' >> > (a few more messages like above) >> > make: *** [.tmp_vmlinux1] Error 1 >> > >> > What's wrong? >> > >> > Thank you, >> > >> > Carlos > > You've got a problem with your .config - in theory, all of the > dependencies for the various source files are correctly recorded in > the Kconfig system, in practice errors happen. > > First identify which files refer to alloc_skb, and where it is > defined, > then look at the Makefiles to see what will cause the object file to > be created. > >From linux-2.6.10 Curious - alloc_skb is defined in net/core/skbuff.c, which (at the top level net tree) depends on CONFIG_NET. The actual error above is from drivers/ieee1394 (i.e. firewire) - which doesn't depend on CONFIG_NET. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: ../binutils-2.15.91.0.2/configure --prefix=/tools \
Patrick <[EMAIL PROTECTED]> wrote in news:[EMAIL PROTECTED]: > [code] > ../binutils-2.15.91.0.2/configure --prefix=/tools \ > --disable-nls > [/code] > > it says to do that, but.. does the "\" after the --prefix=/tools mean > anything, or matter, or is it just because the PDF had to fit the > command in 2 lines or someting... The "\" character is the shell line conitnuation character - meaning the shell will not execute the command yet, as you have more data to type. It's primarily done in the book due to space considerations and to control line- wrapping. IF you copy & paste the commands from the book, they will work fine, as well as if you type them verbatim (i.e. still multi-line). However, if you don't have the command on multiple lines, you should *remove* the "\" from the command. For example: ../binutils-2.15.91.0.2/configure --prefix=/tools --disable-nls is correct, whereas: ../binutils-2.15.91.0.2/configure --prefix=/tools / --disable-nls is not (and actually produces unexpected results). -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Hard link -style package management
Tapio Kelloniemi <[EMAIL PROTECTED]> wrote in news:d0n8np$4bi$1 @belgarath.linuxfromscratch.org: > Philipp Tölke <[EMAIL PROTECTED]> wrote: >>Tapio Kelloniemi wrote: >>> There are many package managers which use symbolic links to track >>> which files belong to which package. I have thought of a similar >>> approach but with hard links instead. >>> [...] >>> ; and what other problems might exist? >> >>There is the problem, that you can't have /usr/src on a different >>partition - which is sometimes wanted. You would have to split it into >>/src /usr/src /opt/src a.s.o, and if a package wants to put files in >>/usr _and_ /etc you can't hardlink every file... > > Symlink approach has also this problem, though it is a bit easier to > solve this in that scheme. However, using multiple package directories > for a single package will work (actually better than symlink scheme's > solutions). The problem is that this needs probably some help from user. > The issue is that hard links *cannot* cross from one filesystem into another (as hardlinks are referenced by inode, and inodes are filesystem specific - so two different files on seperate filesystems can have the same inode) Soft links (also known as symbolic links) *can* cross filesystems, as they store the path to the linked file, not it's inode. So if: / /dev/hda1 /usr/dev/hdb4 then ln -s /usr/lib/libfoo.so.x.y.z /lib will work, but ln /usr/lib/libfoo.so.x.y.z /lib will not (and can't ever under the current hardlink semantics) -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Chroot problems
ray <[EMAIL PROTECTED]> wrote in news:269921d905031616426770f0f9 @mail.gmail.com: > I've gotten through chapter 5 with no problems, but I have trouble > once I'm in the chroot environment. No apparent problems mounting the > kernel virtual filesystems, goes into chroot fine, but once I'm in it > I have trouble. I enter a command, the command goes through and does > it's thing, but then it hangs and won't return to the prompt. Can't > kill it, either, it just sits there doing nothing. ps shows it as > defunct. So I can't do any of chapter 6 because of this. My host is > Ubuntu. Any thoughts on what I must do? > This is a known issue - bash has a bug, fixed in the WCONTINUED patch you can find on the LFS website. Just rebuild Chapter 5 bash with this patch, and the hangs should disappear. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: building LFS-6.0 from BootCD, some observations
Joe Kohler <[EMAIL PROTECTED]> wrote in news:423906BC.6030605 @wright.edu: > > In the .bashrc file on page 31 PATH =/tools/bin (no such file on the cd > so it must end up on the ramdisk or else it dosen't exist). Moving ahead > to the middle of page 41: > Binutils installs its assembler and linker in two locations, /tools/bin > and ... The problem for me is how to get all of this on the hard drive > and not the ramdisk. Can I just replace /tools with something like > $LFS/mytools (making sure $LFS=/mnt/lfs) ? > > I think what has happened is that the authors began the book with the > idea of using a standard distribution (see top of page 3) and then > added the cd later (see page 10). Both cases use the term "host". These > are quite different. > > Did you do the build using only the LFS6.0 cd and a clean disk? > > Joe > Your missing some important steps in the setting up phase of the book. Specifically, the point where you create the /tools symlink into $LFS/tools - this is where the /tools comes from, and the fact that it is a symlink means it will actually install into $LFS/tools. You should re-read Chapter 4, this has all the relevant details and explnations. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: why not /tools/lib?
"jcnet" <[EMAIL PROTECTED]> wrote in news:[EMAIL PROTECTED]: >> In chapter 6 Glibc will be installed before Binutils is compiled for >> the last time. Is is only *after* installation of Glibc when the >> "make install" is executed, so the make command in the >> binutils-build/ld directory is merely a preparation. >> It is not really used until chapter 6 after installation of the Glibc >> package. Everything that is installed in chapter 6 is there to stay, >> and the /tools directory does not belong in the final LFS system. > > I guess the /use/bin and /bin directory belong to the host system. > but why LFS change the host system's libraries. > Am i misunderstand it ? > > Here's a (simplified) explanation of the LFS build process: We are going to be building our new LFS system on a seperate parition (and even possibly a seperate hard drive). As such, when we boot into our new LFS system, the old host system will not be available. This means we have to make very sure that nothing in our finished LFS system points to or uses anything from the host system. This is achieved by doing a two-stage build. First (Chapter 5), we build a compiler, C library, and some other tools required to build the main LFS system. We place these packages into the /tools directory, so they are seperate from the host system, and we use a symlink ($LFS/tools) so we can still see these packages once we are in step 2 (Chapter 6). In order to make the Chapter 5 build work properly, we change the "ld" command (part of binutils) to find system libraries in the /tools/lib directory. Once we have done this, we can then build the rest of the Chapter 5 packages, ready for the chroot into Chapter 6. Because we are already in the binutils directory, we also change the configuration of "ld" back to /lib:/usr/lib, but we *don't* install the changes - we do this later in Chapter 6. This is the reason you are instructed to not remove the binutils build and source directories in Chapter 5. We *could* change the "ld" configuration back to /lib:/usr/lib as the last step in Chapter 5 - that may be something the editors will consider, in order to reduce confusion. Note that although we install the Chapter 6 "ld" configuration to point to /usr:/usr/lib, because we are in a chroot environment, it will actually use $LFS/usr:$LFS/usr/lib, so will not interfere with the host system at all. Clear as mud? ;) -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Problem .. Bootin' from LILO
"Mukesh Kaushal" <[EMAIL PROTECTED]> wrote in news:[EMAIL PROTECTED]: > Hello there, > I've created my own customized linux system usin' LILO bootloader and > I've > also made a live CD of it.This is also workin' well on every machine > but > on some machines it is showin' followin' msg n' not bootin': > --- > L 99 99 99 99 99 99 99 99 99 99 > 99 99 99 99 99 99 99 99 99 99 99 > 99 99 99 99 99 > --- > Could anyone suggest me why it is happenin' > > Thanx in advance > > >From the lilo man page: Errors 99 and 9A usually mean the map file (-m or map=) is not readable, likely because LILO was not re-run after some system change, or there is a geometry mis-match between what LILO used (lilo -v3 to display) and what is actually being used by the BIOS (one of the lilo diagnostic disks, available in the source distribution, may be needed to diagnose this problem). -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Problem .. Bootin' from LILO
"Mukesh Kaushal" <[EMAIL PROTECTED]> wrote in news:[EMAIL PROTECTED]: > Thanx Steve, > May be you are right.But there is no map file in my file system.Now .. > what 'll I do?? Shall I install lilo again. > Except this in the live CD the TERM variable is also not being set so > that `clear' command is not workin'. > Plz tell me where I 'll have to set TERM n' PS variable?? > > Thanx for ur nice reply Please don't post replies at the top of the message (top-posting) As the man page states, you should run lilo -v3 on the affected system, and compare the results to the BIOS settings - they need to match. Setting the TERM and PS variables is normally done in either the /etc/profile or /etc/bashrc files. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
RE: Problem .. Bootin' from LILO
"ULStudent:Donal.Farrell" <[EMAIL PROTECTED]> wrote in news:[EMAIL PROTECTED]: > Hi Mukesh, I am unable to help you with your problem, however, if you > had the time, I'd like to know how to make a LiveCD. I've got LFS 6.0 > working, and I'm going through BLFS at the moment. Is there any > documentation abot how to make an LFS build into a LiveCD? > > Thanks in advance, and I hope you get an answer to your own question > Donal Farrell > If you go to the Hints link from the main linuxfromscratch.org website, you will find several documents desribing the steps needed to make a CD bootable LFS system. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: [EMAIL PROTECTED] ?
