Re: 7 MB linux with lighttpd & Busybox
satish patel wrote: > I have created small 7 MB linux with lighttpd and busybox installed I have Very impressive. Good job! Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- 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
Valter Douglas Lisbôa Jr. wrote: > Em Tuesday 30 September 2008 11:04:06 Valter Douglas Lisbôa Jr. escreveu: >> Em Monday 29 September 2008 09:47:35 Jeremy Henty escreveu: >> I try to find by man, but it's poorly documented, I did think for a while >> that /etc/defaults/useradd could change this behaviour, but no. My file >> have GROUP=100 and 100 is the GID of group users and it always create a new >> group if the -u is nor expecified. >> >> I try a strings on /usr/sbin/useradd searching a possible parameter >> on /etc/login.defs, but no clear hint how to change it by this way. Going >> to source code... >> >> -- >> Valter Douglas Lisbôa Jr. >> Sócio-Diretor >> Trenix - IT Solutions >> "Nossas Idéias, suas Soluções!" Is that "Our Ideas, your Solutions"? >> www.trenix.com.br >> [EMAIL PROTECTED] > > The patch completelly ignore the GROUP= on /etc/defaults/useradd, to be > truth, > almost the whole patch code is about ignore this and does not have any > options to change the behaviour, neither in /etc/login.defs. I use a version of Fedora, and started to reply, but forebore until now. I saw nothing in my /etc/defaults/useradd. The man page documents the "-n" option to prevent the creation of a user specific group. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Build RPM from LFS
satish patel wrote: > Hi > > I want to make my own installed like other distor but I want to do > this with RPMS so how to build RPM from install LFS system ?? Is there any > document on LFS which help to build RPMS from LFS source The way to build RPMs does not depend on what the source is. The RPM is independent of what it is building. You make a spec file which describes the build, and let RPM do the build for you. The output is two RPMs, one for source, one for built object. You need to study RPM and understand what it does. It simply executes the commands you would in order to do the build, but under automatic control, as contained in the spec file. Your question is really about RPM, and unrelated to LFS. I am no RPM expert, though I have built some RPMs from scratch. Perhaps I can give you some help in private e-mail. Or, perhaps there is an RPM expert here who would be better and willing to give you some help. Making a collection of RPMs with a proper set of dependency checks in them is not a simple task. I suggest you also study yum or some other install manager, like perhaps apt. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: gcc failure
Rodolfo Perez wrote: [...] > After all I don't know how to "unset any environmental > variables ..:-( According to the LFS Book, I have to do it only when I'm $ = $ export To find out what is set, $ set | less Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: I DID IT...
GMail wrote: > Finally... > > This monster that has been eating my time and patience is up. > > IPL'd my version of LFS at 20:00...woohoo. Initial Program Load! Congrats! > Now whats nextNASA? Just one step at a time, and you'll get there! Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Lots of Errors gcc make check
Ralph Porter wrote: > Thank you Ken, > > I did indeed have all the old source directories, although, I was > compiling in a new build dir. Seems strange that the presence of the > old source for other packages would effect the compile. The sources get modified during the configure step at least. Some of them get created. Usually, the Makefile gets created as well. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Question @ 10,000 feet
Alexander Haley wrote: [...] > What is to stop me from telling glibc to install itself into > /usr/weird/path/foo and gcc into /bar/zap/ .. and then somehow > configuring them to understand their relationship? Is that even > feasible? Would doing this somehow create a deeper understanding (for > me) of how gcc and glibc fit together? Nothing should prevent you from doing that, which I am aware of. [...] > Basically, the fundamental thing that bugs me is ... I type 'make > install' and scads of files arrive on the file system ... and I really > don't quite know their role, purpose or importance ... Do I really > need to know the purpose of each and every library file that is > installed? Probably not .. but, I am irked that I'm typing 'make > install' and just crossing my fingers that the system is getting it > right (of course the system often gets it right .. but does it > teach me? no. or at least, not yet.) What do you want to learn from it? Essentially, this particular question is just a matter of organization. There COULD be just one big library containing everything anyone could ever want in a Standad C Library, and the kitchen sink, as well. However, not many C programs use, for example, the arcsine function. So, the math related functions are collected together into a library of related functions. There is no compelling reason for doing that. It's for the convenience of the maintainers of the libraries. The fellow who knows how best to write a super fast strlen(.) may not be the one who knows best how to write a fast efficient and accurate asin(.) function, and vice versa. So, the library gets split up into "related" pieces. It makes it less unweildy. (Weird. It seems like I should be saying "It makes it more weildy", but that isn't the idiom. Anyway...) Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: A Suggestion For A Simple Package Manager
Frank Peters wrote: > Hello LFS users, > > The easy management of installed software packages is always > an important concern. After compiling, the "make install" command > does not help the user at all in knowing where the installed files > are located. The major Linux distributions have invented many types > of package management schemes but most of those are very complex. I've worked in software at various levels and in both small and large companies (one was just another guy and I, one was a large multinational telecomm company) for over 20 years. IME, version management is a complex problem, and no simple solution will do everything necessary, and of course complex solutions by definition have their own problems. I would argue that even the complex solutions don't do everything necessary. I'm not presently in the mood to start a package manager flame war. I'm sure each has its own adherents for various reasons. Each package manager is a compromise somewhere between simplicity, ease of use, and completeness. A good package manager is, IMO, a necessity. It's not a matter of just "binary distros", which I've come to conclude are not the right way to go. I've fought too many battles with distro management teams who would not listen to the needs of users. The single most important "feature" of LFS which attracts me is "my machine MY rules". Unfortunately, that almost precludes having a comprehensive package manager. If LFS team chooses to support a package manager, then that becomes the supported one, and then it's no longer "my machine, MY rules". So, the conundrum. LFS can't support every package manager out there, because it would just be too much effort. LFS can't just support one of them, because then it won't attain the "my machine, MY rules" goal. ISTM that this is a philosophical problem which must be addressed. One way of addressing it would be to abandon "my machine, MY rules" as an absolute principle, and support, say RPM, or APT, or whatever, turning LFS into a sort of GENTOO type source distro. Another would be for LFS simply to say (as it has) "we don't support package management, do it for yourself. The only 'package' we support is exactly the collection of software described in The Book. Read Book, Book Good!" Another would be for users to step forward and create the SPEC files (or whatever) for the entire distro, and manage it for themselves, with LFS simply hosting the product for ease of download. Another would be for individuals to install LFS and BLFS or whatever "by the book", and then build their own SPEC files or whatever for any additional apps they install, or request the owners of the apps to do so, and treat the basic LFS/BLFS as something which doesn't get maintained, but rather rebuilt when necessary. That also means, of course, that /etc/... has to be redone, CUPS reconfigured, etc. every time the system gets rebuilt, and one has to go through the laborious (and time consuming) rebuild (though "automated LFS" may help there somewhat). There may be another solution, but I don't see it. Until someone is willing to put in the effort to create the environment necessary for real package management (which is considerable) I don't think that yet another "here's a simple way to do package management" is going to be too helpful. Anyway, ISTM that before discussing Yet Another Simple Package Manager, the philosophy behind any package management has to be decided. So far, the LFS decision is "we don't do package management". If that's the final decision, then I don't think we need more package managers. We've already got too many, IMO. For those who really do want a package manager (and I consider one to be necessary) then step up and produce the necessary control environment for your favorite one, like RPM or APT. That, or switch to something like CentOS, grab the source, and build that, after making whatever changes you need. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Best Version to Build?
I started LFS 6.3, and got to the point where I was ready to make it bootable, but got interrupted (several months ago). I'm now ready to take the plunge again, and resume or restart building. So, I'm asking the development team which version is considered the best to use at the moment. Is 6.3 now considered "retired"? Is 6.4 considered sufficiently "stable"? I see that there is a 6.4 stable, and a 6.4 development. I've been using the "alongside" note, and building on a working running FC2 system. Thanks for the advice. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: A Suggestion For A Simple Package Manager
Frank Peters wrote: > On Tue, 17 Mar 2009 11:35:40 -0600 > Mike McCarty wrote: > >> ISTM that this is a philosophical problem which must be addressed. >> One way of addressing it would be to abandon "my machine, MY rules" >> as an absolute principle, and support, say RPM, or APT, or whatever, >> turning LFS into a sort of GENTOO type source distro. >> >> Another would be for LFS simply to say (as it has) "we don't support >> package management, do it for yourself. The only 'package' we support >> is exactly the collection of software described in The Book. Read >> Book, Book Good!" >> > > A package manager does not have to break the "my machine, my rules" > philosophy. I believe it would, actually. There are wars fought over which package manager "works right". Of course, no manager is capable of meeting the needs of everyone. I come from a background where RPM fits naturally with the kind of package management and version control I'm accustomed to. I am also aware that there are people who hate RPM. I'm not familiar enough with APT (though I've used it for install and maintenance of a Debian system, I've not built any packages) to know how it works, but I'm aware that there are APT people who wouldn't touch RPM with a ten foot pole. I've seen flame wars fought over whether another GUI (Synaptic?) on top of apt doesn't do violence to the purity of the tool. Such endless arguments constitute one of the reasons I'm no longer active on the support e-mail echoes for the distros I've maintained. > My idea for a package manager would fulfill two basic requirements: > > 1) Allow the user to build a package from source using a > pre-established set of CFLAGS, configure options, and other > variables. > > 2) Allow the user to know what is on his system and where > the files are located by creating a simple text database. That's certainly part of what is needed. > The dependency issue does not need to be addressed because > most followers of LFS will eventually learn enough about > their Linux system to know the dependencies for most > packages. The important thing is to have a quick and Ah, we disagree here. What I'm looking for is "my machine, MY rules", not "I want to become an expert on all aspects of Linux build and system maintenance". If I had a decent source distro which actually gave me the control I want over what features got built into the object, and which objects got built and installed, then I'd use it. I haven't found one. Becoming a Linux "fiddler" isn't a goal of mine. It seems that it's beginning to wear on you. Maintaining a distro is a lot of work. I'm impressed with the amount of labor which gets contributed, essentially free, by the maintainers of the "no charge" distros, like Fedora, Debian, Puppy, and (yes) LFS. [...] > A package manager -- or perhaps it should be called a build > manager -- is sorely needed. Perhaps not immediately, but > sooner or later the LFS follower will come to realize the > inherent pains of constant manual maintenance. Or perhaps not. :-) I would indeed call what you described a build manager, not a package manager. I also agree that one is needed. However, a build manager and a package manager, while related, and both necessary, seem distinctly different to me. A build manager would create packages using the package manager. Version control and dependency management are a part of package management which are both outside the purview of build management. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
LFS 6.4 Book Errata Errata
I see that a couple of files wouldn't retrieve, and found the "errata". I note that the errata section contains, well not exactly an erratum, but perhaps an omission. The replacement URL ftp://ftp.lfs-matrix.net/pub/lfs/lfs-packages/development/glibc-2.8-20080929.tar.bz2 is actually a symbolic link. If you use wget, then you'll need to use "--retr-symlinks" when loading that file, or point to the actual target ftp://ftp.lfs-matrix.net/pub/lfs/conglomeration/glibc/glibc-2.8-20080929.tar.bz2 Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Best Version to Build?
