Re: [blfs-dev] Kernel issues
> Answering Alain: I haven't looked at iwd yet, but we may need that > also. > > Finally, we may need a new page: "About Wireless" I agree to both. For what it's worth, in my builds, I hand craft the svn checkout of lfs and blfs to, for the former, use the latest long-term kernel (currently 4.14.89) and set glibc's --enable-kernel=4.14.75 (a few version below what's running on the linode) instead of 3.2 but I do agree the glibc setting is not relevant to wireless. Haven't had the chance to build libnl and wpa_supplicant later on because the first few build had illegal instructions in libffi and gmp (the linode use a haswell or a broadwell xeon and none of my computers are that recent) but I did make the necessary changes to compile using generic settings in this build. Alain -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Kernel issues
On Wed, 2018-12-19 at 10:37 -0600, Bruce Dubbs via blfs-dev wrote: > On 12/19/2018 10:00 AM, Bruce Dubbs wrote: > > On 12/19/2018 02:55 AM, Xi Ruoyao via blfs-dev wrote: > > > On 2018-12-18 15:11 -0600, Bruce Dubbs via blfs-dev: > > > > I have figured out my original problem with loading the > > > > firmware for > > > > my > > > > wireless and bluetooth. It seems that the drivers for those > > > > were > > > > not > > > > enabled. Evidently I used the wrong base system when building > > > > the > > > > kernel. > > > > > > > > I still do have the message about regulatory.db. Now that the > > > > interface, wlan0, is recognized by the kernel, I can set it up > > > > with > > > > > > > > # ifconfig wlan0 up > > > > > > > > but I also get: > > > > > > > > # iwlist wlan0 scanning > > > > wlan0 Interface doesn't support scanning. > > > > > > > > and I suspect that has to do with the regulatory.db stuff in > > > > cfg80211. > > > > I was able to find: > > > > > > > > https://mirrors.edge.kernel.org/pub/software/network/wireless-r > > > > egdb/wireless-regdb-2018.10.24.tar.xz > > > > > > > > > > > > > > > > Arch has build instructions: > > > > > > > > https://git.archlinux.org/svntogit/packages.git/tree/trunk/PKGB > > > > UILD?h=packages/wireless-regdb > > > > > > > > but it needs the crda package: > > > > > > > > https://www.archlinux.org/packages/core/x86_64/crda/ > > > > > > crda is only used to verify the regulatory.bin file. To keep it > > > simple > > > and stupid we can just install regulatory.db, regulatory.db.p7s, > > > and > > > the man page regulatory.db.5. > > > > > > With regulatory.db directly loaded into the kernel /sbin/crda > > > seems > > > useless. I moved it away and my wpa_supplicant can still change > > > the > > > regdom to CN (automatically, matching my OpenWRT router). > > > > Do you just add regulatory.db to the kernel option > > CONFIG_EXTRA_FIRMWARE > > or is there more to it? > > > > I do have regulatory.db in /lib/firmware and in the kernel source > > directory firmware/. I'm not sure what the .p7s file is. Possibly > > a > > crypto signature. > > I did get the regulatory.db to load by adding both the .db and .p7s > files in CONFIG_EXTRA_FIRMWARE: > > [0.568710] microcode: sig=0x306c3, pf=0x10, revision=0x22 > [0.568923] microcode: Microcode Update Driver: v2.2. > [0.568944] sched_clock: Marking stable (556629458, > 12282386)->(584360640, -15448796) > [0.569404] registered taskstats version 1 > [0.569406] Loading compiled-in X.509 certificates > [0.570431] Magic number: 14:89:490 > [0.570664] cfg80211: Loading compiled-in X.509 certificates for > regulatory database > [0.571735] cfg80211: Loaded X.509 cert 'sforshee: > 00b28ddf47aef9cea7' > > > Also how do you set the regdom? Do you use iw? > > Adding the db is not enough. I don't know hos to set the regdom > other > than with: > > # iw reg set US > > but I haven't built iw yet. > > -- Bruce Hi Bruce, Have you looked into the iwd daemon (https://iwd.wiki.kernel.org/). The thing is supposed to replace wpa_supplicant I'll be installing it on the LFS+BLFS setup I'm building at the moment on a linode server (I recently ordered a 8GB 4 core vm from linode which is used to do repeat build of various LFS+BLFS setups). Alain p.s. posting from gmail because mail.vocatus.pub doesn't want me to send any mails. -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Gnome-online-account breakage with owncloud plugin