Wang Weiping <[EMAIL PROTECTED]> wrote in
news:[EMAIL PROTECTED]:
> this is the first i try LFS . so I hava lots of
> questions.
>
> after i build glibc , try the test:
>
> --
> echo 'main(){}' > dummy.c
> cc dummy.c
> readelf -l a.out | grep ': /tools
> -
>
> than the wrong message show me :
>
> --
> cc dummy.c -o dummy
> /tools/lib/libc.so.6: undefined reference to
> [EMAIL PROTECTED]'
> collect2: ld returned 1 exit status
> make: *** [dummy] Error 1
> --
>
Which version of glibc? and of LFS? This error has been seen recently with
the latest glibc snapshots and some other packages, but should not occur
with the specified version in LFS 6.0
-- -
Steve Crosby
--
http://linuxfromscratch.org/mailman/listinfo/lfs-support
FAQ: http://www.linuxfromscratch.org/lfs/faq.html
Unsubscribe: See the above information page
Re: No such file or directory
Jim Gifford <[EMAIL PROTECTED]> wrote in news:[EMAIL PROTECTED]: > Alexander V. Shepetko wrote: > >>Hi! >> >>ls -l /tools/bin/perl >>-rwxr-xr-x 2 root root 1035129 Mar 27 20:01 /tools/bin/perl >> >>But, >> >>/tools/bin/perl >>bash: /tools/bin/perl: No such file or directory >> >>Why this happens? >> >> >> > What does ls -l / look like? > Interesting - your /tools/bin/perl is hardlinked to something else (The "2" in the ls listing indicates the number of links to the file). I don't have a /tools directory handy, but I'm fairly sure that's not normal. An ls -li /tool/bin/perl will give you the inode of the linked file, which you can then ls -liR /tools | grep to find the other copy of the file. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: No such file or directory
Steve Crosby <[EMAIL PROTECTED]> wrote in news:[EMAIL PROTECTED]: > > Interesting - your /tools/bin/perl is hardlinked to something else > (The "2" in the ls listing indicates the number of links to the file). > > I don't have a /tools directory handy, but I'm fairly sure that's not > normal. > > An ls -li /tool/bin/perl will give you the inode of the linked file, > which you can then ls -liR /tools | grep to find the other > copy of the file. > *sigh* - never assume ;) of course, /tools/bin/perl is hardlinked to /tools/bin/perl sorry for the noise -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Problem in Chapter 6 - glibc zoneinfo Missing
On 13 May 2005, you wrote in lfs.support: > Hi, > > While building chapter 6 glibc v2.3.5, I get the following error: > > bash-3.00# cp --remove-destination \ > /usr/share/zoneinfo/America/Thunder_Bay /etc/localtime > cp: cannot stat `/usr/share/zoneinfo/America/Thunder_Bay': No such > file or directory > > Hmm..., so I check... > > bash-3.00# ls -l /usr/share/zoneinfo/ > total 24 > -rw-r--r-- 1 root root 4029 May 12 13:47 iso3166.tab > -rw-r--r-- 1 root root 17591 May 12 13:47 zone.tab > For some reason the "zic" (ZoneInfo Compiler) has failed to run, which is what generates all the zoneinfo files. Check your config.log and installation logs for more clues. There was a thread about this some time (months?) back, but I don't remember the cause - -- Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: root user instead of LFS - Please help
On 28 May 2005, you wrote in lfs.support: > Hello All, > > In section 4.3 (Add the LFS user) of the book, I don't > want to add a user LFS. I'm automating my installation > using a simple bash script and when it does this: > > su - lfs > > My automation stops. :( I like to be able to continue > my installation in root. I am not interested in using > ALFS either. This is a school project. > Split your install into three scripts. As root, run script one script one runs su lfs -c
Re: root user instead of LFS - Please help
On 29 May 2005, you wrote in lfs.support: > Hi Steve, > > This is perfect! That's exactly what I want to do. > But, I'm having trouble running what you mentioned. > > In script1.sh I have: > su lfs -c ~/script2.sh > su root -c ~/script3.sh > > In script2.sh I have: > mkdir script2 > exit > > In script3.sh I have: > mkdir script3 > exit > > Is this what you mean for me to do? Thank you ALL for > your help & suggestions. > > Bobby trouble in what sense? are the directories not created? hint - put a "set -vx" at the top of each script, which will display the commands as they run. - -- Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: LFS suddenly not working!
On 05 Jun 2005, you wrote in lfs.support: > On Sat, 4 Jun 2005, Matthew Burgess wrote: > >> >> Ken, FWIW my scripts do the following to avoid scanning paths that won't >> (or shouldn't!) see any changes during a build: >> >> find . -xdev ! -path "./logs/*" ! -path "./home/*" ! -path "./proc/*" \ >>! -path "./sources/*" ! -path "./tmp/*" ! -path "./dev/*" \ >>! -path "./var/log/*" ! -path "./srv/*" \ >>-newer timestamp >> > > Cool. I've always found the findutils info pages a bit hard to > navigate, thanks for the example. Should save a bit of run time too. > > Ken I use the following script fu, originally from a script developed by Gerard from .profile: PRUNEPATH="/sys /dev/pts /proc /root /home /tmp /tools /usr/src" PRUNEREGEX=`echo $PRUNEPATH | sed -e 's,^,\\\(^,' -e 's, ,$\\\)\\\| \\\(^,g' -e 's,$,$\\\),'` in build scripts: touch /tmp/.timestamp echo "" >> /tmp/packages.log find / -regex "$PRUNEREGEX" -prune -o -cnewer /tmp/.timestamp -type f >> /tmp/packages.log - -- Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Can LFS be copied from 1 hard drive to another after the build?