Ken Moffat wrote: Thanks very much for the reply. > On Tue, Mar 17, 2009 at 12:20:30PM -0600, Mike McCarty wrote: [...] >> So, I'm asking the development team which version is considered >> the best to use at the moment. Is 6.3 now considered "retired"? >> Is 6.4 considered sufficiently "stable"? I see that there is >> a 6.4 stable, and a 6.4 development. [...] > I've got one machine where I sometimes still run BLFS-6.3 : the > desktop packages are old, but apart from upgrading the gimp (for > functionality) and firefox (that's a box where I'm now using the > ubuntu version of firefox2 - see my BLFS-support post from last > week) it works (and is currently using 2.6.28 and later kernels). > > Unfortunately, BLFS still has a way to go before all the versions > are upgraded for 6.4. OTOH, if you describe FC2 as "working" you > probably aren't keen on using the "latest and greatest" versions ;-) Yes, one of the really off-putting things about FC is the churn. It wasn't the best distro for me, but I got a contract to do a Linux port of some SCO and some Windows stuff, and they wanted me to use FC, so... I got to experience the joy of the FC2 defect which wiped out the partition table. I needed this machine to dual boot Windows XP and FC2 to ensure I didn't break anything in the WXP version during the port. So, I got to learn about some dual boot issues with Compaq computers (they go into "recovery" mode if the MBR isn't as shipped, so I use the Windows Boot Manager in XP to boot GRUB off of another partition) and also about XP recovery along the way :-) > I expect everyone will discover things they don't like in their > first LFS/BLFS build, and therefore that system will have a > comparatively short life. I don't recommend that people use scripts > the first time they build LFS, but it's probably a good idea to I did the 6.3 (up to the point of making bootable) without the aid of any scripts. It took a few days. :-) I had to restart once, and clobbered my running system. Wiped out /dev when I rm'd the chrooted environment while it was still bound to the "real" /dev. I realized that I had still got the bound mount just a moment after pressing the return key :-( Recovered by booting Knoppix LiveCD, copying /dev to my hard drive, rebooting, and being up enough to find out how to rebuild /dev without reinstalling. Learned a lot more about Linux recovery than I ever wanted to know :-) > script the BLFS packages. Once you have the scripts, updating them > for a new build is comparatively simple (a few things move around, a > few get added, sometimes something can drop out). Are you saying that the BLFS for 6.3 will run fine with 6.4? I'd think so. Also, as you note, I really don't care about the L&G GUI stuff. I do my maintenance using a CLI, anyway. I might want a later GIMP and maybe EOG (mine can't print). > Therefore, I suggest you boot the LFS-6.3 system, upgrade the kernel > (2.6.27 will have longer-term support from upstream), and use it to > test out the packages you intend to use on it. When you decide it > has served its purpose, build a newer system. Using 6.4 and BLFS from 6.3? Or are you suggesting to wait for the 6.4 release of BLFS? > ĸen ^^ What is this stuff I see in some posts? Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: verify build files for LFS
support wrote: > > I don't quite get what you are after, if you are downloading a file from > the authors site, and you also want an md5 from the same site to confirm > the download, its kind of pointless. If the site has been hacked and > the original source replaced with something else, it stands to reason > that the md5 (on the same site) would be compromised also. The download > will reach you in original condition thanks to the fact that tcp/ip does > error checking as it goes. As for instructions on confirming the md5 on Umm, TCP/IP does do error checking. That does not guarantee an error free download. The error checking used is rather weak. I forget the name of the checksum used, but essentially it's sum_of_wordss mod(65535) and sum_of_sums_of_words mod(65535). It's named for the guy who suggested it, and IIRC his name starts with an "F". Anyway, it's very weak, and was intended to be "easy" to implement. Today, that header is considered by all to be a mess, and very difficult on modern machines (i.e. with word sizes > 8 bits) to implement efficiently due to odd byte addressing required, resulting in often needing to copy the entire message in order to prepend some stuff. I've experienced corrupted files downloaded a few times. > each page, ideally you should do it before you begin at all, its > pointless getting halfway through the project just to find you have a > bad copy of something. It would be nice to have a single file which listed all the md5 (prefereably sha1) sums for all the files which could be used as input for automated test. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: verify build files for LFS
Chris Staub wrote: > Mike McCarty wrote: >> It would be nice to have a single file which listed all the md5 >> (prefereably sha1) sums for all the files which could be used as >> input for automated test. >> >> Mike > > Uh, you mean like this? - > ftp://ftp.lfs-matrix.net/pub/lfs/lfs-packages/development/MD5SUMS > > Every dir in the LFS download ftp has an MD5SUMS file listing md5sums > for all files in that dir. http://www.linuxfromscratch.org/lfs/downloads/stable/ Has no such file, and this is the natural place to put it, don't you think? However, I did find ftp://ftp.lfs-matrix.net/pub/lfs/lfs-packages/6.4/SHA1SUMS which looks like it might be the place to look. The Book might at least point there. Thanks for the pointer. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: verify build files for LFS
Mike McCarty wrote: > support wrote: >> I don't quite get what you are after, if you are downloading a file from >> the authors site, and you also want an md5 from the same site to confirm >> the download, its kind of pointless. If the site has been hacked and >> the original source replaced with something else, it stands to reason >> that the md5 (on the same site) would be compromised also. The download >> will reach you in original condition thanks to the fact that tcp/ip does >> error checking as it goes. As for instructions on confirming the md5 on > > Umm, TCP/IP does do error checking. That does not guarantee an error > free download. The error checking used is rather weak. I forget the > name of the checksum used, but essentially it's sum_of_wordss mod(65535) > and sum_of_sums_of_words mod(65535). It's named for the guy who > suggested it, and IIRC his name starts with an "F". Anyway, it's very I recalled it! "Fletcher's Checksum" Sheesh! It's pretty much universally hated by everyone who uses it. I just did this: $ sha1sum -c ../../../6.4/SHA1SUMS 2>&1 | grep -v OK inetutils-1.5.tar.gz: FAILED sha1sum: WARNING: 1 of 85 computed checksums did NOT match Hmm. OTOH: $ gunzip -tv inetutils-1.5.tar.gz inetutils-1.5.tar.gz:OK So, which file is in error, the tar.gz or the SHA1SUMS file? Gonna have to look into this. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: verify build files for LFS
Mike McCarty wrote: > Mike McCarty wrote: >> support wrote: >>> I don't quite get what you are after, if you are downloading a file from >>> the authors site, and you also want an md5 from the same site to confirm >>> the download, its kind of pointless. If the site has been hacked and >>> the original source replaced with something else, it stands to reason >>> that the md5 (on the same site) would be compromised also. The download >>> will reach you in original condition thanks to the fact that tcp/ip does >>> error checking as it goes. As for instructions on confirming the md5 on >> Umm, TCP/IP does do error checking. That does not guarantee an error >> free download. The error checking used is rather weak. I forget the >> name of the checksum used, but essentially it's sum_of_wordss mod(65535) >> and sum_of_sums_of_words mod(65535). It's named for the guy who >> suggested it, and IIRC his name starts with an "F". Anyway, it's very > > I recalled it! "Fletcher's Checksum" Sheesh! It's pretty much > universally hated by everyone who uses it. > > I just did this: > > $ sha1sum -c ../../../6.4/SHA1SUMS 2>&1 | grep -v OK > inetutils-1.5.tar.gz: FAILED > sha1sum: WARNING: 1 of 85 computed checksums did NOT match > > Hmm. OTOH: > > $ gunzip -tv inetutils-1.5.tar.gz > inetutils-1.5.tar.gz:OK > > So, which file is in error, the tar.gz or the SHA1SUMS file? > > Gonna have to look into this. Ok, gunzip thinks there is no error, but when I pulled that file from ftp://ftp.lfs-matrix.net/pub/lfs/lfs-packages/6.4/inetutils-1.5.tar.gz it's different from the wget-list file's URL target http://ftp.gnu.org/gnu/inetutils/inetutils-1.5.tar.gz So, which one is correct? the one from the GNU project page, which does not match the SHA1SUMS file, or the one from the LFS page, which does? Presumably, any difference between what GNU provides and what LFS actually needs would be provided by a corresponding patch file. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Best Version to Build?
genericmailli...@gmail.com wrote: > On Wednesday 18 March 2009 12:07:43 Mike McCarty wrote: >> I got to learn about some dual boot >> issues with Compaq computers (they go into "recovery" mode if the >> MBR isn't as shipped, so I use the Windows Boot Manager in XP to >> boot GRUB off of another partition) and also about XP recovery >> along the way > > Do you still have any of the info about the Compaq recovery/MBR > issue? I would really like to get that info have you have it. Yes, I do. Is this sufficiently LFS related to continue to discuss it here, or should we move to private e-mail? It might be remotely LFS related, as in "How to make LFS play with WinXP on a multiboot system which won't permit a Linux loader in the MBR." Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: GCC-4.3.2 - Pass 2 - expect -c "spawn ls"
SKOC(DOPOLE Tomá wrote: > Hello everybody, > > I have booted LFS live CD and now I am in the chapter 5.12. GCC-4.3.2 - Pass > 2. > I have some question about this command: > lfs:/mnt/lfs/sources/gcc-4.3.2$ expect -c "spawn ls" > > I got this response: > spawn ls > but in the book is written this response: > The system has no more ptys. > Ask your system administrator to create more. You got some possibly useful answers. I thought I'd explain the meaning of the error. A pty is a "pseudo terminal" used for connecting the standard input and output files when a program is run by another program which wants to trick the program running into thinking it is connected to a terminal, but control the input and capture the output. In this case, "expect" is running "spawn" which runs "ls", and your system has not enough (possibly none) pseudo terminals to make this work. > I want to ask you about this difference outputs. > > And my second question is about possibility interrupt compiling and after > some time to continuing. > I have ssh daemon started in target machine and I am compiling my LFS > remotely. > I tried to start screen, but i give this error: > $ screen > Cannot open your terminal '/dev/pts/1' - please check. This looks like /dev/pts is an ordinary directory instead of a collection of devices mapped to your terminal, as pointed out in another message. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: A Suggestion For A Simple Package Manager
Frank Peters wrote: > On Wed, 18 Mar 2009 14:46:08 -0500 > Tushar Teredesai wrote: > >> Because that goes against the Unix developer mentality. Providing an >> "checkinstall" style functionality is not the purpose of make. >> According to the man page: "The purpose of the make utility is to >> determine automatically which pieces of a large program need to be >> recompiled, and issue the commands to recompile them." >> > > Well, the make program *does* indicate where each file has been > installed. If you examine the console output of make, all Often Makefiles don't do the install themselves, but rather either us "cp" or "install" to put things where they go, when one uses "make install". I don't like the idea of making "make" have purposeful output to track versions and installs. ISTM that one should use a special tool, like "install" for that. "make" is supposed to be a tool which other tools can use, not a Swiss Army Knife. Having "install" track what has been put where sounds like a reasonable idea, and if it provides specially tagged output upon request, that could be captured and used, perhaps. I'm not an "install" expert. Maybe it already does that. > the install information is there but it is mixed in with all > sorts of other messages. It would be possible to redirect That's not clear to me. > the make output to a text file and then edit the file to remove > all the extraneous messages and thereby produce a usable install > log. > > A very useful addition, IMO, to the make program would be > to automatically redirect just the install information to a file > as an option. This should not be too difficult to accomplish > but it may have to done along with the other build utilities > such as autoconf, automake, etc. > > Frank Peters > -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: A Suggestion For A Simple Package Manager
Chris Staub wrote: [...] > Also, as far as I know, there's really nothing special about > "install"...it's just another Makefile target, though obviously one of > the commonly-used ones. As a target, yes, however there is an "install" program, and that's what I was mentioning. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: A Suggestion For A Simple Package Manager
DJ Lucas wrote: [...] > Unfortunately, you'd need to do the same with cp, ln, mkdir, mknod, mv, > rename, rm, etc. Suddenly, the approach by installwatch, CheckInstall, > and other like approaches, makes quite a bit more sense. I don't recall saying that those approaches didn't make sense. I also didn't recommend modifying "install". I said it makes more sense to modify "install" than it does to modify "make". Modifying either tool doesn't really appeal to me. > Additionally, there is a reason that the vast majority of makefiles > today support DESTDIR. The following is what I (and others) have > proposed several times for the book. We simply install to the DESTDIR I've used similar techniques in order to find what to put into the %files section of RPM spec files. [...] > In this manor, we've managed to explain the complexities of PM, give > real examples, and still avoid using any particular PM by default. > Dependencies are already covered by the book, the only thing left is > upgrading. Concerning upgrades, Perl (specifically perldoc) is the only Version control, dependency management, and package management are not simple, and doing them by hand is prone to mistakes, and vastly unappealing to me. I've been doing it for over twenty years with a variety of tools, some expensive, some not, and I have yet to find one which does everything one needs in a simple and easy to use manner. I think this thread is likely reaching the end of its useful life. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: verify build files for LFS
Chris Staub wrote: > Mike McCarty wrote: >> Ok, gunzip thinks there is no error, but when I pulled that >> file from >> >> ftp://ftp.lfs-matrix.net/pub/lfs/lfs-packages/6.4/inetutils-1.5.tar.gz >> >> it's different from the wget-list file's URL target >> >> http://ftp.gnu.org/gnu/inetutils/inetutils-1.5.tar.gz >> >> So, which one is correct? the one from the GNU project page, which >> does not match the SHA1SUMS file, or the one from the LFS page, >> which does? Presumably, any difference between what GNU provides >> and what LFS actually needs would be provided by a corresponding >> patch file. >> >> Mike > > I just downloaded the inetutils 1.5 tarball from the LFS ftp, and from > gnu, and both are identical. Also, the SHA1SUM given in the SHA1SUMS > file in the 6.4 dir on the LFS download site matches what I get for both > ineutils tarballs... That's odd. I got different files, and gunzip didn't complain that either was corrupt. That's confusing. [...] > Why it's saying "FAILED" for you, I don't know. It's because the files have different content. I just re pulled from both sites, and $ ls -l gnu-site/inetutils-1.5.tar.gz lfs-site/inetutils-1.5.tar.gz -rw-rw-r-- 1 jmccarty jmccarty 1390529 Jun 29 2007 gnu-site/inetutils-1.5.tar.gz -rw-rw-r-- 1 jmccarty jmccarty 1390529 Jan 4 22:20 lfs-site/inetutils-1.5.tar.gz $ diff -s gnu-site/inetutils-1.5.tar.gz lfs-site/inetutils-1.5.tar.gz Files gnu-site/inetutils-1.5.tar.gz and lfs-site/inetutils-1.5.tar.gz are identical Using the link you helpfully provided: $ wget -N -c -nv ftp://ftp.lfs-matrix.net/pub/lfs/lfs-packages/6.4/SHA1SUMS 02:54:59 URL: ftp://ftp.lfs-matrix.net/pub/lfs/lfs-packages/6.4/SHA1SUMS [11,704] -> ".listing" [1] 02:55:00 URL: ftp://ftp.lfs-matrix.net/pub/lfs/lfs-packages/6.4/SHA1SUMS [5,653] -> "SHA1SUMS" [1] $ grep inetutils-1.5.tar.gz SHA1SUMS 825834b94cd387b2d088ef1cfe727de824b9589e inetutils-1.5.tar.gz $ sha1sum lfs-site/inetutils-1.5.tar.gz 825834b94cd387b2d088ef1cfe727de824b9589e lfs-site/inetutils-1.5.tar.gz $ sha1sum gnu-site/inetutils-1.5.tar.gz 825834b94cd387b2d088ef1cfe727de824b9589e gnu-site/inetutils-1.5.tar.gz Perhaps the SHA1SUMS file is incorrect? Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Backup of LFS
Ryan Isaacs wrote: > I have a fresh LFS 6.4 installation, all I've done is add 1 user. Is > there a preferred way to back it up (and restore)? > > Would it be sufficient to just use the liveCD, mount my paritition, > and 'cp -ar', or do I need to do something more arcane like 'dd'? It depends upon what your goals are. Do you want a quick'n'dirty way to put the disc back the way it was? If so, then the "dd" mentioned in another post is one easy way to do that. If you are asking more about generally how to back up Linux, then there are several systems. Many like Amanda or Bacula. I've used tar with various hand generated scripts. Lately, I've changed over to using yakup, which seems to work reasonably well, allows incrementals, checks the results. One thing I don't like about it is that it doesn't put a TOC on the backup discs themselves, only in the backup repository. Another thing I don't like about it is that it stores the entire directory structure, even if no files have changed. So, I get megabyte sized TOCs which are essentially just a list of my entire directory structure, with a few hundred bytes of file names. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: How to upgrade a package in a linux distribution?