> Anyway, please report if there is any more breakage from poppler. LibreOffice 6.1.3.2 at the moment. I decided to postpone the build and look at it later. The rest should go easy peachy. Alain -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Gnome-online-account breakage with owncloud plugin
Le dimanche 25 novembre 2018 à 15:53 -0600, Douglas R. Reno a écrit : > On 11/25/18 3:52 PM, Alain Toussaint wrote: > > > My hope is to get gnome-online-accounts updated to 3.30.x within > > > the > > > next day or two. I'll reply with the revision number once I'm > > > done, > > > but > > > please try using the 3.30.x series instead. > > > > > > > > > Thank you for the report. > > > > > > > Just checked out and there is a 3.31.2 version (of evolution at > > least > > but I expect the other to follow suit, http://ftp.acc.umu.se/pub/GN > > OME/ > > sources/evolution/3.31/ ). Want me to try 3.31.2? > > > > Alain > > I'd prefer 3.30 if you can, but please try 3.31.x to see if it > works. > Thank you! Just to report back that I installed as many packages as I can in the gnome 3.30 series and they all built fine with a little massaging for some packages and lot of massaging for other packages. That include out of BLFS packages (like googles' protobuf and libphonenumber for extras functionalities out of evolution-data-server...there are other too). It's building in a chroot as it is now and once it's done, I'll report back on usability testing & stability but there are some heavyweight in the pipeline (LibreOffice & Firefox). Alain -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Poppler breakage
> > Doing something like this will help downstream to make their code > > conform to your expectations and not have developers incorrectly > > blame > > poppler for breakage when you need to make changes." > > > > - > > And the answer: > https://gitlab.freedesktop.org/poppler/poppler/issues/675#note_82490 > >-- Bruce And I'm thinking of a possible solution but: 1-: it would require some heavy lifting in downstream packages. 2-: I, personnally, don't want to make any promises but I have a plan on getting a 8GB 6-core linode which will be used as a continuous integration system to build LFS/BLFS linux images suitable for a range of systems including my, and friends' desktop, server images for the other nanodes that I use and a few other computing devices. Still no promise because so far, I'm lagging on my current commitments. Alain -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] perl module issues
Le lundi 26 novembre 2018 à 12:24 -0600, Bruce Dubbs via blfs-dev a écrit : > Reference: > http://wiki.linuxfromscratch.org/blfs/ticket/11246 (git and shipped > perl > modules) > http://wiki.linuxfromscratch.org/blfs/ticket/11295 (Error-0.17027 > (Perl > module)) > > I've been looking at these tickets, but want to raise them here for > more > visibility. In #11246, Ken prefers to remove Error from the perl > modules because it is only used by git and git provides a copy in > /usr/share/perl5/Git. > > I agree. Actually, I think we can just archive > general/prog/perl-modules/perl-error.xml, mark #11295 as > overcomebyevents, and mark #11246 fixed. > > I do not see a need to put a comment in the git section. That sets > a > precedent to document every package that has some sort of shell > script > or an executable in /usr/libexec or in a dedicated location. > > Opinions? > >-- Bruce Entirely agreed to archiving perl-error.xml Alain -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] dconf 0.30 need bash-completion dependency
> https://github.com/scop/bash-completion/releases/download/2.8/bash-co > mpletion-2.8.tar.xz > > But it did not used to be required. If it is now, it should be easy > to circumvent. > > --DJ Maybe, but given that I'm the type of "shoot first, ask question later" at this moment, I installed it and will investigate more after having built the whole stack (it's been close to a week and a half I've been building). Alain p.s. gnome entire desktop, LibreOffice and Firefox with all optional and suggested dependencies. -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-dev] bogofilter update