On 15 Jun 2005, you wrote in lfs.support: > > On Wed, Jun 15, 2005 at 08:36:32AM +0100, Andrew Benton wrote: > >> I find tar to be quite effective >> >> cd /mnt >> tar cjf lfs.tar.bz2 lfs > > Rsync is cool if you have installed it. Mount the empty partition on > /mnt/lfs-new and run: > > rsync -a /mnt/lfs/ /mnt/lfs-new > > (NB: the trailing slash on /mnt/lfs/ is important, don't leave it > out!) > > Cheers, > > Jeremy Henty > > Just as a note for the unwary - make sure you don't have proc, sysfs, dev et al mounted in /mnt/lfs when you create the tarball or copy the files to the new partition ;) - -- Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Question on directory created
On 6/27/05, Stephen Liu <[EMAIL PROTECTED]> wrote: > Hi folks, > > I have 'LFS' partition permanently mounted as; > > # cat /etc/fstab > /dev/hda6 /mnt/lfsext3 > defaults1 2 okay, that means you have already created and mounted the LFS parition > > According to following URL > http://www.sg.linuxfromscratch.org/lfs/view/6.0/chapter02/mounting.html > > I proceeded > # mkdir -p $LFS /mnt/LFS > # mkdir /mnt/LFS/$LFS/usr > > # ls -al /mnt/LFS > total 24 > drwxr-xr-x 3 root root 4096 Jun 27 07:40 . > drwxr-xr-x 8 root root 4096 Jun 27 07:39 .. > drwxr-xr-x 2 root root 4096 Jun 27 07:40 usr > > Only /usr is there, '$LFS' not found. Whether > /mnt/LFS is '$LFS'? > Okay, you only need the mkdir /mnt/LFS/$LFS/usr if you are going to have your LFS /usr directories on a seperate filesystem, which does not appear to be the case, so you don't need to follow this step. Also, the instructions you are refering to in the URL above are for mounting the new LFS partition, which it appears you have already completed. - -- Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: 16 Bash Patches found
On 6/27/05, Joern Abatz <[EMAIL PROTECTED]> wrote: > I happened to search for 'bash' once on freshmeat.net, hit no.8 was > 'Gnu bash', the download location is ftp.cwru.edu/pub/bash, and there > are 16 patches there. I think maybe the download location in the book > is just not mirroring the directory. So it might actually be > bash-3.0.16 by now. > > Joern Bash does not change it's version number when applying patches - the LFS supplied bash-3.0-fixes-3.patch contains all the known official bash patches - you can find this patch in the LFS patches project -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Question on directory created
On 6/27/05, Stephen Liu <[EMAIL PROTECTED]> wrote: > > Whether jumping to following section; > 3. Packages and Patches > http://www.sg.linuxfromscratch.org/lfs/view/6.0/chapter03/chapter03.html > > to run > > # mkdir /mnt/$LFS/sources > # chmod a+wt /mnt/$LFS/sources > (what is the flag'a+wt'? Is it timestamp?) > First, the command shown in the URL is as follows: mkdir $LFS/sources you should not have the /mnt/$LFS/sources as you have typed. The $LFS is a variable that you should have set to equal /mnt/LFS - try the following and see if the output is correct. If not, you need to go back some steps and confirm you followed the instructions correctly regarding the LFS variable. echo $LFS should produce /mnt/LFS In answer to the specific question, a+wt is an option that is passed to the chmod command - specifcally, to add the w (write) and t (sticky bit) attributes to a (all = user, group and other). man chmod will give you more information. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Question on directory created
On 6/27/05, Stephen Liu <[EMAIL PROTECTED]> wrote: > Hi Steve, > > - snip - > > > # echo $LFS > Empty printout > > # echo LFS > LFS > > Whether to remove following entry from /etc/fstab > /dev/hda6/mnt/lfsext3defaults 1 2 > > and restart from > 2.4. Mounting the New Partition > http://www.sg.linuxfromscratch.org/lfs/view/6.0/chapter02/mounting.html > Sort of. Just unmount the existing LFS partition (umount /mnt/lfs) and start again from Chapter 2.4 - pay particular attention to this command export LFS=/mnt/lfs you should then do a echo $LFS and expect to get /mnt/lfs as the output. If not, let us know, otherwise you should be able to continue the commands in sequence. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Problem in running "ln -s $LFS/tools /"
On 6/28/05, Stephen Liu <[EMAIL PROTECTED]> wrote: > Hi Andrew, > > > # chown lfs $LFS/sources > chown: cannot access `/sources': No such file or > directory > > I'm held here again. Any advice? > When you "su" you load a new environment, and your $LFS variable is lost, so the chown is attempting to access /sources, instead of /mnt/lfs/sources. do a "export LFS=/mnt/lfs" and run the chown again. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Problem in running "ln -s $LFS/tools /"
On 6/28/05, Stephen Liu <[EMAIL PROTECTED]> wrote: > bash: umash: command not found > lfs:~$ > > Up to here I have no idea whether I have committed any > mistake because no indication of success. > > What is "bash: umash: command not found"? Is it a > WARNING? You've made a typo - that should be "umask", not "umash". When you run the "source" command, bash is reporting that it cannot find any command called "umash". -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Fail to create 'binutils-build' directory
On 6/29/05, Stephen Liu <[EMAIL PROTECTED]> wrote: > However I still don't understand how following 3 > hidden files exist on host FC3 > The Stage 1 (Chapter 5) building of LFS compiles and *installs* several packages that already exist on you FC3 host system. In order to prevent the possibility of damaging your host system, you are directed to create the LFS user (on your host system), and use this user to compile and install the software. Because the LFS user does not have permissions on your host system, you can't accidentally overwrite important components of your host system. The three "hidden" files are part of the login environment of the LFS user, which is created in the host system, so the files should appear where they are. Once you start Chapter 6, you will be using the chroot command to create an environment that will also protect your host system, although this time you will be the root user. Given the scope of your recent questions, I strongly suggest you should at a minimum view the Recommended Reading section of the LFS book, and read the documents referenced there - that way you will have a better understanding of the commands you are asked to run while building LFS. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: chroot problem (Chapter 6.3)
On 6/30/05, David Ciecierski <[EMAIL PROTECTED]> wrote: > PS: Just for the record: I followed the instructions exactly except I > used GCC 3.4.4, binutils 2.16.1 and Glibc 2.3.5 with apropriate patches > obtained from LFS Development site. Using GLIBC2.3.5 requires an additional patch for bash - the WCONTINUED patch, which is in the LFS patches repository - recompile bash with that patch and all should be fine. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Held on 5.9. Adjusting the Toolchain
On 6/30/05, Stephen Liu <[EMAIL PROTECTED]> wrote: > Hi folks, > > I encountered problem on > > 5.9. Adjusting the Toolchain > http://www.sg.linuxfromscratch.org/lfs/view/6.0/chapter05/adjusting.html > > I tried following steps; > > lfs:/mnt/lfs/sources/glibc-build$ make -C ld install > make: *** ld: No such file or directory. Stop. > on which directory I have to run the command line. Directly above the make -C ld install command is the following sentence: The linker, adjusted at the end of the first pass of Binutils, is installed by running the following command from within the binutils-build directory: -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: lfs 6.0 and bash 3.0: should i apply the patch?