petrele.chen wrote: [...] > And about LFS, I doesn't agree with you in some place. First, LFS is > a way to build your own OS _or_ learn the construction of a Linux OS, > and what you end up it with is not important at all. So, where the > question came from doesn't matter at all. the only thing matters is the > solution. You can't expect the LFS support group to be experts in system administration for every distro of Linux on the planet. If your question is about LFS, then you can expect expert responses. If it's about some other distro, then you need to ask people who reasonably can be expected to be experts in that distro. That usually means asking in a support group for that distro. There are certain requirements listed for the system used to build LFS. If you try to use one which does not meet those minimal requirements, then you are more or less on your own, I'm afraid. The LFS support team supplies a system which does meet the minimal requirements, namely the LFS LiveCD. THAT they can be expected to support and help with. Perhaps that's the way you should go. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: verify build files for LFS
Mike McCarty wrote: [...] > That's odd. I got different files, and gunzip didn't complain > that either was corrupt. That's confusing. Or so I thought. Now, I'm not so sure. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: USB Hard Drive issues
Jack Stone wrote: > Randy McMurchy wrote: >> The 'mount -a' command will only mount those filesystems >> identified in /etc/fstab. And typically you would not put >> a USB device in /etc/fstab. Have you identified the USB >> drive in that file? If so, how do you ensure that the drive >> always has the same device filename every time? >> > Sorry I should have made that clear. Yes the device does have an entry > in fstab. So far it is always given the same name (sde), but I don't > know if I can rely on that. You cannot. It depends on the order in which the USB gets enumerated, and what gets found. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: A Suggestion For A Simple Package Manager
Tobias Gasser wrote: >> A good package manager is, IMO, a necessity. > NO > > > i'll explain why i NEVER will use any package-manager: The choices you make are of course your own. I would not presume to tell you what you should do for your own machine. I stated what I consider to be necessary for my machines, and my machines only. That's why I put the qualifier "IMO" into that statement. I'm a little surprised at the vehemence of your statement. Why are you so excited by my opinion which applies only to the machines I own that you used four exclamation marks? > have a look at the configure options with php5: [...] > which version do you declare the one and only to be used for the > package-management?? I don't declare any particular version to be the one and only to be used for package management. I have used both of the terms version control (which is what you allude to here) package management (which is related, but an independent concept) You seem to be conflating them. > with (b)lfs i have full control on what is installed on a system. i With a proper version control manager, package manager, and build control and integration system, you would have complete control over every full system you built. > really dislike packages like "php5" + "php-mysql" + "php-pgsql" + > "php-whatever-you-wish" (and each one again with -devel). i just compile > php with all the options i need on a particular machine. ISTM that you have relatively little contact with reasonable version control, package management, and build control systems. It shouldn't even be necessary to compile or build the image(s) necessary for a given machine on a machine of the same architecture, let alone for it to be the same machine. Most of the work I've done with such types of build systems was compiling for an OS (embedded proprietary RTOS) and machine architecture (microcontroller) which were both different from those on the development machine (often Solaris on a Sparc). In fact, often multiple targets with different machine architectures were often specified, and the software required to run properly on all of them, and multiple target builds might be required for a single release of a subsytem (package). What is necessary is a full and correct specification of what a given target build requires, and some tools which know how to interpret that specification and create the necessary packages, and integrate them into a deliverable product which is guaranteed to meet the specifications. Once the requirements for a given build have been specified, the build should not require further intervention, other than to command it to be done.[*] The result of the build should be all components necessary for the target to boot and run. One of the "packages" may be an installer which will update a target to the latest release level. With a good front end, one can use some relatively easy to use tools to generate the specification of the target. Of course, creation of such a collection of tools is a significant effort, as is maintenance of it, as well as the maintenance of the specifications, though those should be relatively static by comparison with the point releases of the components comprising the deliverable. [...] > greetings from switzerland Greetings from the U.S.A. [*] Of course, test is still required to verify that the functional and performance requirements are actually met, on the target, that is, to verify proper operation. I mean that the /build/ is guaranteed to meet its specifications, not that the built image meets is requirements. That's another matter altogether. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: verify build files for LFS
Mike McCarty wrote: [...] > $ grep inetutils-1.5.tar.gz SHA1SUMS > 825834b94cd387b2d088ef1cfe727de824b9589e inetutils-1.5.tar.gz > > > $ sha1sum lfs-site/inetutils-1.5.tar.gz > 825834b94cd387b2d088ef1cfe727de824b9589e lfs-site/inetutils-1.5.tar.gz > $ sha1sum gnu-site/inetutils-1.5.tar.gz > 825834b94cd387b2d088ef1cfe727de824b9589e gnu-site/inetutils-1.5.tar.gz > > Perhaps the SHA1SUMS file is incorrect? No one cared to comment? Perhaps I didn't state it quite as strongly as I should. The SHA1SUMS file is incorrect. Using the URL you provided... $ mkdir development $ cd development $ wget -N -c -nv ftp://ftp.lfs-matrix.net/pub/lfs/lfs-packages/development/SHA1SUMS 09:00:21 URL: ftp://ftp.lfs-matrix.net/pub/lfs/lfs-packages/development/SHA1SUMS [11,512] -> ".listing" [1] 09:00:22 URL: ftp://ftp.lfs-matrix.net/pub/lfs/lfs-packages/development/SHA1SUMS [5,543] -> "SHA1SUMS" [1] $ grep inetutils-1.5.tar.gz SHA1SUMS f2dc4d7fa5f5590ea71260c464f60eb3b1fc7505 ./inetutils-1.5.tar.gz However, using the stable URL $ cd .. $ mkdir stable $ cd stable $ wget -N -c -nv ftp://ftp.lfs-matrix.net/pub/lfs/lfs-packages/stable/SHA1SUMS 09:02:57 URL: ftp://ftp.lfs-matrix.net/pub/lfs/lfs-packages/stable/SHA1SUMS [11,704] -> ".listing" [1] 09:02:57 URL: ftp://ftp.lfs-matrix.net/pub/lfs/lfs-packages/stable/SHA1SUMS [5,653] -> "SHA1SUMS" [1] $ grep inetutils-1.5.tar.gz SHA1SUMS 825834b94cd387b2d088ef1cfe727de824b9589e inetutils-1.5.tar.gz They can't both be correct. The one in "stable" seems to match the actual source file. The one in "development" seems wrong. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: chroot problem
rajesh britto wrote: > hi, > i entered into the chroot environment and then i tried to install the > packages. but it shows the error: > > root:~/linux-2.6.27.4# pwd > /root/linux-2.6.27.4 > root:~/linux-2.6.27.4# make mrproper > make: gcc: Command not found > root:~/linux-2.6.27.4# > > dont know how to proceed. help me in this issue. > > Note: I am in the chroot environment not in the systems root directory. What is your PATH variable? $ set | grep PATH What happens when you look down the pieces of the PATH variable? Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: A Suggestion For A Simple Package Manager
Tobias Gasser wrote: [...] > for me, a package is something wich can be used to install something A package is anything which gets "packaged" for delivery. Usually a package incorporates more than one item in it. > where as the package manager helps resolving all (or at least most) > dependencies. the manager has options to remove or update a package too, > maybe even an option to find unneeded packages. This is a reasonable working definition from the viewpoint of the user, it does not describe a package manager from the perspective of the developer. > this definition (as i was convinced to be correct) implies the > compile-time options for any package to be set by the package builder. No. A package does not necessarily contain binaries. A package may contain sources. As an example, a tarball containing configure, a Makefile, etc. comprises a package, with most or all package management done manually. In another vein, a package does not necessarily even result in programs. One can have a documentation package, for example, which only contains documents. Red Hat uses a package to ship the /dev directory. A package manager is a collection of programs which helps both the maintainer and user of the package ensure that the package contains the proper versions of installable objects, helps track what objects are installed and their versions, and track dependencies between them, and uninstalls and tracks dependencies when they are uninstalled. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: I managed to wipe out my host /dev/ directory
Dan Tran wrote: > Hello, > > I am a first time user of LFS, and was away from linux related wor for > quite some time. > > I am now at chroot phase where I managed to wipe out my $LFS/dev > directory which was still bound to the actual /dev Well, great minds get stuck in the same ruts :-) I did the same thing, when I decided I needed to restart the build from the beginning, and rm'd the build environment. I had exited the chroot, but forgotten to unmount the bound mount. So, there ya go. I realized what I had done the moment I lifted my hand from the keyboard. Sinking feeling in the pit of my stomache. > My system is still functional, i just can not login ::( Also, don't shut down. Find out what RPM supplies /dev on your system. First, find the version. I'll show you the output for my machine, and that may help. Be sure to substitute the appropriate versions and file names when you run yours, as they are very ulikely to be the same. $ rpm -qf /dev dev-3.3.13-1 That tells you the name of the RPM which supplies /dev. In my case it's dev-3.3.13-1.i386.rpm Next, you need to find a copy of that RPM. $ locate dev-3.3.13-1 /home/jmccarty/packages/rpm/dev-3.3.13-1.i386.rpm I'll suppose that you don't have a copy of that on your machine, as I didn't. I found mine on rpm.pbone.net and got it from ftp://ftp.muug.mb.ca/mirror/fedoralegacy/fedora/2/os/i386/dev-3.3.13-1.i386.rpm You'll need to find yours and download it, then install. # rpm -i --force dev-3.3.13-1.i386.rpm Note carefully the "#". That must be done as root. At this point, if all is well, your system is in a reasonable state, but it hasn't built the entries. You can try to do MKNODs yourself, or you can take a (very) deep breath and reboot. I'd make sure I had some sort of rescue disc, and another machine I could boot to continue to get help, or at least be able to get help when booted from the rescue disc before I did anything to the machine other than download the RPM. I had already rebooted and discovered that the system was, indeed, hosed, and completely unable to boot. I used KNOPPIX and simply copied the (working) /dev to my machine to get it to a mostly bootable state, but unable to mount USB devices, the printer was gone, etc. I used that hobbled state to download and install. System recovery is always scary. I hope you have a good backup. If you can, you might want to make one now, before doing anything else. Good luck. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: I managed to wipe out my host /dev/ directory
Dan Tran wrote: > Hi Mike, > > what is your original distribution? RedHat ES5? Fedora Core 2 > I am able do some primative shell commands thru myweb console ( lucky i guess) > > + rpm -qf /dev > filesystem-2.4.0-1 > + uname -a > Linux fortidb 2.6.18-53.el5PAE #1 SMP Wed Oct 10 16:48:18 EDT 2007 > i686 i686 i386 GNU/Linux > finished: SUCCESS Try doing the same thing on another one you know will give you the rest of the name for the RPM. Your RPM is going to be named something like filesystem-2.4.0-1.i386.rpm or filesystem-2.4.0-1.i686.rpm etc. So, look for something which will tell you the whole name. My system has /etc supplied by filesystem-2.2.4-1. > Thanks for for advice I hope it helps. I'll poke around and see what I can find out about RH 5.0 and names for RPMs and maybe a URL for you. Perhaps we should take this off the LFS list, since it's getting pretty far OT by now. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
JHALFS
I am going to try a build of 6.4 using QEMU to run a virtual machine booting the LFS 6.3 LiveCD. I've got it set up so that QEMU sees the ISO image as the CD-ROM, and sees another actual CD-ROM as a VFAT RO file system, containing the documentation directory of the book for 6.4, and another disc file as /dev/hdb. I've successfully booted, and verified that I can access the book, and run fdisk to partition and install an ext3 file system on /dev/hdb1. I now want to get started getting JHALFS up and running, but it seems to want a path to BOOK.XML, which doesn't appear to exist for the 6.4 version of the book. What should I configure JHALFS to use? have a directory with this view (from the host system) $ ls /mnt/cdrom/6.4/extracted/BOOK appendicesINSTALL aux-file-data.sh lfs-bootscripts-20081031.tar.bz2 bootscripts make-aux-files.sh chapter01 Makefile chapter02 obfuscate.sh chapter03 packages.ent chapter04 patches.ent chapter05 process-scripts.sh chapter06 prologue chapter07 README chapter08 stylesheets chapter09 tidy.conf general.ent udev-config imagesudev-config-20081015.tar.bz2 index.xml Do I need to pass JHALFS "index.xml" as the XML version of the book? If not, then what, exactly, is JHALFS expecting as input? Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
UML Support Anyone?