Hello, It seem like bogofilter's source has been updated while keeping the same version. New md5sum is supposed to be: 94aee3ced558c2a354547083735f2ba7 Alain -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-dev] dconf 0.30 need bash-completion dependency
Hello, just FYI, dconf 0.30 now require a package named bash-completion for which is found on github: https://github.com/scop/bash-completion.git but for which, debian and other distro package it out in tarball form from this git tree. I used git clone -b 2.8 https://github.com/scop/bash-completion.git to download and install Alain p.s. dconf itself is needed by gdm 3.30.2 -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] cups variables for C++
> If specifying CXX=g++ will disable those warnings, that is good. > But does it, or will cups decide to anyway use clang++ ? Nope, cups will use g++ when CXX=g++ Alain -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-dev] LibreOffice 6.1.3.2, was Re: Gnome-online-account breakage with owncloud plugin
> 3.30 it is then. After sending the message, I did think about the > possibility of odd releases being development release while even > release are stable ones... I also took the occasion to modify my ALFS setup to install LibreOffice 6.1.3.2 while at it. If LO work fine on my setup, I'll modify the various xml files relevant as well as the package.ent (which is already done for gnome and gnome.ent before feeding it to jhalfs) file and submit back. Alain -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Gnome-online-account breakage with owncloud plugin
Le dimanche 25 novembre 2018 à 15:53 -0600, Douglas R. Reno a écrit : > On 11/25/18 3:52 PM, Alain Toussaint wrote: > > > My hope is to get gnome-online-accounts updated to 3.30.x within > > > the > > > next day or two. I'll reply with the revision number once I'm > > > done, > > > but > > > please try using the 3.30.x series instead. > > > > > > > > > Thank you for the report. > > > > > > > Just checked out and there is a 3.31.2 version (of evolution at > > least > > but I expect the other to follow suit, http://ftp.acc.umu.se/pub/GN > > OME/ > > sources/evolution/3.31/ ). Want me to try 3.31.2? > > > > Alain > > I'd prefer 3.30 if you can, but please try 3.31.x to see if it > works. > Thank you! > 3.30 it is then. After sending the message, I did think about the possibility of odd releases being development release while even release are stable ones... Alain -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Gnome-online-account breakage with owncloud plugin
> My hope is to get gnome-online-accounts updated to 3.30.x within the > next day or two. I'll reply with the revision number once I'm done, > but > please try using the 3.30.x series instead. > > > Thank you for the report. > Just checked out and there is a 3.31.2 version (of evolution at least but I expect the other to follow suit, http://ftp.acc.umu.se/pub/GNOME/ sources/evolution/3.31/ ). Want me to try 3.31.2? Alain -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-dev] Quick detail about OpenJDK
Hello, Since the start of the week, I've been building a gnome desktop based Linux system with all the recommended and optional packages included via the ALFS build tool and for OpenJDK, the build took 894 minutes (yes, nearly 15h) only to fail because 7 tests out of the testsuite failed. Easy fix for me was to remove the test run from the build script. Last detail: 2 cores pentium with 4GB of ram and everything compile using one core while still being chrooted. This the reason why it takes forever to build. Alain -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-dev] cups variables for C++
Hello, It might be useful to add CXX=g++ to the configure script of cups: before: CC=gcc \ ./configure --libdir=/usr/lib\ --with-rcdir=/tmp/cupsinit \ --with-system-groups=lpadmin \ --with-docdir=/usr/share/cups/doc-2.2.8 && make after: CC=gcc CXX=g++ \ ./configure --libdir=/usr/lib\ --with-rcdir=/tmp/cupsinit \ --with-system-groups=lpadmin \ --with-docdir=/usr/share/cups/doc-2.2.8 && make Some of the files are in C++ and clang++ have issues when having CFLAGS and CXXFLAGS turning on flags that work for gcc/g++ and not in clang/clang++ Alain -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
[blfs-dev] rpcsvc-proto: md5sums doesn't correspond