On 7/2/05, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote: > I'm using rev 6.0 of the book. I'm in chapter 5. > > Section 5.29. Bash-3.0 doesn't say to apply the patch, but when I > downloaded the required patches it came with two: > >bash-3.0-avoid_WCONTINUED-1.patch >bash-3.0-fixes-1.patch > > Should they be applied or not? > bash-3.0-WCONTINUED fixes a bug with glibc2.3.5 and bash - since LFS6.0 installs glibc2.3.4, it's not strictly necessary, but will not hurt. bash-fixes contains 16 or so different patches for bugs in the bash code. Although not required to complete the build, it is recommended. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Backtracking HotPlug and Udev
On 7/2/05, Matthew Burgess <[EMAIL PROTECTED]> wrote: > rules file should be safe enough to do. udev-059 was recently released, > though it included major changes which we haven't been able to assess > the affects of yet, so I'd strongly suggest you stick with udev-058 for now. > just building now with udev-059. One change is that udevstart now launches udev as a process (with a --daemon flag), so you can't umount /dev without stopping that process (/dev is reported as busy), so we might need to modify the "Booting into LFS" section. Other than some new significant rule capbilities, it looks like the new version will work in our current setup - if I find any additional issues I report them. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Backtracking HotPlug and Udev (SOLVED)
On 7/3/05, Matthew Burgess <[EMAIL PROTECTED]> wrote: > Well, this is an interesting point. I was going to point you at > http://www.linuxfromscratch.org/lfs/view/6.1-pre1/chapter07/udev.html > (7.4.2, paragraph 4 in particular), but that completely fails to mention > how the driver modules are actually loaded, in order to register > themselves in /sys). > > Anyone care to hit me with the cluebat on this one? So far the chain of > events is: > >From udev-047, udevd is capable of handling hotplug events from modules, as well as devices, and knows to call the appropriate hotplug rc script related to the module. (at module load, a hotplug event is generated by the kernel, and calls whatever is defined as the hotplug handler). The udev package contains some (out of date) docs regarding event handling and modules - the Changelog specifies the fact that udevd now handles module events. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: patch versions do not match source versions?
On 7/4/05, sean machin <[EMAIL PROTECTED]> wrote: > Hi All, > > > http://www.linuxfromscratch.org/patches/lfs/6.0/gcc-3.4.1-specs-1.patch > > seem to apply to gcc version 3.4.0 NOT 3.4.1 which is the version > I am trying to build. > > What should I do about this? Try to find patches which match the As long as the *name* of the patch matches the version you are building, you will be fine. I believe you are referring to the comments in the patch itself, which indicate the patch was initially against 3.4.0? - this is correct, but the patch should apply fine to 3.4.1 (although there may be some "fuzz" or "offsets" - this is normal, and does not indicate a failure). > > Also, am I correct in assuming that all the patches listed > in chapter 3 should be applied before building each package? > No. Only apply the patches as directed by the book (each package install section indicates if a patch is required. The reason is that some patches are only applicable later in the book, or have no benefit in earlier sections. The instructions are fairly explicit about when a patch is needed however. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Problems during compilation of binutils ...
On 7/4/05, Shane Shields <[EMAIL PROTECTED]> wrote: > R.Grasbon wrote: > > > && rm -f $file && PATH=../src:$PATH no -o $file > > Looks to me to be a typo here > The configure script for binutils has replaced a missing gettext command with the value "no". Either your host has an incomplete gettext installation, or it's broken. You can either install gettext on your host distribution, or use --disable-nls on the configure command line for binutils if you don't need language support. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: I have no name!
On 7/8/05, Stephen Liu <[EMAIL PROTECTED]> wrote: > Hi folks, > > Continued on "6.3. Entering the Chroot Environment" > http://www.sg.linuxfromscratch.org/...r06/chroot.html > > > I have no name!:/# > > Please advise how to proceed. Whether enter a name > here? OR just ignore it and continue on "6.4. Changing > Ownership" > Look at the last paragraph on the section you just finished (6.3) Note that the bash prompt will say "I have no name!" This is normal because the /etc/passwd file has not been created yet. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: How to build an entire LFS system using uclibc instead of glibc
On 7/10/05, Razvan Cristian <[EMAIL PROTECTED]> wrote: > I have built a LFS system, and I would like to find > out how I could build it using uclibc instead of > glibc. > > I'm not interested on putting it on a floppy or > something, I just want to build a system which should > be as large as 30-40 MB. > As I've seen, binaries compiled against uclibc are > very small. You want a full LFS system in 30-40Mb? I'm not sure even uclibc will give you that. I have a dedicated firewall running mostly LFS (no gettext) with webserver/proxy and various IDS tools that has had gcc, headers, et al removed, and it comes in at 24Mb with glibc (35Mb normally, but I butcher the perl libraries to only have needed files, and also remove every binary I don't need). > If there's a how to on replacing the glibc used in the > LFS book with uClibc library, please tell me. > Did anyone try this? How big was the system? > Which are the necessary steps to achieve this? The Hardened Linux From Scratch (HLFS) project has a early alpha book on how to use uClibc with LFS - you can skip the "security" bits (SSP, etc) and see if that suits your needs. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: udevsend and hotplug
On 7/12/05, kareemy <[EMAIL PROTECTED]> wrote:
> On 7/11/05, Dan McGhee <[EMAIL PROTECTED]> wrote:
> > kareemy wrote:
> >
>
> Thank you. I have read that. It provided some insight as to how
> udev+hotplug are supposed to work but didn't quite solve my issue. In
> the meantime I found two possible solutions that have worked for me.
>
> 1) Keep /sbin/udevsend as the hotplug handler. Compile the
> run_directory extras in the udev source code by issuing the command
> "make EXTRAS=extras/run_directory". This will generate a
> udev_run_hotplugd binary. Copy that binary to /sbin and add the
> following rule to 50-udev.rules:
>
> ENV{UDEVD_EVENT}=="1", RUN+="/sbin/udev_run_hotplugd"
>
This is the correct option - from udev-059 onwards, the hotplug
handling is being done differently, with the udev rules being expanded
to allow calling of hotplug scripts from the udev rules file. It will
take a while to get a properly grip on the new functionality.
Ideally, the hotplug scripts will go away entirely, and be replaced by
udev rules that can call modprobe directly, etc. - that will take some
time however.
-- -
Steve Crosby
--
http://linuxfromscratch.org/mailman/listinfo/lfs-support
FAQ: http://www.linuxfromscratch.org/lfs/faq.html
Unsubscribe: See the above information page
Re: Help me to build LFS
On 7/17/05, Stephen Liu <[EMAIL PROTECTED]> wrote: > Hi bolt, > > I have been trying to build LFS 6.0 on a FC3 host box > without success. At this juction I can't say the > problems coming from the host. I solved problem after > problem on building LFS 6.0. Finally I decided In fairness to the LFS project, the majority of the issues you reported to the list were of your own making, specifically missing steps clearly outlined in the book, and leaving and failing to re-enter the build environment properly on numerous occasions, causing the host contamination you experienced. It was suggested several times to review the "Recommended Reading" list for LFS to increase your understanding of the basic commands and concepts that need to be understood before you start an LFS build. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Minimal installation
On 7/22/05, Andrei Dore <[EMAIL PROTECTED]> wrote:
> quite afraid that I will be more over 100 Mb when I
> will finish :(. As I need the whole system to run on a
> 256 MB compact flash, this will be a real problem.