I am attempting to use the UML hint, and have successfully booted linux, but am having troubles logging in as any user. I'm getting PAM failures. I've tried sending e-mail to the originator of the hint, but the e-mail address in the hint is no longer valid. Is there anyone who supports this hint, or is willing to try to help me overcome the snag? Thanks! Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: JHALFS
Thomas Pegg wrote: > Mike McCarty wrote: [...] >> an ext3 file system on /dev/hdb1. I now want to get started >> getting JHALFS up and running, but it seems to want a path >> to BOOK.XML, which doesn't appear to exist for the 6.4 >> version of the book. What should I configure JHALFS to use? >> have a directory with this view (from the host system) > > Not sure why it would be looking for BOOK.XML, it should look for index.xml Ok, that makes sense. I'm not getting an error, I'm running the configurator and it's prompting for the location of the book. The default entry for edit has /usr/share/LFS-BOOK-6.3.XML in it. I need to know what to enter there. I suppose that just the path to the index.xml file will do. Does that need to be in a writable directory? > Could you post the exact error your getting? > > What version of jhalfs are you using, run jhalfs -v to find out. That should be 2.3.1, however, I'm in the process of simply doing a "make", so there is not yet any script to run. It's the one on the LiveCD 6.3. What you gave me may be enough to go on. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: JHALFS
Thomas Pegg wrote: Thanks for your reply! > Not sure why it would be looking for BOOK.XML, it should look for index.xml > > Could you post the exact error your getting? Right after boot, I did # su - jhalfs $ startx and since then I've been using sudo for all root level commands necessary, otherwise all work done as jhalfs. Ok, I've copied to the mounted disc image to get a writable version and done a "make". It runs a configurator which has menu driven entry. I entered the paths I want to use, and told it to save the configuration. It then begins to verify that the toolset is adequate (it is) and then I get this: [QUOTE MODE ON] The variable "" value is invalid, rerun make and fix your configuration settings. [QUOTE MODE OFF] When I do an ls on that full path it shows as existing, and having permissions 755 rwxr-xr-x, but being in a ro file system. > What version of jhalfs are you using, run jhalfs -v to find out. Not possible, it doesn't exist, but the directory it's in is /home/jhalfs/jhalfs-2.3.1 (originally, now copied). Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: JHALFS
Thomas Pegg wrote: > On Thu, Mar 26, 2009 at 10:39 PM, Mike McCarty > wrote: >> [QUOTE MODE ON] >> The variable "" value >> is invalid, rerun >> make and fix your configuration settings. >> [QUOTE MODE OFF] > > Ok retry it like this "/media/hdb1/6.4/extracted/BOOK", but without > the quotes I put in. > You don't want index.xml in there, jhalfs adds that automatically. > Now if that doesn't work, then I'm kinda stumped. It likes it, and it's starting to extract commands. Nothing executed yet, but it looks good so far. >> When I do an ls on that full path it shows as existing, >> and having permissions 755 rwxr-xr-x, but being in a >> ro file system. >> >>> What version of jhalfs are you using, run jhalfs -v to find out. >> Not possible, it doesn't exist, but the directory it's in >> is /home/jhalfs/jhalfs-2.3.1 (originally, now copied). > > It should be possible within that directory should be be a shell > script named jhalfs. > ./jhalfs -v [QUOTE MODE ON] $ ./jhalfs -v "jhalfs 2.3.1" builder tool (stable) $Rev: 3468 $ $Date: 2007-08-30 10:57:38 -0600 (Thu, 30 Aug 2007) $ Written by George Boudreau and Manuel Canales Esparcia, plus several contributions. Based on an idea from Jeremy Huntwork This set of files[sic] are published under the Gnu General Public License, Version 2. exit [QUOTE MODE OFF] The "[sic]" is because of the grammar error. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: JHALFS
Thomas Pegg wrote: > > Ok retry it like this "/media/hdb1/6.4/extracted/BOOK", but without > the quotes I put in. > You don't want index.xml in there, jhalfs adds that automatically. > Now if that doesn't work, then I'm kinda stumped. Ok, it processed the book, seemingly, but it still can't build. It apparently has no concept that the source files might not actually be in the same place as the build directory. I'd like to leave them on the CD-ROM, and not copy everything to the miniature disc. It is going to copy from the sources directory to a temp directory anyway, so I don't see why they are conflated. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
A Modest Proposal for JHALFS Enhancement
In the vein of being able to build small systems, I hereby make this modest proposal for an enhancement to JHALFS. Currently, the system builds in a subdirectory of the configurable sources directory, which must, therefore, be on read/write storage. I suggest that the actual directory in which builds take place also be configurable, and not be required to be a sub directory of the sources directory. This would have two desirable consequences. First, the sources could reside on an immutable medium, like a CD-ROM, and not consume precious (on small systems) disc space. Second, the sources would be immutable, and hence not subject to accidental modification. In this wise, the disc space used during the builds would be used only when a build was actually taking place, and could be reclaimed after each build (at least when temporary builds are taking place). Since LFS is essentially a source distribution (if one may even call it a distribution), there is no need to keep the objects in the build directories to be preserved, but only the installed objects. The original compressed source tarballs could continue to reside on an immutable medium. I trow that the additional effort would be well worth it, and the extra maintenance effort would likewise be well repaid. Thank you for your consideration. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: A Modest Proposal for JHALFS Enhancement
Bruce Dubbs wrote: > Mike McCarty wrote: [...] >> This would have two desirable consequences. First, the sources >> could reside on an immutable medium, like a CD-ROM, > > Nothing prevents you keeping a copy on CD-ROM now. What I mean is it would not take up disc space. I have a system which I'd like to build on, which has a small hard disc. > and not >> consume precious (on small systems) disc space. Second, the >> sources would be immutable, and hence not subject to accidental >> modification. > > This doesn't make sense to me. If you want to put it on a small system, > build > it on a workstation and copy it to the small system. It sounds like a > painful > process to build on a small system, especially when there is no need. After > all, if you do this, you don't need to copy *any* sources to the small system. Your observation makes a lot of sense for people who have the resources to do what you propose. However, I'm a laid off telecomm engineer putting together a system out of pieces of left over stuff I have lying around. Since I don't have anyone who has stepped up and volunteered to supply a system with more resources, I need to use what I have, which is cramped for disc space. It would help if the sources didn't have to reside on the same medium with the built images. > There are other considerations. The build process takes a lot of room. > glibc > takes 1.2G, but the final system is only about 250M and a *lot* less if you > strip it down. > > That said, I'm sure the JHALFS team will look closely at any patches you want > to > submit. I'll see what I can do. It's not clear to me how the tool sets up the directory to do the decompresion into, but I'll investigate it. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
JHALFS with QEMU 6.3 LiveCD Fails
I've created a small (I now realize too small, but that's not the problem) QEMU disc and am running the 6.3 r2160 build of the LiveCD as QEMU's CD-ROM. I configured and started JHALFS, which ran for a while, then gets an unrecoverable error. I'm trying to build LFS 6.4 from the 6.4 book. Here's how I have the disc partitioned: jhalfs [ /media/hda1/build/jhalfs ]$ sudo fdisk -l /dev/hda Disk /dev/hda: 2516 MB, 2516582400 bytes 255 heads, 63 sectors/track, 305 cylinders Units = cylinders f 16065 * 512 = 8225280 bytes Device Boot Start End Blocks Id System /dev/hda1 *1 243 1951866 83 Linux /dev/hda2 244 305 498015 82 Linux swap / Solaris I've mounted /dev/hda1 on /media/hda1, and enabled swap on the other partition. I've created /media/hda1/build, and put the sources in the right place, and made a copy of the jhalfs-2.3.1 directory using cp -a so I have a writable copy. I did a make, and it seemingly successfully created Makefile. I did another make, and it successfully ran up through the creation of gcc. However, it failed attempting to make the API headers. I snipped off the relevant message (somewhat edited for readability): Building target 031-linux-headers [ + +++- make: *** [031-linux-headers] Error 1 make: *** [mk_LUSER] Error 2 I'm not out of emulated disc, as df shows /dev/hda1 is only 40% full. Reading the man page for make didn't help much. However, poking around found jhalfs/logs/031-linux-headers Fri Mar 27 20:37:33 CDT 2009 KB: 348280 /media/hda1/build make[1]: Entering directory `/media/hda1/build/sources/linux-2.6.27.4' make[1]: Leaving directory `/media/hda1/build/sources/linux-2.6.27.4' cp: cannot stat `../kernel-config': No such file or directory So, it looks like a required file is not present. I don't recall whether I configured it to build the kernel, but this does look like a possible defect in the tool. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: JHALFS with QEMU 6.3 LiveCD Fails
Bruce Dubbs wrote: > Mike McCarty wrote: > >> make[1]: Entering directory `/media/hda1/build/sources/linux-2.6.27.4' >> make[1]: Leaving directory `/media/hda1/build/sources/linux-2.6.27.4' >> cp: cannot stat `../kernel-config': No such file or directory >> >> So, it looks like a required file is not present. I don't recall whether >> I configured it to build the kernel, but this does look like a possible >> defect in the tool. > > You have to create kernel-config yourself. There is no way to know in > advance > what hardware you have. Ok, I'll go look in the book and read the relevant section. Is there any way to restart the build, short of starting over from the beginning? Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: JHALFS with QEMU 6.3 LiveCD Fails
Bruce Dubbs wrote: > Mike McCarty wrote: > >> make[1]: Entering directory `/media/hda1/build/sources/linux-2.6.27.4' >> make[1]: Leaving directory `/media/hda1/build/sources/linux-2.6.27.4' >> cp: cannot stat `../kernel-config': No such file or directory >> >> So, it looks like a required file is not present. I don't recall whether >> I configured it to build the kernel, but this does look like a possible >> defect in the tool. > > You have to create kernel-config yourself. There is no way to know in > advance > what hardware you have. Ok, I'm still confused. This is the Chapter 5 build, and I see nothing in there about kernel-config. If the commands are simply extracted from the book, then I don't see where the book describes how to construct that file. I also read the Chapter 6 build instructions for the Linux API, and see nothing about how to create that file. So, I still don't know what to do about this. The book does not describe building that file in either Chapter 5 or Chapter 6. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: JHALFS with QEMU 6.3 LiveCD Fails
Tony Sauri wrote: > On Tue, 31 Mar 2009 10:22, Mike McCarty wrote: > >>> Mike McCarty wrote: > >> So, I still don't know what to do about this. The book does not >> describe building that file in either Chapter 5 or Chapter 6. >> >> Mike > > The point Bruce refers to is covered in the jhalfs 2.3.1 README file. Thanks very muchly! I'll go read that (again). I missed it somehow. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: GCC-4.3.2 Pass1 - requires GMP 4.1+ and MPFR 2.3.0+
Thomas Anderson wrote: > Hi, > I was working on the compilation of GCC (first pass) when after the > following command: > > CC="gcc -B/usr/bin/" ../gcc-4.3.2/configure --prefix=/tools > --with-local-prefix=/tools --disable-nls --disable-shared > --disable-libssp --enable-languages=c > > I received the following error message: > > configure: error: Building GCC requires GMP 4.1+ and MPFR 2.3.0+. > Try the --with-gmp and/or --with-mpfr options to specify their locations. > > The book did not explicitly say to compile the mpfr and gmp packages > (they are unpacked - as user lfs - into their specified directories) I am currently at the end of Chapter 5, and had no problem. I'm using the LiveCD 6.3 r 2160. Did you do the rename specified after the unpacks? Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Package User package management
I've been at the end of Chapter 5 pondering package management, and have finally decided to give the package user method a try. I don't like overburdening /etc/passwd with a lot of junk entries, but then the benefits look pretty good. In fact, it looks like a good idea even if another package manager is used to manage the installs and database entry. One slightly confusing point, though, is why the package management tools don't get installed using the package user philosophy. Why aren't the tools themselves installed into the chroot environment using the temporary tools in /tools? ISTM that the first package to get installed should be package-user, not linux-libc-headers. Is there some reason that isn't done, that I've overlooked? Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Package User package management
Support wrote: > Mike McCarty wrote: >> I've been at the end of Chapter 5 pondering package management, >> and have finally decided to give the package user method a try. >> I don't like overburdening /etc/passwd with a lot of junk entries, >> but then the benefits look pretty good. In fact, it looks like >> a good idea even if another package manager is used to manage >> the installs and database entry. >> >> One slightly confusing point, though, is why the package management >> tools don't get installed using the package user philosophy. Why >> aren't the tools themselves installed into the chroot environment >> using the temporary tools in /tools? ISTM that the first package >> to get installed should be package-user, not linux-libc-headers. >> >> Is there some reason that isn't done, that I've overlooked? >> >> Mike >> > Chapter was just the temporary system, you begin packaging everything in > chapter 6 if you want package management. Only thing I don't like about That's what I'm asking about. I guess I wasn't clear. I don't understand why the first package installed in Chapter 6, per the Hint, is libc-headers instead of the package manager. ISTM that the first package installed in the "real" system, that is the chroot environment, should be whatever package manager is being used, and it should be installed using itself. > the create user accounts method is exactly what you state, if > /etc/passwd gets too full, running your eye over it to ensure nothing > new and suspicious has cropped up becomes much harder. Just my 2 cents :) I guess there is no perfect system. I've used several, and each has advantages and disadvantages. Many of the better ones are not so suited for LFS because they are heavily optimized toward the "build one, install many" world, which is suboptimal for most LFS systems, I trow. I like to use tripwire, which catches changes to /etc/passwd and /etc/shadow, and one can also write a simple special tool to watch for changes to those files, which ignore entries in some file maintained by the package manager. That might be a reasonable solution. It could run as a cron job, and e-mail root when it finds "unusual" changes that is, changes not already white listed by the package manager. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Package User package management