> As far as I've understood, the key is to remove
> unneeded packages.
You can remove the following, depending on what you want to do:
+++TimeZone and Locales (if you only need english)
- (copy the one needed timezone to /etc instead of ln -s)
/usr/share/{zoneinfo,locale}
/usr/lib/locale
/usr/share/i18n
+++Compiling tools (if you will not be building any packages from source)
/usr/share/{automake*,aclocal*,autoconf,libtool}
/usr/bin/{make,automake*,aclocal*,autoconf,autom4te,autoheader}
/usr/bin/{autoreconf,autoscan,autoupdate,ifnamesi,libtool*}
/{,usr/}lib/*.la
/usr/lib/{crt*.o,?crt1.o,gcc-lib}
/lib/cpp
/usr/bin/{c++*,g++,cc,gcov,gcc*,cpp,i?86-pc-linux-{g++,c++,gcc*}}
/usr/include
/usr/lib/*.a
+++Documentation
/usr/share/man
/usr/man
/usr/share/tex{info,mf}
/usr/share/groff
/usr/share/{doc,info}
Those are the major ones, but there are hundreds of other files that
can be removed, especially in /{,usr/}{s}bin - use the man pages to
decide if you need it or not for your finished system. And if you
don't need Perl, thats another 35Mb or so that you can "save". Note of
caution - some of the bootscripts use commands in /bin, so
double-check them before you remove commands you don't need - the boot
scripts may need them too.
With some other minor enhancements, I have a system running at 24Mb
from CD-ROM that has a DNS\DHCP server, IPtables firewall, ssh server,
web server and web proxy amongst other things - by using uClibc and
converting some perl apps to native C code, I expect to get that below
15Mb easily.
NOTE: Once you have removed the compiler tools, you are stuck with the
packages you have now (unless you use a binary installer) - so make
sure you have installed everything you want!
One day, I'll write a hint about how I made this ;)
-- -
Steve Crosby
--
http://linuxfromscratch.org/mailman/listinfo/lfs-support
FAQ: http://www.linuxfromscratch.org/lfs/faq.html
Unsubscribe: See the above information page
Re: /tools/bin/env: /tools/bin/bash: No such file or directory
On 7/23/05, Stephen Liu <[EMAIL PROTECTED]> wrote: > Hi Chris, > > > > # ldd /tools/bin/bash > > > linux-gate.so.1 => (0x00d2) > > > libtermcap.so.2 => /lib/libtermcap.so.2 > > > (0x00ac4000) > > > libdl.so.2 => /lib/libdl.so.2 (0x009d3000) > > > libc.so.6 => /lib/tls/libc.so.6 > > (0x00882000) > > > /lib/ld-linux.so.2 (0x00868000) > > > > I assume you were logged on as the lfs user when you > > did this...? If so, > > that's not good - nothing in /tools should be > > pointing to anything in /lib. > > Following steps were performed after starting the PC > Stephen, When you built the gcc compiler in Chapter 5, there was a step that changed the compiler so applications built after gcc would use the /tools directory. This step has not worked in your case, since the bash executable is quite clearly using the host libraries (which are not available in the chroot step). You need to (as Chris pointed out), start from Chapter 5 again, and try and see where you went wrong. Points to check include the "adjusting the linker" step, and the "setup LFS user environment" step. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: internet connection with liveccd
On 7/24/05, Jeremy Huntwork <[EMAIL PROTECTED]> wrote: > Alexander E. Patrakov wrote: > > it is hard to do so with the unknown future RedHat extensions in mind, > > but it either has to be done or too-new hosts should be blacklisted > > (thus making LFS live CD the only officially supported host - but that's > > against the spirit of LFS). > > To some extent, I agree with your suggestion here, Alexander. Since FC > came out, there's been several run-ins with LFS. I think, to a large > extent, simply because FC is too bleeding-edge. And though LFS trunk > tries to stay up-to-speed, it takes longer for us to successfully > incorporate new technology than it does for a project like FC to produce > and release (for several reasons). So, at the least, I think it might > be a good idea to discourage using the latest version of FC somewhere in > the book. Another option is to apply the FC2\3 workaround permanently - change the book instructions to download, compile and use e2fsprogs from LFS to create the LFS partition always, and never use the host tools. Although that would work for all e2fs problems, I can forsee other FC related tool issues, so might not be as useful as it could be. Yet Another Option: Have a "Host Quirks" page - document known Host based issues and suggested workarounds - we could maintain that as an online only page referred to in the book (Check the page for known issues with your Host system before building) - this would be fairly easy to maintain, since the number of quirks is not high, and we could test and update relatively easily when a new host OS is released. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: internet connection with liveccd
On 7/25/05, Matthew Burgess <[EMAIL PROTECTED]> wrote: > Gerard Beekmans wrote: > > steve crosby wrote: > > > >> Another option is to apply the FC2\3 workaround permanently - change > >> the book instructions to download, compile and use e2fsprogs from LFS > >> to create the LFS partition always, and never use the host tools. > >> > >> Although that would work for all e2fs problems, I can forsee other FC > >> related tool issues, so might not be as useful as it could be. > > > > However, at least it would be documented. That's more than we can say > > right now. It might miss other issues and those should simply be added > > as they are discovered. > > Err, I thought this particular issue was already covered adequately by > the "Note" box at > http://www.linuxfromscratch.org/lfs/view/6.1/chapter02/creatingfilesystem.html. > Yes, but given the frequency of the issue with recent FC releases, it may be better to *always* apply the workaround, regardless of the host - which will eliminate the problem completely. We can then add other workarounds as required. At least, that's what I think Gerard meant. >From an education standpoint, it gives the reader exposure to the c;m;mi model eary (before they start building LFS), although we already expect that level of knowledge from an LFS reader. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Re: small LFS distribution
On 8/9/05, Feng Zhou <[EMAIL PROTECTED]> wrote: > > Feng Zhou wrote these words on 08/08/05 00:31 CST: > > > > > I try to build a LFS linux system on disk-on-chip (under 128M) with > > > apache, php, perl support. I need help to do so. > > > > What have you tried to do so far? > > I have tried using LFS6.1. But It's too big to fit the distribution in > 128M system. > > > > Where are you having trouble? > > I am not sure what are min packages to install in order to make the system > run. > > > > What *exactly* do you "need help to do so" with? > > I am looking for the howto to build Embedded Linux Systems using LFS. > > There is no specific hint available for that. Look in the hints section of the linuxfromscratch website for the "stripped down LFS", which will describe things you can do to reduce the final size of a LFS system. You will also have to be more specific about the end system - do you need it to compile things - if not then gcc, the auto and libtools, bison and flex are not required - however you should still build them, and remove the files at the end, otherwise you can impact the actual building process. With care, and depending on what your end system needs to be functional for your purpose, it's possible to get a standard LFS6.1 system stripped down to less than 20Mb, plus whatever additional package your system needs. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: how to compile NPTL using gcc 3.3.3 in 2.4.18 kernel
On 8/9/05, yang lin <[EMAIL PROTECTED]> wrote: > then I have a question: in RH9 with kernel > 2.4.20,/lib/tls/libpthread.so.0 print the info : > "NPTL 0.29 by Ulrich Drepper > it seem that the RH9 support the NPTL threadlib on gcc 3.2.2 in kernel 2.4.20 > > so ,why cannt lfs-5.1.1 support it ? Because RedHat pays full time programmers to modify the linux kernel to support NTPL - in doing so, they also modify glibc, gcc, and many other packages so they all work together (mostly) nicely. This is Linux From Scratch, not RedHat From Scratch, so thats why we don't provide support for using NTPL with Linux 2.4.x kernels. The effort to include the changes RedHat (and other distributions) provide to stock packages is simply too much for a very small volunteer community. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: file permissions
On 8/11/05, Michiel Faber <[EMAIL PROTECTED]> wrote: > > this is what i do: > > rm -R patch-2.5.4 > use (with care!) rm -rf -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: file permissions
On 8/12/05, Michiel Faber <[EMAIL PROTECTED]> wrote: > Thanks for the help. > > rm -rf indeed worked but is not a solution to the wrong (i think at > least) file permisions. > Even without configure, make and make install, the questions pop up. > Umask is set to 022. > > Can i solve the problem of file permissions after untarring a file? > You can't fix the tarball without extracting it, fixing the permissions (as below) and then re-creating the tarfile. To "fix" the permissions, use chmod -R +w . That will add the "write" permission bit to all files in all subdirectories, and thus an rm -r will not ask before removing it. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: make check failed first time, passed second time with no change
On 8/15/05, Doug Ronne <[EMAIL PROTECTED]> wrote: > I am using Version 7.0-cross-lfs-20050814-x86_64 > > In section 10.3 when I ran make check, it failed on libc.cat (in > catgets/de/libc.cat). I went away, accidentally closed the terminal, > and ran it again to get the message so that I could send it in... and > it passed. Does this mean I am good to go? And why would it fail > once and pass a second time? > the make command simply continues from where it last stopped - it doesn't rerun the previous tests unless you manually tell it to. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Kernel Config Question: SMP support?
On 8/17/05, Joseph M Dupre (AVAB Inc.) <[EMAIL PROTECTED]> wrote: > I have a Dell sc420, which, as far as I can tell, only has one > processor. There is no spot on the motherboard for a second > processor. > > However, if I disable SMP support, the kernel issues a warning: > kernel: WARNING: NR_CPUS limit of 1 reached. Processor ignored. > > If I enable SMP support, the kernel initializes not one, but two > processors: > kernel: Total of 2 processors activated (11075.58 BogoMIPS). > > What is this second processor??? Where is this SMP MP-table that the > kernel is finding, and why is it on a single processor machine? > Newer Intel Pentium 4 chips have a feature known as HyperThreading - look on google for lots of info. The result is the kernel sees two CPU's - enabling SMP is required to use the HyperThreading capability. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: sysklogd question
On 8/18/05, David Fix <[EMAIL PROTECTED]> wrote: > How can I go about NOT having the stuff show up in the sys.log file, yet > appear in the cron.log file? Here is what my /etc/syslog.conf looks like: > > *.*;auth,authpriv.none -/var/log/sys.log In the line above, do the following: *.*;auth,authpriv,cron.none -/var/log/sys.log This means log everytyhing (*.*), and also log auth, authpriv and cron logs at a level of "none" - which effectivley means don't log them at all. man syslog.conf for more information on rules and exclusions. -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Final step 9.3 then no reboot
NFIG_HW_CONSOLE=y # CONFIG_SERIAL_NONSTANDARD is not set # # Serial drivers # # CONFIG_SERIAL_8250 is not set # # Non-8250 serial port support # CONFIG_UNIX98_PTYS=y # CONFIG_LEGACY_PTYS is not set # # IPMI # # CONFIG_IPMI_HANDLER is not set # # Watchdog Cards # # CONFIG_WATCHDOG is not set # CONFIG_HW_RANDOM is not set # CONFIG_NVRAM is not set # CONFIG_RTC is not set # CONFIG_GEN_RTC is not set # CONFIG_DTLK is not set # CONFIG_R3964 is not set # CONFIG_APPLICOM is not set # # Ftape, the floppy tape device driver # # CONFIG_FTAPE is not set # CONFIG_AGP is not set # CONFIG_DRM is not set # CONFIG_MWAVE is not set # CONFIG_RAW_DRIVER is not set # CONFIG_HANGCHECK_TIMER is not set # # I2C support # # CONFIG_I2C is not set # # Dallas's 1-wire bus # # CONFIG_W1 is not set # # Misc devices # # # Multimedia devices # # CONFIG_VIDEO_DEV is not set # # Digital Video Broadcasting Devices # # CONFIG_DVB is not set # # Graphics support # # CONFIG_FB is not set # CONFIG_VIDEO_SELECT is not set # # Console display driver support # CONFIG_VGA_CONSOLE=y CONFIG_DUMMY_CONSOLE=y # # Sound # # CONFIG_SOUND is not set # # USB support # # CONFIG_USB is not set CONFIG_USB_ARCH_HAS_HCD=y CONFIG_USB_ARCH_HAS_OHCI=y # # NOTE: USB_STORAGE enables SCSI, and 'SCSI disk support' may also be needed; see USB_STORAGE Help for more information # # # USB Gadget Support # # CONFIG_USB_GADGET is not set # # MMC/SD Card support # # CONFIG_MMC is not set # # File systems # CONFIG_EXT2_FS=y # CONFIG_EXT2_FS_XATTR is not set # CONFIG_EXT3_FS is not