Support wrote: > Mike McCarty wrote: >> That's what I'm asking about. I guess I wasn't clear. I don't understand >> why the first package installed in Chapter 6, per the Hint, is >> libc-headers instead of the package manager. ISTM that the first >> package installed in the "real" system, that is the chroot environment, >> should be whatever package manager is being used, and it should be >> installed using itself. >> >> > > Almost, you need to install your package manager at the end of chapter 5 > rather than start of 6, because then you have the package manager built > as part of the toolchain, ready for use in the final system construction. That's what I mentioned in the first message, which I didn't make clear, I guess. In between Chapters 5 and 6, one would build & install in /tools a temporary copy of the manager. Then, after the initial setup in Chapter 6, the first package built & intalled would be the package manager itself, using the temporary copy in /tools to install the permanent version into the chroot environment, that is, into what will eventually be the permanent system. The Hint for Package Users doesn't follow this, but leaves the package manager outside the control of any package management. This seems a little inconsistent, so I was wondering whether there was something I had not understood. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Package User package management
Support wrote: > Mike McCarty wrote: >> That's what I mentioned in the first message, which I didn't make >> clear, I guess. In between Chapters 5 and 6, one would >> build & install in /tools a temporary copy of the manager. Then, after >> the initial setup in Chapter 6, the first package built & intalled >> would be the package manager itself, using the temporary copy in >> /tools to install the permanent version into the chroot environment, >> that is, into what will eventually be the permanent system. >> >> Mike >> I see now that there are actually two somewhat related issues here, which I'm conflating. One is that the PM never gets put under its own control, ever, and the other is when the PM gets put under its own control. It might be easier (see below) for various PM techniques to defer placing the PM under its own control (doing an actual install as opposed to placing it into /tools somewhere where it can be run) until later in Chapter 6 than at the very beginning, and you seem to be addressing this. The other is, that with the Hints I've been reading, the PM never actually gets installed, that is, placed under its own control, it just gets stuck into the PATH where it'll get executed on an as needed basis. This doesn't seem right, somehow. If ping is so important that its version needs to be tracked, then isn't the PM even more important? (If you believe in PM at all, that is). I think that one would want the PM under control so that versions could be checked, changes noted, appropriate back-outs designed into it, etc. In other words, it could be managed as a package, rather than just a bunch of executables, data files, and directory structures, and could be uninstalled and a fall back to a previous version could be done, if needed, etc. > I see your thinking, but in theory (following the dependency tracking > malarky) you can't install the package manager first in chapter 6, > because the package manager would have dependencies of its own which > must be satisfied in order to install the package manager (yep, its a > headache). So: However, the specific package manager I have in mind is the Package User package manager, which (AFAICT) has all its dependencies already satisfied (except for su, which the hint describes how to put in there). > Chapter 5, build entire toolchain > End of chapter 5, build deps for package manager and then the package > manager itself [And place it into /tools, using any convenient technique (package management not necessary in the temporary build).] > Chapter 6, build packages of each section, in the order they appear in > the book, then install each package after you've built it You mean install it using the temp package manager in /tools, I suppose. > As for when you build the package manager for the final system, you > could do this at any point in chapter 6 where the deps of the package > manager have been satisfied, however, as you still have access to the Our thinking on this matter coincides, except that AFAICT the Package User tool set is a collection of scripts, and the only unfulfilled dependency I know of is su, which is easily put into the /tools location during Chapter 5. Other package managers might have more dependencies, which might not be met. In that case one has three options, I guess defer installing the package manager until all deps are met and just use the temporary tool in /tools until that happens, then build and install the "real" PM, and use it from then on make sure to build all dependencies the PM has in Chapter 5 and install them into /tools, then use the PM in /tools to do the install, using a --forced install, and use the "real" PM to do all the other installs "fake" an install by hand However, the Hint for Package Users never puts the PM under the control of the PM, ever. In fact, I don't seem to recall any of the Hints covering this topic, that is, placing the PM under its own control so its versions can be tracked, upgrades performed, etc. > one in the toolchain, I'd see no advantage of working out when its best > to do this, I'd just finished chapter 6 and then build my package manager. In any case, ISTM that even when the package manager has other dependencies, they could be built in Chapter 5 into the /tools area, and then it could still be the first thing run. However, even whenever the package manager gets built, it should be used to do its own install. That requires two installations, as do other parts of the tool chain, and ISTM that is best handled by building the deps (if any) in Chapter 5. If the package manager tracks dependencies, then the first install would have to be a --forced install, so that, even if its depend
Re: Package User package management
Tushar Teredesai wrote: > > If you want, you can install the package manager as a package > controlled by the package manager. Since the package manager in the > package user's approach does not have any build dependencies (only > run-time dependencies) it can be the first package to be installed in > chroot. I do that in my installation. Okay, that's what I thought. Thanks for the reply! > Of course if you have a package manager that has build-time > dependencies, the package manager should be installed later in the > process. Yes. However, I believe it should still be installed using the PM. I realize that thoughts on this matter differ. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: LFS 6.4 Book HTML
genericmailli...@gmail.com wrote: > > You don't need to change the book number just make the changes in > the book. On the page where the people will download the book just > warn them to check there for the newest version. This is what other > major open source developers do. Why do you want to make it > complicated. You are not empressing any one by doing it the old > hard way. What you propose sounds like a major train wreck type disaster in version control. > To deliberately limit yourself in that manner with a live flowing > electronic document is just foolish. Why limit your flow of > communication by following the traditions of an old technology that > does not have a better choice? The better choice is to have consistency and good version control. The book itself is part of an entire supported released system, not just a bunch of bits on a disc drive. You apparently have little experience with maintaining something like the enormous project LFS is, and the effort it takes to keep everything in sync in a project like LFS. What, on the surface, seems like stubborn reluctance to do things "the obviously better way" involves very much more than you imagine. The LFS book is used to create scripts, and requires, as another mentioned, creating directories, ensuring that no other defects were introduced, doing regression testing, etc., in order to make sure that everything is consistent. Having an errata page available seems like the perfect solution to sequestering the frequently changing parts from the parts which require careful analysis and quality checks. If the errata page gets broken, then it just gets fixed, and we go on. When the next release comes out, and we think we have enough changes to do a full rebuild and regression check, we incorporate all the known errata into the newly released version. It makes perfect sense to anyone who has to maintain and support a large project. The documentation is part of the revision controlled released product, not an afterthought on the web, which may or may not match the software package as a whole anymore, due to "fixes". This sort of thing is standard practice with all major projects, and has nothing to do with the medium in which the components get realized. It has everything to do with quality assurance and version control. Also, as pointed out, if one actually reads the book (interesting thought, actually reading it, as opposed to searching for interesting bits and then trying them out), one notices the errata notice right there. I had no problem finding it. True, even people reading things don't always notice what they just read, and may forget that they saw it. However, a quick recheck, or even an e-mail here, will soon put one on the right track. I trow you will not forget about the errata page in future, and neither will anyone else who has read this thread. Of course, in a few months, someone else will probably overlook the pointer in the book. Life goes on. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: LFS 6.4 Book HTML
genericmailli...@gmail.com wrote: [...] > Why do so many computer geeks want to make things complicated? I am not a member of the LFS support, development, or maintenance teams and efforts, and do not speak for any of them. I speak only for myself. I will allow that I am what you probably refer to as a "computer geek", since I've been team leader for projects involving as many as twenty software engineers and lasting for periods of up to two years for initial development, and participated in projects involving hundreds of engineers for even longer periods. I wrote my first computer program in 1967 or so. You are displaying complete ignorance of what the terms "Quality Control" and "Version Control" mean. I suggest that you study a little bit. The reason we want to do things the way you call "complicated" is because it vastly uncomplicates a number of other things, of which you seem to be completely ignorant. Experience with doing things the "simple" way has convinced those of us who actually have maintained large projects that it is much more complicated than what you perceive to be the "complicated" way. "Everything should be made as simple as possible, and no simpler." A.Einstein You want to make things simpler than is possible. It's like squeezing a balloon. If you squeeze out complexity in one area, it bulges into others. Supporting a static product, with static, known defects, each of which has a static solution is much simpler than supporting a non controlled fluid product which has no static defects, since nothing about it is static. A static product has, by its nature, a fixed finite number of (in principle) knowable and fixable defects. A fluid product, by its nature, has no finite limit to the number and kinds of defects it possesses. I think the comment that the type of volatility you seem to desire is better served by the "dev" non version of the book, software, and project, is spot on, and perhaps you should get your stuff there. Be aware, however, that some of that stuff is likely broken in places, the procedures described may not actually work as written, and support may be almost non existent, and may just get a response like "Yeah, we know that's broken; we expect to have it working an a couple of days." In fact, it is what you would get from LFS as the final product if they did things the way you suggest. Until you've walked a mile or so in the support team's moccasins, I suggest you not criticize their careful decisions based upon their painful experiences trying to do things the way that seems correct to you, and seeing them fail. In a project like this, coordination of effort is vital, and can be achieved, as far as my experience tells me, only through fixed static releases. Otherwise the support and maintenance teams have an impossible task. Large projects need careful separation and coordination of development, support, and maintenance. Sometimes the people can wear two or even all three of the hats, but the efforts need to be distinct. The errata sheets are a part of the maintenance effort, while the book itself is part of the support team's stuff (though the support team probably makes recommendations for the maintenance effort). You are suggesting conflating distinct efforts, which may be carried out by different teams of people. Doing as you suggest also, in a project like this, would likely involve the development team. That's my $0.02 USD worth. Since this has now gotten seriously OT for LFS, I think I'll not respond to this thread any further here. Let's take this off list or to /dev/null. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: LFS 6.4 Book HTML
genericmailli...@gmail.com wrote: > On Monday 06 April 2009 08:57:28 Mike McCarty wrote: >> You are displaying complete ignorance of what the terms >> "Quality Control" and "Version Control" mean. I suggest that >> you study a little bit. > > Why do you want to be insulting? That is rude and abrasive. > > When I used the word geek I was not being insulting because I do not > regard the term geek as insulting or degrading, although some do. Ok, fair enough. Where I come from "geek" is indeed an insulting term. I apologize. [...] > Did you mean quality control or quality assurance? Quality control. Assurance is something done "after the fact" so to speak. Control is done "during the fact". The definitions you quote seem pretty adequate, and better written than I could have done. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: LFS 6.4 Book HTML