set # CONFIG_JBD is not set # CONFIG_REISERFS_FS is not set # CONFIG_JFS_FS is not set # CONFIG_XFS_FS is not set # CONFIG_MINIX_FS is not set # CONFIG_ROMFS_FS is not set # CONFIG_QUOTA is not set CONFIG_DNOTIFY=y # CONFIG_AUTOFS_FS is not set # CONFIG_AUTOFS4_FS is not set # # CD-ROM/DVD Filesystems # CONFIG_ISO9660_FS=y CONFIG_JOLIET=y # CONFIG_ZISOFS is not set CONFIG_UDF_FS=y CONFIG_UDF_NLS=y # # DOS/FAT/NT Filesystems # # CONFIG_MSDOS_FS is not set # CONFIG_VFAT_FS is not set # CONFIG_NTFS_FS is not set # # Pseudo filesystems # CONFIG_PROC_FS=y CONFIG_PROC_KCORE=y CONFIG_SYSFS=y # CONFIG_DEVPTS_FS_XATTR is not set CONFIG_TMPFS=y # CONFIG_TMPFS_XATTR is not set # CONFIG_HUGETLBFS is not set # CONFIG_HUGETLB_PAGE is not set CONFIG_RAMFS=y # # Miscellaneous filesystems # # CONFIG_HFSPLUS_FS is not set # CONFIG_CRAMFS is not set # CONFIG_VXFS_FS is not set # CONFIG_HPFS_FS is not set # CONFIG_QNX4FS_FS is not set # CONFIG_SYSV_FS is not set # CONFIG_UFS_FS is not set # # Network File Systems # # CONFIG_NFS_FS is not set # CONFIG_NFSD is not set # CONFIG_EXPORTFS is not set # CONFIG_SMB_FS is not set # CONFIG_CIFS is not set # CONFIG_NCP_FS is not set # CONFIG_CODA_FS is not set # # Partition Types # # CONFIG_PARTITION_ADVANCED is not set CONFIG_MSDOS_PARTITION=y # # Native Language Support # CONFIG_NLS=y CONFIG_NLS_DEFAULT="iso8859-1" CONFIG_NLS_CODEPAGE_437=y # CONFIG_NLS_CODEPAGE_737 is not set # CONFIG_NLS_CODEPAGE_775 is not set # CONFIG_NLS_CODEPAGE_850 is not set # CONFIG_NLS_CODEPAGE_852 is not set # CONFIG_NLS_CODEPAGE_855 is not set # CONFIG_NLS_CODEPAGE_857 is not set # CONFIG_NLS_CODEPAGE_860 is not set # CONFIG_NLS_CODEPAGE_861 is not set # CONFIG_NLS_CODEPAGE_862 is not set # CONFIG_NLS_CODEPAGE_863 is not set # CONFIG_NLS_CODEPAGE_864 is not set # CONFIG_NLS_CODEPAGE_865 is not set # CONFIG_NLS_CODEPAGE_866 is not set # CONFIG_NLS_CODEPAGE_869 is not set # CONFIG_NLS_CODEPAGE_936 is not set # CONFIG_NLS_CODEPAGE_950 is not set # CONFIG_NLS_CODEPAGE_932 is not set # CONFIG_NLS_CODEPAGE_949 is not set # CONFIG_NLS_CODEPAGE_874 is not set # CONFIG_NLS_ISO8859_8 is not set # CONFIG_NLS_CODEPAGE_1250 is not set # CONFIG_NLS_CODEPAGE_1251 is not set # CONFIG_NLS_ASCII is not set CONFIG_NLS_ISO8859_1=y # CONFIG_NLS_ISO8859_2 is not set # CONFIG_NLS_ISO8859_3 is not set # CONFIG_NLS_ISO8859_4 is not set # CONFIG_NLS_ISO8859_5 is not set # CONFIG_NLS_ISO8859_6 is not set # CONFIG_NLS_ISO8859_7 is not set # CONFIG_NLS_ISO8859_9 is not set # CONFIG_NLS_ISO8859_13 is not set # CONFIG_NLS_ISO8859_14 is not set # CONFIG_NLS_ISO8859_15 is not set # CONFIG_NLS_KOI8_R is not set # CONFIG_NLS_KOI8_U is not set # CONFIG_NLS_UTF8 is not set # # Kernel hacking # # CONFIG_DEBUG_KERNEL is not set CONFIG_FRAME_POINTER=y CONFIG_EARLY_PRINTK=y CONFIG_4KSTACKS=y # # Security options # # CONFIG_KEYS is not set # CONFIG_SECURITY is not set # # Cryptographic options # # CONFIG_CRYPTO is not set # # Library routines # # CONFIG_CRC_CCITT is not set CONFIG_CRC32=y CONFIG_LIBCRC32C=y CONFIG_GENERIC_HARDIRQS=y CONFIG_GENERIC_IRQ_PROBE=y CONFIG_X86_BIOS_REBOOT=y CONFIG_PC=y -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: ip forwarding and iptables
On 5/16/06, Andrew Benton <[EMAIL PROTECTED]> wrote: Angel Tsankov wrote: > I have 2 PCs: one configured as gateway (PC1) and the other one (PC2) > configured to use PC1 as gateway. PC1 runs a LFS. It has ip forwarding > enabled (e.g. by echo 'net.ipv4.ip_forward = 1' >> /etc/sysctl.conf). > As far as I understand, I do not need to do anything else to make the > kernel route traffic to and from PC2, right? No, you need to use iptables to handle NAT/masquerading Although this is the wrong forum (BLFS is where you need to ask these questions, as Andy pointed out), I thought I'd correct some of the statements for the archives. Enabling the ip_forwarding sysctl turns your PC1 into a router. If one of the networks that PC1 is connected to is the Internet, then there may be additional work involved on PC1 to allow traffic to enter and exit the Internet network properly (such as NAT). This depends on your network configuration, target networks, routers and address space, so can't be answered as a simple yes or no. > However, if I have one PC more - PC3, and I do not wnat to route traffic > to and from it I need to configure the kernel, e.g. with > the help of iptables. Now if I do so, i.e. use iptables to configure the > kernel, save the iptables configuration, setup the system > to reload it at startup (using the init.d scripts), is there any moment > (during system startup) when ip forwarding has been enabled > but the iptables configuration has not yet been loaded and traffic could > be routed to and from PC3? No, without iptables nothing would be forwarded. If you're worried about it then don't do the echo 1 > /proc/sys/net/ipv4/ip_forward until after you've set your firewall rules. Yes. Once ip_forwarding has been turned on, your PC1 is a router, and will happily deliver traffic from and from any machines. Whether or not the traffic does anything, can get a response or is routed correctly may depend on additional configuration however. The correct solution is to enable ip_forwarding only after you have made any additional configuration changes to your PC1 networking (such as iptables, etc). -- -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Are the gcc build files still needed after section 5.4?
On 6/5/06, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote: On Sun, 4 Jun 2006, Randy McMurchy wrote: > [EMAIL PROTECTED] wrote these words on 06/04/06 19:02 CST: > >> On the other hand the link created at the end of section 5.4 >> points /tools/bin/cc at the gcc in the working directory, not at >> /tools/bin/gcc. > > Are you certain about that? :-) Dead certain. Look: You've made one of the common mistakes (and this is\was in the FAQ) in misunderstanding the ln command. Specifically that relative paths in a symlink use the path of the link as the search path, not the current working directory. so ln -sv gcc /tools/bin/cc will result in a /tools/bin/cc symlink pointing to gcc (which, since the cc file is in /tools/bin, means effectively /tools/bin/gcc will be used) as per the following console extract. # mkdir -p /tools/bin # touch /tools/bin/gcc # ln -sv gcc /tools/bin/cc create symbolic link `/tools/bin/cc' to `gcc' # ls -l /tools/bin total 0 lrwxrwxrwx1 root root3 Jun 5 20:05 cc -> gcc -rw-r--r--1 root root0 Jun 5 20:03 gcc -- -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Are the gcc build files still needed after section 5.4?