genericmailli...@gmail.com wrote: > For 30 years I have been involved in different parts of Could we please take this off the list? I'm not going to respond further here. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: LFS 6.4 Book HTML
Simon Geard wrote: > On Mon, 2009-04-06 at 07:08 -0500, Mike McCarty wrote: >> This sort of thing is standard practice with all major projects, >> and has nothing to do with the medium in which the components >> get realized. It has everything to do with quality assurance >> and version control. > > Actually, I disagree - most projects don't provide errata pages, they I suppose that's correct. The "errata pages" don't go to the customer, they go to the support group, along with the work arounds. > provide support releases with the defects fixed, and discourage people > from using versions without the fixes. That's pretty much the opposite Yes, I agree. > of what LFS does - we expect them to follow instructions that are known > to have problems, and to use a separate errata document to know when > they'll run into those problems. > > If we were *really* following standard practice, we'd be applying the > errata changes against the 6.4 book, and putting a 6.4.1 book on the > website to replace it. Can't dispute. You'd do a "point release". Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: LFS 6.4 Book HTML
Simon Geard wrote: > On Tue, 2009-04-07 at 10:48 -0500, Mike McCarty wrote: >> Simon Geard wrote: >>> If we were *really* following standard practice, we'd be applying the >>> errata changes against the 6.4 book, and putting a 6.4.1 book on the >>> website to replace it. >> Can't dispute. You'd do a "point release". > > Right, which I assume is what the other poster was arguing for. The idea I don't think so, though that's possible. > of an errata page is very much a legacy of hardcopy publishing - for I trow not. > purely electronic documents, the software model of fix and re-release > makes much more sense. > > No idea how big a deal it would be to do, but in theory it's highly > automatible... A point release should go through full QA, which is why it would be a big deal, require full directory structure creation, regresion testing, etc. Someone would have to go through the full build following the book exactly. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Kernel Configuration
Is kernel configuration something which can be run not on the target system? I'm not referring to a cross-build. I don't have easy access to the information necessary to do the configuration right now on the target machine. I'd like to try jhalfs. I tried earlier, and even though I told it not to build the kernel, it insisted that it needed the .config for it. So, I wonder whether I couldn't run the configuration on another machine and transfer the resulting .config to the target and then proceed with the build. Doe the configurator query the running system? I don't see how that would be useful, so my guess is that one can do that. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Kernel Configuration
Trent Shea wrote: > On Wednesday 08 April 2009 12:11:29 Mike McCarty wrote: >> I'd like to try jhalfs. I tried >> earlier, and even though I told it not to build the kernel, it >> insisted that it needed the .config for it. > > What version of jhalfs are you using? I've been using svn for a number of > months (currently, 3526,) and I've never run into this. 2.3.1 In any case, I'd like to use jhalfs, even to build the kernel eventually, and that means I need to run the configurator. So, can I do that on another machine? Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Kernel Configuration
Trent Shea wrote: > > You're wondering if you can get the kernel config file from another machine? > If so, I cant' see any reason not to; I've done it a number of times, I'd > just > make sure that it gets created from the same kernel version. Yes, that's what I thought. I "made" the configurator and ran it on one machine, and saved the output to an "alternate" config file, for transfer to the target (different) machine. I don't see why that won't work, but if there were a problem, I didn't want to be posting here about how I followed the book, except I did something boneheaded that anyone familiar with building the kernel a few times would know wouldn't work. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
JHALFS: Permission Denied
I"ve been using Linux for several years, and various other *NIXes since about 1990 or so, working as a software engineer, so I feel fairly comfortable with the build environment. I've build LFS once before, and didn't want to go through it by hand again, so I'm trying jhalfs on a "blank" system I built up from pieces of this and that. I'm using jhalfs-2.3.1 on the LiveCD 6.3 r2160, book version 6.4. I boot the LiveCD... # mkdir /mnt/build_dir # mount /dev/hda1 /mnt/build_dir # chown jhalfs:jhalfs /mnt/build_dir # swapon /dev/hda2 # su - jhalfs $ startx I want to use a few windows, mostly so I can read the book while I do the build, and also read the jhalfs README etc., and look at things side-by-side. The windowing system starts up, and I create the following directories: /mnt/build_dir/BOOK /mnt/build_dir/sources and populate them. Next, I copy (so I have a persistent version) /home/jhalfs/jhalfs-2.3.1 $ cp -a /home/jhalfs/jhalfs-2.3.1 /mnt/build_dir I then unpacked the kernel sources into a temporary area and ran the configurator, to create /mnt/build_dir/kernel-config Next, I cd to /mnt/jhalfs-2.3.1 and "make". After I'm satisfied with the configuration parameters, I save them, and then tell it to go. It starts up, and verifies that it thinks it has a reasonable situation, and then asks me if I'm satisfied. I am, so it starts to run... Creating Makefile... START LFS/master.sh:line 381:cd: /mnt/build_dir/jhalfs/lfs-commands: No such file or directory Ok, $ mkdir -pv /mnt/build_dir/jhalfs/lfs-commands creates two directories, one under the other. Do another make, and tell it to "go" Now it stops No rule to make target makefile-functions $ locate makefile-functions finds one in /home/jhalfs/jhalfs-2.3.1/common, which I copy to /mnt/build_dir/jhalfs Do another make, and it stops again... Building target 020-creatingtoolsdir ln:creating symbolic link '/tools': Permission denied I look at the sudoers file, and all looks ok. I have no problem doing sudo from the command line. The user is in there as no password required. Ok.. $ sudo chmod 775 / $ sudo chown root:jhalfs / $ make This fails trying to create /mnt/build_dir/tools, which already exists. Ok $ rmdir ../tools $ make Now it fails with Building target 021-addinguser groupadd: unable to lock group file useradd: invalid numeric argument 'lfs' So, it looks like jhalfs (user) either can't do sudo (but he can, because I'm he, and I did a few) or the commands are not wrapped in sudo. This could possibly be because of the lack of a proper makefile-functions I guess, but I don't know. The $ locate only turned up one version of that, though maybe more exist: $ sudo find / -name "makefile-functions" /home/jhalfs/jhalfs-2.3.1/common/makefile-functions /mnt/build_dir/jhalfs-2.3.1/common/makefile-functions /mnt/build_dir/jhalfs/makefile-functions I tried mkdiring a "common" directory in there, and mving the makefile-functions into it, thinking maybe there was a path problem, but that only made it complain it couldn't find the makefile-functions file. Ok, now what? I can't just go on babying it along. Obviously, something about my setup was unanticipated by the tool, and I confused it somehow. What have I done wrong? Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
JHALFS 2.3.1 on LiveCD 6.3 r2160 is broken
I have struggled for a few days, a few hours at a time, to make the JHALFS shipped on (in) the LiveCD 6.3 r2160 work, and finally decided that it is just broken. So, I went to the website, and downloaded the current jhalfs-2.3.1.tar.bz2, transferred it to the disc on the build target, and ran it. It now runs without some warnings it got before, and is (apparently) successfully building an LFS system. The version which is actually on the disc, in the /home/jhalfs/jhalfs-2.3.1 directory is apparently broken. What do I need to do to make a defect report? Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: JHALFS 2.3.1 on LiveCD 6.3 r2160 is broken
Bruce Dubbs wrote: > Mike McCarty wrote: >> I have struggled for a few days, a few hours at a time, to make >> the JHALFS shipped on (in) the LiveCD 6.3 r2160 work, and finally >> decided that it is just broken. So, I went to the website, and >> downloaded the current jhalfs-2.3.1.tar.bz2, transferred it to >> the disc on the build target, and ran it. It now runs without >> some warnings it got before, and is (apparently) successfully >> building an LFS system. >> >> The version which is actually on the disc, in the >> /home/jhalfs/jhalfs-2.3.1 directory is apparently broken. >> >> What do I need to do to make a defect report? > > The "defect" is in the LiveCD Yes, of course. > and that is not being supported any more due to > lack of manpower. The CD is good for a rescue disk or as a basis for a new Ah, I see. > build, but the sources and support tools like jhalfs are out of date and > current > packages have to be downloaded. Ok, well, I've done that, so I guess I've got what can be gotten. Umm, would some additional manpower help? Perhaps I could contribute some. At the very least, I might be able to create a respin with fixed jhalfs, if that's not too difficult. I've got some machines sitting around idle, and I might be willing to use one for that. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Kernel panic - not syncing: Attempted to kill init!
ForrestG wrote: > Hi all, > > I succesfuly create lfs at my home PC so I copy all lfs files to CF > card. At cf card it was installed grub so I make menu.lst and I put > CF card into another PC. Linux kernel succesfuly booted and mounted > root directory but ist stopped with this kernel panic. > > init used greatest stark depth: 6000 bytes left Is that possibly a typo error for "stack". It looks like you are perhaps out of memory. How much RAM is on that machine? Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Kernel panic - not syncing: Attempted to kill init!
ForrestG wrote: > I have 256MB ram but motherboard detect only 128MB Well, that looks like a problem, but it doesn't look like the cause of your boot failure. The error report looks like you are using too much stack. If so, then the problem is not insufficient RAM. Unless you use a RAM disc, and that's eaten everything. Do you have a RAM disc, and if so how big is it? Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Report on JHALFS with LiveCD 6.3 r2160
Sorry this is longish, but I didn't want to post piecemeal as I went, with "this failed" and what I did, then later "oh, this failed, too" etc. Too much clutter. Well, I have fiddled this and that, and somewhat gotten JHALFS to work with the LiveCd as the host distribution. The first hurdle is that the JHALFS furnished on the LiveCD is broken. So, I installed the tarball from the release repository for jhalfs-2.3.1. During the configuration run, some of the MD5 sums didn't match. Well, that is a known problem with the book, I think. I changed to $BUILD_DIR/jhalfs and $ make mk_CHROOT to get it to that point, I hoped. It ran for some time, then failed trying to build Chapter 5 042-coreutils. After checking the logs, I found that it was running the tests, including the expensive tests, though I configured only to run the Chapter 6 tests, and only on the necessary parts (gcc et al.). Anyway, all but one test passed, so I investigated what happened. It was trying to verify that it could get name resolution for www.gnu.org fromhttp www.ibm.com fromhttps microsoft.com fromhttp google.org fromldap As this computer has no network connections at this time, this is an understandable failure. It's not so clear why it was running all the tests. I edited the $BUILD_DIR/jhalfs/lfs-commands/Chapter05/042-coreutils script to comment out the make for the tests, and restarted. It then failed on 085-perl, again during make test, this time during Chapter 6. In this case, the cause of the failure is somewhat more obscure (at least to me). The log reports _test_tty Error 1 ext/Sys/Syslog/t/syslog[...] Failed test 'setlogsock() should be true: ''' I didn't write down all of the second line. The syntax of the third line above is exactly as printed, which looks a little strange. It looks like it's failing some sort of socket attach when trying to make sure it can issue system logs. Anyway, the solution was as before, I edited the $BUILD_DIR/jhalfs/lfs-commands/Chapter06/085-perl script to omit the tests. Then it ran to 115-udev, which tar failed insisting it couldn't find ../udev-config-20081015.tar.bz2. That file exists in the repository I told jhalfs to use ($BUILD_DIR/sources-6.4/). So, it appears that the book parser didn't find that as a source file to copy for some reason, because tar is right, it isn't in $BUILD_DIR/sources. Anyway, I copied it by hand, and now the system thinks it has completed Chapter 6. However, as a quick check, I ran the tests on the compiler suggested in the book. I $ sudo chrooted as instructed to do in the "after this point use the modified chroot command" and built the "dummy" program as instructed in section 6.14.1. This failed in certain incompatibilities for some reason. First, it complained about "-W1,--verbose" being unrecognized. I changed that to "-W1 --verbose" (one), and it complained about "-W1". I changed it to "-Wl --verbose" (ell), and it complained also. I changed it to "-Wall --verbose" and it was happier, and yielded the a.out file. I did the readelf command, and got the expected output. However, the grep of the dummy.log file does not show the expected "succeeded". In fact, nowhere in the file does that string occur. Those files were the ones requested, and they do exist with the path shown in the book. Also, I changed the file dummy.c using the new vi to "hello world" with appropriate includes, and that runs and does what we want. The grep for the include file paths looks correct. So, jhalfs-2.3.1 "almost" works with LFS 6.4, but not quite. I don't understand why the dummy.log file does not contain exactly the expected output, nor why cc doesn't like the command string in the book. Inside the chroot jail # /usr/bin/cc --version cc (GCC) 4.3.2 Outside the jail $ /usr/bin/cc --version cc (GCC) 4.1.2 So it appears that I'm running exactly the version specified, but not getting the expected output. I rechecked the errata, and there is nothing about that topic. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
LFS 6.4 Lives!
Well, JHALFS took some work (see my earlier report) and it took me four times configuring and building the kernel to get one which would recognize all the disc drives, but LFS 6.4 now successfully boots and runs, and I can mount each of my CD-ROM drives. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Mount USB Hard Disc
Ok, now that my system boots and can basically run, I want to be able to mount USB discs, as I have one. However, it seems not to recognize that. I guess I need to configure at least one more thing into the kernel. Is there anyone here who can tell me what I left out? Thanks! When that works, I'll have to start working on my ethernet devices. And after that, on to BLFS! Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Mount USB Hard Disc
Bruce Dubbs wrote: Thanks very much for the reply. > Mike McCarty wrote: >> Ok, now that my system boots and can basically run, I want to >> be able to mount USB discs, as I have one. However, it seems >> not to recognize that. I guess I need to configure at least >> one more thing into the kernel. Is there anyone here who can >> tell me what I left out? > > You should have configured the kernel for USB and probably VFAT. Ok, I have those, but compiled as modules. Probably if I forced loading those modules, it would work. I guess I'll reconfigure and recompile with "y" instead "m". > CONFIG_FAT_DEFAULT_IOCHARSET="iso8859-1" Umm, I have "ascii" here. > CONFIG_NTFS_FS=y > > I recommend compiling in drivers and only using kernel modules when you have > to. Right, makes sense for stuff like this. Then the driver is there regardless of whether discovery found the need for it during boot. > Plugging in a memory stick will be handled by udev. The stick will show up > as > /dev/uba (I think), but you will have to mount it manually: I also have "CONFIG_USB_ANNOUNCE_NEW_DEVICES=y" but dmesg didn't show anything when I plugged in the cable. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Suggestion for Probe Program
I have just built LFS 6.4 and am in the process of ironing out some kinks. I ran a simple minded performance check, by simply copying the LiveCD to /dev/null, and comparing that with /dev/zero -> /dev/null. dd if= of=/dev/null bs=1024 count=636340 Source realusersys -- - 52x CD 31m6.06s0m0.176s0m4.918s That looks like I've got non-optimal driver there. Anyway, I'm going to try this test again booted with the LiveCD and see how fast it runs. Could just be the small block size. In an attempt to discover what is under the hood on this machine without peeling back the hood and trying to decipher chip markings, I might use /proc when the system is booted with the LiveCD. ISTM that for someone who knew what he was doing, it wouldn't be too hard to write something that could convert that info into recommendations for kernel config. Just a thought. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Vulnerabilities in udev
Ken Moffat wrote: [...] > There are two vulnerabilities in versions of udev before udev-141. Thanks very much for the heads up. > (i.) For all previous versions, netlink messages can be received > from local users, allowing privilege escalation. CVE-2009-1185 > > (ii.) There is a potential buffer overflow in the util_path_encode > function - rated as a denial of service. This function was > introduced comparatively recently (somewhere between versions 114 > and 124) so it does not apply to older versions. CVE-2009-1186 > > All users who run udev are recommended to upgrade and reboot. Why? What I see there shows two vulnerabilities indeed, but perhaps not for everyone. ISTM that they require a hostile local user, or at least one with a running local agent. I don't see how my LFS machine is vulnerable if no serial cable is connected no network cable is connected no PLIP is running or connected nobody lives in my house who wants to do my machine mischief I am not expert, so I perhaps am not able to see how the vulnerabilities listed affect my machine. Could you be more specific about how the vulnerabilities are subject to exploit? I'd appreciate that very much. IOW, I'd like to see something which would allow us to evaluate what our exposure might be. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Vulnerabilities in udev
Ken Moffat wrote: Thanks for your kind reply. > On Mon, Apr 27, 2009 at 12:53:41PM -0500, Mike McCarty wrote: [...] >> I am not expert, so I perhaps am not able to see how the vulnerabilities >> listed affect my machine. Could you be more specific about how the >> vulnerabilities are subject to exploit? I'd appreciate that very much. >> IOW, I'd like to see something which would allow us to evaluate what >> our exposure might be. >> >> Mike > If, like many of us, you only have a single human user then you can > do a risk assessment and decide you don't need to update. Nobody can > recommend running known-vulnerable software, but for _everything_ on > your LFS box you make your own choices. Well, naturally, but what I had in mind was how to evaluate my exposure. > If you have multiple human users, it is generally a good idea to > mistrust them when you are in your sysadmin role. Yes, of course. That is good advice. I actually also mistrust *myself*, which is why I don't log in as root, normally just do a few sudo's, and if necessary su to root for a little while. Not that I think I am malicious (to my *own* machine, anyway :-) ) but because everyone makes mistakes, and I'm not immune. > I'm not an expert either, and unlike regular distros we can't > subscribe somebody to the full-disclosure list where at least one > proof-of-concept has apparently circulated. Well, you see there are two exposures involved, the obvious one possible exploit of known vulnerability and the less obvious one replacing working code with with defective code The first exposure is relatively easy to evaluate; the latter is less so, but exists nonetheless. I like to hear that a given patch or other fix has "burnt in" for a while, especially where exposure due to the know vulnerability has low or even nonexistent possibility of exploit. I was hoping to get more information about how to evaluate my exposure. Thanks again! Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Vulnerabilities in udev
Bruce Dubbs wrote: Thanks for your reply. > Mike McCarty wrote: > >> I am not expert, so I perhaps am not able to see how the vulnerabilities >> listed affect my machine. Could you be more specific about how the >> vulnerabilities are subject to exploit? I'd appreciate that very much. >> IOW, I'd like to see something which would allow us to evaluate what >> our exposure might be. > > You're right Mike, not all vulnerabilities are equal. However it is good > practice to fix known vulnerabilities. If, for instance, you decided to run > a It is also good practice not to replace otherwise working code with possibly defective code, especially if the possibility of exploit is small to non existent. I was hoping to get information to enable me to evaluate my risk to exploit. > web server or even give yourself the capability to ssh into the system from > outside your home and there was a problem with that server software, a local > vulnerability could then lead to a root compromise. Yes, certainly. Neither of those is anything I ever intend to do. ISTM that the exposure my machine has is nil at present, and I see no reason to risk running unseasoned changes unless one can demonstrate actual possibility of exploit. For that reason, I am wary of publishing blanket recommendations for all users to replace working software simply because there is a known vulnerability. A vulnerability with no possibility of exploit is not a liability. Unseasoned code is a greater risk in that circumstance. Thanks also for the instructions! Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Vulnerabilities in udev
Bruce Dubbs wrote: > Mike McCarty wrote: [...] >> I was hoping to get more information about how to evaluate my exposure. > > Look at the source of the patch. The header says that the changes are from > upstream. They will be in future versions of the code. To evaluate the > vulnerability, the header says it fixes CVE-2009-1185 and CVE-2009-1186. > Google > that and you can read all about it. Thanks! Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Vulnerabilities in udev
Agathoklis D. Hatzimanikas wrote: > On Mon, Apr 27, at 02:52 Bruce Dubbs wrote: >> Mike McCarty wrote: >> [...] >>> >>> I was hoping to get more information about how to evaluate my exposure. >> Look at the source of the patch. The header says that the changes are from >> upstream. They will be in future versions of the code. To evaluate the >> vulnerability, the header says it fixes CVE-2009-1185 and CVE-2009-1186. >> Google >> that and you can read all about it. > > This is pretty serious. > > http://cve.mitre.org/cgi-bin/cvename.cgi?name=cve-2009-1185 > http://c-skills.blogspot.com/2009/04/udev-trickery-cve-2009-1185-and-cve.html > http://xorl.wordpress.com/2009/04/17/cve-2009-1185-linux-udev-path-encoding/ I'll look into that. Thanks! > > And for Mike, > > LFS is actually teaching administration, so it has the obligation and > the duty to teach the user to follow religiously the recommendations as > far it concerns security problems; in fact I would like to use the > *preach* expression instead of teach, to emphasize how the administrator > should take seriously the security domain. Umm, while I agree in principle with taking things like this seriously, an important part of administration is knowing not only what vulnerabilities there are, but what one's own exposure may be. So, simply issuing serious warnings and recommending all to upgrade doesn't give the full story. > So while the individual can choose, by looking to the security reports, > not to fix the vulnerabilities in her machine, in LFS we have *no* other > choice than to report them and publish fixes (if available), no matter how > critical they are. We are talking here about practices. No argument here about publishing, and I appreciate it. I'm not convinced that best practice is always to accept security updates immediately, or even ever (necessarily). I've seen security updates retracted. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Error while compiling util-linux-ng-2.14.1 LFS Chapter 5.32
jnbut...@jnbutler.com wrote: [...] > Wish me luck! Good luck! Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: util-linux-ng-2.14.1
jnbut...@jnbutler.com wrote: [about an include file problem] > I followed the LFS book instructions to the letter. You _thought_ you followed it to the letter. > Shouldn't it have installed ncurses correctly? Yes. > I went back and extracted ncurses again and followed > the book instructions to the letter, and that time it > worked and util-linux compiled and installed correctly. Apparently, you did something differently this time. What the difference may be could be ascertained only if either you have an absolutely perfect and reliable memory, or you have a machine kept log of the actions you took both times. > I am in chapter 6 now and so far have got up to mpfr with > no problems. Will continue building packages this evening. Good luck! It seems you're having some good success. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Nice Kernel Reference on Web
http://lwn.net/images/pdf/LDD3/ch00.pdf through http://lwn.net/images/pdf/LDD3/ch18.pdf Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: error glibc
Anders Bolinder wrote: > hello > im in secion 6.9 (lfs 6.4) > im using glibc 2.9-20090216 insted of the version mention in the book (i > couldnt find that version) I found it with no problem. However, if you still can't, then I can e-mail you a copy. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
JHALFS didn't install Shadow
I recently used JHALFS to build the 6.4 book, using QEMU as the build "machine", that is, a hardware emulator, and the 6.3r2160 LiveCD ISO image. I used the tricks I used before when on a "real" machine to take out the tests that failed (and were not intended to run, anyway). Well, it did not build or install shadow, so I had no useradd, no ability to log in as root, etc. It took a little bit of reading to figure out what happened, and then I rebooted the "machine" using the LiveCD image, and did the build and install by hand, after which it now boots and runs just fine (such as it is, of course, not much of a system). Anyway, I didn't encounter this when running directly on an old HP XL844 machine, and thought I'd pass that along. When I get a little more time, I'll go have a look at the JHALFS logs and report back more fully. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Cannot open root device ""..please append a correct "root=" boot option...
Mauricio Henriquez wrote: > Hi guys, > > I just finish my third LFS system this time using LFS 6.4 ...as usual I > make all the process in a vmware machine becouse I use it for teching > porpuses, but now at the end I have a kernel panic due to the fact that > it can't mount the root filesystem, the error is the one at subject: > > Cannot open root device ""..please append a correct "root=" boot > option...here are all the available partitions: > [2.871] 0b00 1048575 sr0 driver: sr > This can be caused by a missing or incorrect driver for the "disc" you are using. IOW, you may be missing the driver for the part of the emulated chipset which runs the "disc". Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: [lfs] error compiling gcc in Pass 1
Ankur Sinha wrote: > On Fri, 2009-06-12 at 13:42 +0530, Aditya Bankar wrote: [...] >> This has to be done inside the gcc's source directory. If you have >> unzipped it outside the gcc's source directory then mpfr would not >> be found. I had done the same thing. The general instructions provide the clue needed there, I think. Each section presupposes that one has untar'd the source, and then cd'd to the resultant directory, before performing any of the instructions associated with the package. > Thanks. > > I had figured it was a small error somewhere.. > > Is there some documentation which provides _reasons_ as to what we are > doing? Or do we learn that as we read on? Umm, I thought the explanation was pretty clear. What is there about needing to build a compiler in order to build that needs explaining? Perhaps you mean the reason one needs to get the associated packages and untar them in the gcc directory? That's simply a matter of how the FSF decided to package gcc. You'd have to ask them, not the LFS team. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: [lfs] error compiling gcc in Pass 1
Justin Mattock wrote: > On Fri, Jun 12, 2009 at 7:13 AM, Mike McCarty > wrote: >> Umm, I thought the explanation was pretty clear. What is there >> about needing to build a compiler in order to build that needs >> explaining? >> >> Perhaps you mean the reason one needs to get the associated >> packages and untar them in the gcc directory? That's simply >> a matter of how the FSF decided to package gcc. You'd have >> to ask them, not the LFS team. >> >> Mike > I didn't know you can put mpfr/gmp in the gcc source tree > and have gcc build them automatically. > (was aware of different languages that can be added). > > I'm happy with compile mpfr, and gmp separately, then > compiling gcc,etc... I didn't intend to say what you infer. I'm talking about building gcc and what (apparently) it considers source dependencies. I don't know why those sources need to be there. I haven't supported gcc since about 1990 or so, when I helped do a port to the Z8000 series of processors. What I intended to say is that one would have to ask the gcc dev team why those things need to be there. I suppose if you wanted to you could go look at the sources to gcc and see what it really wants from them. Then perhaps you could figure out why they are packaged separately, even though the C language build (appparently) depends upon them. IOW, I was expressing ignorance on this point, and pointing to where one might get the information. The LFS team isn't in the business of supporting gcc, and I don't suppose it's reasonable to ask them to know why the gcc support team decided to package the required sources in the way they did. Sorry for any confusion I caused. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Help: ownership problem of $LFS/tools
Ken Moffat wrote: [...] > Can you get a root login shell by using 'sudo /bin/bash', and from there do > things in the way that the book says ? > > I understand the attraction of 'sudo' for regular distributions, but > configuring > /etc/sudoers is always a compromise between functioality and security. When > the aim is to use a distro as a host to build a new LFS system, actually > becoming root is far more straightforward. There may be an issue here, where 'sudo' is not actually a separate program, in which case "bash sudo " is different from "sudo bash " Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: ncurses unexpected EOF while looking for matching ``'
Esben Stien wrote: > Bruce Dubbs writes: > >> It's really doubtful that the compiled version of bash on the Live >> CD is the problem since it has been around since Dec 2007 and this >> is the first report of a problem. > > Well, I grabbed that live CD just yesterday, and I built LFS 6.4 and > then encountered this problem. I applied the patch and after that, it > worked. I did a complete build using the 6.3 LiveCD both on real hardware and with QEMU, and both went through. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: ncurses unexpected EOF while looking for matching ``'
Esben Stien wrote: > Mike McCarty writes: > >> I did a complete build using the 6.3 LiveCD both on real hardware and >> with QEMU, and both went through. > > You did a 6.4 build?. Yes. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: glibc-2.5.1 installation fails
William Immendorf wrote: > On Tue, Jun 23, 2009 at 12:15 PM, khaled gouaich > wrote: >> I use ubuntu 8.04 & LFS book 6.3 > Best solution: Don't use 6.3. That book is obslete, and it's trouble > following that book. Try ether 6.4, or if you are adventurous, try a > development version. I agree with the conclusion, but not the reasoning. If the same versions are used as specifically mentioned in the 6.3 book, they'll work today just like they did when the book was "new". Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: glibc-2.5.1 installation fails
William Immendorf wrote: > On Tue, Jun 23, 2009 at 8:15 PM, Mike McCarty > wrote: >> If the same versions are used as specifically mentioned in the 6.3 book, >> they'll work today >> just like they did when the book was "new". > That's WRONG!! They will not work with recent systems. Try it and see. Perhaps you know something I do not. I did read the release notes for several packages' differences between the versions used in book 6.3 and 6.4, and I didn't notice any additional architectures added. I saw defect repairs, and some security hole closures. Could you please point to a specific package version difference between the 6.3 book and 6.4 book which shows an architecture which got added? I'd be interested to see that. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: boot LFS panic !