On 6/6/06, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote: > Confusion Re-read the bit about search path for sources. That's the bit that's not in the man. The syntax is also different between hard and soft links, but we won't go into that (yet!). I may have introduced some confusion - hope the following clarifies things. You seem to be confusing two seperate things about how symlinks work: 1. *where* the symlink file is placed in the filesystem by the ln command 2. *how* the symlink is resolved by the Operating System. For 1), the type of path you provide the ln command determines *where* the resulting symlink file is created. For example cd / mkdir -p tmp/ln-test touch /tmp/ln-test/afile now we have a file in /tmp/ln-test called "afile" create a link to this file two different ways, using absolute and relative paths Relative 1: === cd / ln -s afile tmp/ln-test/bfile This creates link bfile --> afile in directory /tmp/ln-test Relative 2: === cd /usr/bin ln -s afile ../../tmp/ln-test/bfile This *also* creates link bfile --> afile in directory /tmp/ln-test Absolute: === cd / ln -s afile /tmp/ln-test/bfile This always creates link bfile --> afile in directory /tmp/ln-test, regardless of your current working directory. Therefore, the location of the symlink file is determined by the path you use. If it's a relative path (not starting with a /) then it will be relative to your current working directory. For 2) The method used to resolve the symlink by the OS, the choice of relative and absolute paths is also relavant. For example /tools/bin/cc --> gcc /tools/bin/cc --> ../../tools/bin/gcc /tools/bin/cc --> /tools/bin/gcc All three of these symlinks point to the *same* file (/tools/bin/gcc). The relative paths are relative to the *location of the symlink file*, not your current working directory. If you were to move the first two symlinks to another directory, they would no longer work (they are then broken or dangling symlinks) - the thrid example will always work regardless of *where* the symlink file is. All three will always work regardless of your current working directory. The discussion of hard vs soft links is way outside the scope of this email thread - suffice it to say the two are very different, and have different rules (both in how they work, and in how they are created). Hope that clears things up - the man page for ln refers to *where* the ln command will place the symlink file - I don't think the ln manpage does (or should) cover how the OS resolves symlinks. -- -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: gcc-4.1.1 pass2 cannot find limits.h
On 6/30/06, f. l. <[EMAIL PROTECTED]> wrote: +#define DYNAMIC_LINKER "/tool/lib/ld-linux.so.2" assuming it's not a typo, this should be "tools", not "tool", correct? -- -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: No http/ftp program included in regular build?
On 12/5/06, Rainer Peter Feller
<[EMAIL PROTECTED]> wrote:
On Tue, 2006-12-05 at 02:12 -0500, Colin Dean wrote:
> Chris Staub wrote:
> > Colin Dean wrote:
> >> I recently finished a build of LFS 6.2, and quickly realized that
> >> there is no way to get the packages for BLFS or other packages from
> >> within a fresh built LFS environment.
> >>
> > LFS includes inetutils, which has several internet programs, including ftp.
>
> Ah, I see that now--having FTP is definitely a boon, but aren't the
> majority of packages available via HTTP?
your distro your rules ...
just install wget ... and yes may be it should be a component of lfs :-)
just use this piece of script fu for http file downloading - I really
should troll the archives and find out the proper attribution for this
code, but it never seems I have enough time ;)
< add to /etc/profile >---
# Function to get the contents of a URL and display to STDOUT
httpget () {
h=${1#http://}
h=${h%%/*}
p=${h#*:}
if [ "$p" = "$h" ]; then
p='80'
fi
f=${1#http://*/}
if [ "$f" = "$1" ]; then
f='/'
else
if [ -z "$f" ]; then
f='/'
else
f="/$f"
fi
fi
( echo -e "GET $f HTTP/1.0\r\nHost: $h\r\n\r\n" 1>&3 & cat 0<&3 )
3<>/dev/tcp/$h/$p \
| (read i; while [ "$(echo $i | tr -d '\r')" != "" ]; do read i; done; cat)
}
# Export the function we have created
export -f httpget
---
the after login\sourceing profile, you can use
httpget http://my.http.file.location/filename.extenstion > filename.extension
no additional binaries required.
--
-- -
Steve Crosby
--
http://linuxfromscratch.org/mailman/listinfo/lfs-support
FAQ: http://www.linuxfromscratch.org/lfs/faq.html
Unsubscribe: See the above information page
Re: No http/ftp program included in regular build?
On 12/5/06, steve crosby <[EMAIL PROTECTED]> wrote: just use this piece of script fu for http file downloading - I really should troll the archives and find out the proper attribution for this code, but it never seems I have enough time ;) made the time - kudos to the original poster, although no idea where he\she got it from ;) http://archives.linuxfromscratch.org/mail-archives/lfs-dev/2002-June/026490.html -- -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: How to set $PATH for fcron jobs?
On 12/30/06, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote: > On 26/12/06 17:28:18, Ag. Hatzimanikas wrote: > > On Tue, Dec 26, at 11:26 [EMAIL PROTECTED] wrote: > > > > > > I'm failing at figuring out how to set $PATH for fcron jobs. I want > > > to be > > > able to start scripts that depend on a modified $PATH. If I'm not > > > mistaken > > > fcron runs as the special user fcron and then changes to the user > > > it is > > Ah, thanks. I thought I'd have to get login set the PATH. Still I'm wondering > where the default PATH is coming from before any bash_profile or bashrc is > read. Some where I read it would be compiled in, just didn't say where. > Sorry for the wrong list, as I was lead to login I decided to take lfs. I'm not sure specifically about fcron, since I don't have the source handy, but most cron implementations *deliberately* reset the executing environment to null for security reasons - that way any nefarious individual cannot modify the default system path and cause a cron job to execute an unexpected action because it's environment has changed. Just manually setup your environment for each script, or better yet, never assume a command location and always hardcode the full path - that way if something moves, you do an upgrade, etc, the script will break rather than silently choose a new binary you weren't expecting it to. Also, the default setting of the path is set by either login (part of shadow, and thus using \etc\login.def to define the default PATH) or sshd (for remote logins only) -- -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Spring Ahead
On 2/3/07, Ken Moffat <[EMAIL PROTECTED]> wrote: > On Fri, Feb 02, 2007 at 04:44:00PM +0100, Barius Drubeck wrote: > > > > Ken's suggestion sounds like a good one, although I would do it > > slightly differently. Build glibc-2.3.6 with the _same_ prefix as > do it. The specific problem is that the construction of the > zoneinfo files is fairly well hidden, at least in my logs, and I > think they are a product of make install. Of course, if they are > there after 'make' then it is even easier. > man zic (timezone compiler) just take your existing zoneinfo source, add the new rules, and compile the files - no need for new glibc install ;) -- -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Dhcpcd and iptables problems (BLFS 6.1)
On 2/8/07, Darcy Roberts <[EMAIL PROTECTED]> wrote: > I've turned on the appropriate modules in the kernel and all appears correct > now. One issue is that the style of the options - everyone shows > "CONFIG_IP_NF_IPTABLES=y" but this isn't the same text as when using > menuconfig. I ended up manually editting .config (bad idea, since ugly > dependencies for options now show up). > Defintely can create issues with dependencies, use the menu based config where you can. If you select the help option for each kernel config item (i.e. ? instead of Y\N\M for make menuconfig) the help will usually tell you the config item name. What you need to have is dependant on how complex your iptables rules are going to be. For example, if your going to be using NAT, you will need some or all of the NAT options - if your never going to adjust the TCPMSS value using iptables, you don't need those option in the kernel. That said, it adds very little additional overhead to the kernel to just select every single option (m for modular, y for static [always built into kernel]) - if your iptables ruleset doesn't need a function it won't be loaded if you selected modules. Note that the latest kernel release (2.6.20) has just significantly re-arranged the netfilter options menu again, obseleting a number of items, and adding new ones. > BTW, how do I limit the size of the kern and sys logs? grab logrotate from the debian ftp mirrors. you run logrotate every so often (usually daily) in cron (you'll need to install the cron daemon too!) and the logrotate configuration files indicate after how long, what size etc to chop the logs smaller, keep x number of backup copies, compress the backups, etc. -- -- - Steve Crosby -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