rony jak wrote: > hi all,william i have successfully complete the LFS system :) and when i > rebooted my pc to new LFS system i got this error > //**ERROR** > Starting up > [3.123320] Root-NFS: No NFS server available , giving up . > [3.123411] VFS: unable to mount root fs via NFS, trying floppy. > [3.132644] Kernel panic - not syncing: VFS: Unable to mount root fs > on unknown-block(2,0) > //**ERROR** > and Caps Lock & Scroll Lock blinking :) > is any way to disable this VFS ? i would like to boot minimal LFS system > thanks & regards This looks like fallback. You likely misconfigured your kernel, and don't have drivers for the disc interface present in your hardware. Carefully check your kernel configuration and make sure that you compiled in the driver for the file system, and for the hardware present. Pay careful attention to the chipset drivers. Do you know your disc chipset? If not, does the host system have lspci? If so, then try running that and see if you get a list of PCI bus devices, including a disc interface. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Can not boot from LFS-6.4
khaled gouaich wrote: > Hi all, > I have a problem with bootig my lfs-6.4, here the out put : > > -VFS: cannot open root device "hda3" or unknown-block (2-0) > Please append a correct "root="; boot option .. > kernel panic-not syncing : VFS :unable to mount root fs on unknown-block (2,0) The kernel is looking for a file system to mount, and can't find it. The most likely cause for this is not building drivers compatible with your actual disc interface hardware into the kernel. Boot your host system, and use lspci or something similar to find out what your hardware chips are (or look inside your machine) and then check your kernel configuration and make sure you include drivers for those chips, and build them in (not make them loadable modules). Another possibility is that you don't have the file system driver built in. [...] Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: LFS 6.4 - Coreutils-6.12 - problems compiling chcon
Justin P. Mattock wrote: > With your host system you probably have > libselinux(and so forth), I recommend to go ahead and > building everything with SELiux support, > then, at the end of the day if you decide to add a policy > you do, if not then you don't. (at least the system has a security > infrastructure). > > Justin P. Mattock I recommend that, for a first build, NO DEVIATIONS are made from the directions in the book. Also, personally, one of the reasons I have built LFS is so I WON'T have SeLinux on my machine. Like everything, there are advantages and disadvantages to having SeLinux, even if it is not "enabled". YMMV Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: LFS 6.4 - Coreutils-6.12 - problems compiling chcon
Russell Stockhammer wrote: > > > I recommend that, for a first build, NO DEVIATIONS are made from >> the directions in the book. > > > I've made no deviations, I've cut and pasted command yet still I am > cursed with some echo of selinux. I was commenting on the recommendation of including SeLinux in a first build. As you can see, it's easy enough to have problems even when making "no deviations". Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: LFS 6.4 - Coreutils-6.12 - problems compiling chcon
Justin P. Mattock wrote: [...] > From what it sounds like you want a SELinux free system. > you might want to try the LFS livecd instead of RH due to RH > having SELinux enabled from the start. > (not sure if the lfs livecd has libselinux installed or not) > but if it doesn't then life for you will be easier, as opposed to a system > have all of the SELinux infrastructure compiled in it. I built LFS 6.3 on a Fedora Core 2 system with SeLinux on it, and was able to create a SeLinux free system. If the steps are followed correctly, then contamination from the development system should not occur. The toolset is built from scratch. It's very similar to cross compiling, actually. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: LFS 6.4 - Coreutils-6.12 - problems compiling chcon
Russell Stockhammer wrote: > OK thanks for everyones help by using coreutils-5.97 I mananged to > get past this step without anymore hassles. Congratulations! Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: making LFS
ajith silva wrote: > LFS 6.4 book mentioned two folders, > > /mnt/lfs/sources > /mnt/lfs/tools > > You have to work in /mnt/lfs/sources for all installation work. All > installation scripts will install required files in /mnt/lfs/tools > automatically when you execute proper commands according to the book > (LFS 6.4). I have buil LFS 6.3 and 6.4 (a few times for 6.4) and I haven't yet built any of them by hand (I did some jhalfs builds of 6.4 which did) by building in $LFS/sources > First consider page 37 of the book where you begin the LFS > installation with binutils-2.18. You have to do fllowing: > > copy binutils- 2. 18. tar. bz2 and binutils-2.18-configure-1.patch file to > /mnt/lfs/sources > then uncompress it as tar -jxf binutils- 2. 18. tar. bz2 > go into the directory binutils- 2. 18 (cd binutils- 2. 18) > issue the command as book says: patch -Np1 -i > ../binutils-2.18-configure-1.patch > then issue following command to create a folder inside /mnt/lfs/sources Well, you state that one has to do that, but I've not done it by hand for either 6.3 or 6.4, and I've done at least three builds by hand. I've lost count... > mkdir -v ../binutils-build > > > Please try to understand this. Your current location should be > /mnt/lfs/source/binutils-2.18/ and when you issue the command "mkdir > -v ../binutils-build" it will create the directory in > /mnt/lfs/sources/ as /mnt/lfs/sources/binutils-build. Well, aside from the error of substituting "/mnt/lfs" for "$LFS", what you wrote is just not what I experience. I don't like, and haven't used, the exact directory structure suggested in the book. > this means > > LFS book does not explain everything (specially location to execute > installation commands) Actually, it does. There is a section which describes "General Instructions" which explains something similar to what you wrote. > you should have basic knowledge about unix system and its command > (specially file system related command behaviour) This is certainly true. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Any issues upgrading LFS 6.3 -> 6.5
Trent Shea wrote: [...] > So, I guess the obvious issue would be the state of BLFS right now, and how > dirty you want to get your hands ;) Precisely. I've built 6.3 once, and 6.4 four times, but so far not BLFS, as I'm waiting for it to catch up. I'd like to see a coherent release of LFS + BLFS. I want to leave the standard distros and build/support my own, since I can't seem to find a distro which isn't bloated with lots of stuff I don't want or need, and find that somehow the distros are not flexible enough to accommodate really customizing the install. However, so far I haven't seen a coherent LFS + BLFS since 6.3. I didn't build BLFS yet, because I was waiting for a coherent 6.4 mate BLFS. Maybe I missed some news on the website? Anyway, if the dev team could concentrate on coherence, I sure would appreciate it, as my current distro is getting a little "long in the tooth" and is no longer supported by anyone, not even me :-) Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: Any issues upgrading LFS 6.3 -> 6.5
Bruce Dubbs wrote: > Randy McMurchy wrote: >> Mike McCarty wrote these words on 07/21/09 12:19 CST: >> >>> However, so far I haven't seen a coherent LFS + BLFS since 6.3. >>> I didn't build BLFS yet, because I was waiting for a coherent >>> 6.4 mate BLFS. >> You may wait a long time. :-) We have very few contributors with >> any spare time right now. >> >> >>> Anyway, if the dev team could concentrate on coherence, I sure >>> would appreciate it, >> See above, same thing applies. >> >> Mike, what I suggest is to just start building BLFS using the most >> recent packages the vendor supplies. If something doesn't work, or >> you have any questions, write in. Someone will respond (me if needed), >> and we'll find a solution. Don't wait for a BLFS release, just jump >> in and start building. >> >> Again, if something doesn't work, or you get stalled, write in and >> let us know. Someone will get you over the hurdle. Chances are you >> won't need much help, if any. > > Let me add that getting a 'current' BLFS is almost impossible. There are > about > 300 packages in BLFS. It each only changed once a year (an many change much > more frequently than that) it would be almost a package a day. It's a moving > target and with lots of help we could get close, but in the long run, > whatever > we produce is only a snapshot. Thanks for the replies, Bruce and Randy. I am not complaining, merely observing. I did offer to do some help with the dev team, if nothing else then just verifying that builds will work, but had no real response. Your observation also applies to LFS, however. Still, you manage to produce a list of coherent packages. I do realize that BLFS is much _larger_ than LFS. Perhaps your suggestion is the best. I'll give some consideration to how much of BLFS I intend to put on my machine(s), and document which versions I used. If I get a coherent package working, at least that would be a list someone else could start from. I mean a version list. Wouldn't guarantee that a larger list would be coherent. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: time command help
RaptorX wrote: > Hi guys, > > This is my first time building LFS 6.4, im in chapter 5.4.1 at the moment > and I wanted to run the benchmark with the time command... > > as far as I understood I should make something like: > > *CC="gcc -B/usr/bin/" time {../binutils-2.18/configure --prefix=/tools > --disable-nls --disable-werror && make && make install}* > > but i keep getting: > * > {../binutils-2.18/configure --prefix=/tools --disable-nls --disable-werror > && make && make install} > :No such file or directory* I believe that what may be missing is the time command. > I have tried several combinations but it simply does not work, I confirmed > that ../binutils-2.18 is correct and that the configure file is in there... > > if I run time {ls} it runs perfectly it gives me the time of execution... Yes, I encountered something like that myself. I believe that the "time" which is getting executed may be a bash builtin. > any ideas what am i doing wrong?... > > PS: > > I tried other commands... here is what i tried. > > time {CC="gcc -B/usr/bin" ../gcc-4.1.2/configure --prefix=/tools > --with-local-prefix=/tools --disable-nls --disable-shared --disable-libssp > --enable-languages=c && make && make install} > bash: {CC=gcc -B/usr/bin: No such file or directory > > real0m0.001s > user0m0.001s > sys 0m0.000s Yes, this won't work. Try something like time bash "{CC=\"gcc -B/usr/bin\" ../ [rest omitted for typing's sake] This times the loading of another shell, which gets called with an appropriate argument. You may need to have some switch to tell bash that you've passed in a string. I forget just what was the trick I used, but it was something like this. > time {'CC="gcc -B/usr/bin" ../gcc-4.1.2/configure --prefix=/tools > --with-local-prefix=/tools --disable-nls --disable-shared --disable-libssp > --enable-languages=c && make && make install'} > bash: {CC="gcc -B/usr/bin" ../gcc-4.1.2/configure --prefix=/tools > --with-local-prefix=/tools --disable-nls --disable-shared --disable-libssp > --enable-languages=c && make && make install}: No such file or directory > Yes, this won't work at all. It tries to tell bash to load a program named "{... Try my workaround and see how it goes for you. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: time command help
RaptorX wrote: > the following commands work fine: > > time ls > > time "ls" > > time 'ls' > > this one doesnt: > > time {ls} The reason for this is that time simply passes its argument to one of the exec() calls. So, "{ls}" is not the name of any command, and it can't be executed. However time bash "{ ls }" should work, because then the command is simply bash, which *is* found, and bash gets the string "{ ls }" as a command, which it can interpret. Those brackets and things are stuff bash knows how to interpret. However, that's a function of bash, not of the program loader. The program loader looks for files in the PATH. > time { ls } > > > so do we really have to enclose in brackets? -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: time command help
RaptorX wrote: > actually in this system: > Slackware 12.2 > > that returns with an error... > > *[~]$ time bash "{ ls }" > bash: { ls }: No such file or directory > > real0m0.002s > user0m0.000s > sys 0m0.002s > [~]$* > > it really seems to be with the version of bash that Slackware uses. Try $ time bash -c "{ ls; }" All these strange issues would go away if the actual time command got installed. As it is, you are simply using the reserved word in bash. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
Re: time command help
Trent Shea wrote: > On Wednesday 22 July 2009 07:46:58 RaptorX wrote: >> the following commands work fine: >> >> time ls >> >> time "ls" >> >> time 'ls' >> >> this one doesnt: >> >> time {ls} >> >> time { ls } >> >> >> so do we really have to enclose in brackets? > > http://tldp.org/LDP/abs/html/functions.html > > time { ls; } > > The first space is required, I believe the last one is just for formatting. > > > Quote: > To achieve this easily, wrap the three commands in a time command like this: > time { ./configure ... && make && make install; }. Everything's there ;). Yes, but not quite the situation at hand... The original question related to a command like this: $ CC="gcc -B/usr/bin/" time { ls; } which fails. However, this works for me... CC="gcc -B/usr/bin/" bash -c 'time { ls; }' I guess that's the workaround I used. Mike -- p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} Oppose globalization and One World Governments like the UN. This message made from 100% recycled bits. You have found the bank of Larn. I speak only for myself, and I am unanimous in that! -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page