Hello, it's been some time that rpcsvc-proto's md5sums doesn't corresponds: alain [ ~/blfs_root/work ]$ md5sum /sources/rpcsvc-proto-1.4.tar.gz df1022ccc5d0aff30b8e505735b4a98b /sources/rpcsvc-proto-1.4.tar.gz in the book, it still refer to an md5sum of: 3b2724bb7d6560ddcb4dbcdb5465de10 Have they updated the sources while still keeping the same version? Alain -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] openssl (sendmail) lfs/blfs discrepancy
Le jeudi 15 novembre 2018 à 09:09 -0500, Jean-Marc Pigeon via blfs-dev a écrit : > Hello, > > > As reported earlier, I am not able to recompile > sendmail-8.15.2. > > It is not fully proved but I have strong indication > it is caused by openssl-1.1.[0i,1] > > Looking about this, I noticed > LFS is openssl-1.1.1 > BLFS is openssl-1.0.2p > (LFS 2018-11-09, BLFS 2018-11-13) > > Is this done on purpose? (I am doubtful about > it) There are two version of openssl because some software in BLFS are not able to work with openssl 1.1, sendmail being one of them, ref: https://mta.openssl.org/pipermail/openssl-users/2016-June/003897.html > Trouble ahead compiling sendmail (and other?) > using opessl-1.1.1? > > Comments?, something I am overlooking? > I don't think so but do try to compile sendmail with openssl-1.0.2p. I'm pretty sure it will work fine :) Alain -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] LFS-8.3, ISO , XFCE, LXQT.
> > > Caution: On a GeForce GT 710, falkon crash as soon > > > trying to display an URL, It was needed to implement the > > > last proprietary Nvidia driver (standard "Nvidia-run" > > > procedure). > > > For other (more recent nvidia card) falkon is working fine. > > > > > > > I refuse to touch nvidia drivers, and after my experience with > > nouveau on a GT710 (worked fine if Xorg not used, e.g. for a > > server, > > disastrous for crashing if Xorg in use) and from reading reports > > about the problems of using current nvidia cards with nouveau, I > > will not again use nvidia cards on a desktop machine. > > > > I was providing "pure data". > > The implication is the nouveau driver was not working A_One > for "old" card as GT 710. > I found no trouble with the nouveau driver (from LFS) with > other nvidia card I have here. > ATI cards seem to be working fine too. > > It is my understanding nvidia card are pretty common, so > my comment was rather to help someone using an, > old and on the side, PC to do some tests (sharing finding, > this is the whole purpose of a mailing list). In my case, I have an old GTX 460 card in this workforce and I can attest that in kernel 4.18.X, nouveau is unusable for me and matter of fact, the 4.18.X serie of kernels are unuseable so I reverted back to one of the long-terms, that is 4.14.78 which work fine under the proprietary drivers. I haven't tested with nouveau under 4.14.78 but I'll see if I can test soon. Alain -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] BLFS Basic
> Actually I prefer gdisk also, but fdisk from LFS can do msdos > partitions > as well so I felt gdisk was redundant. Just my 0.02$ in that regard but for {u}efi, I prefer dealing with gdisk as opposed to parted but then again, I'm not a new user. Alain -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] keyutils regression tests
> Note that the kernel option descriptions say > >If you are unsure as to whether this is required, answer N. > > for CONFIG_BIG_KEYS and CONFIG_KEY_DH_OPERATIONS. > > > Opinions? There's a tangent to that that some time ago that Linus refused a patch regarding blk-io scheduling because of a lack of explanation text regarding the option itself so I'd be inclined to file a bug report on the kernel documentation itself if it isn't explanatory enough...but then again, this is my 2 grain of salt. Alain -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
Re: [blfs-dev] Last call for BLFS-8.3
Le vendredi 31 août 2018 à 11:42 -0500, Bruce Dubbs via blfs-dev a écrit : > We now have all packages in BLFS tagged for 8.3, There are still a > couple of outstanding tickets but they can be put off if they are > not > done in the next day. I can't vouch for dovecot 2.3.2.1 because of my linking issues and also, I have yet to put it up on my server (@vocatus.pub) but the test suite ran fine. I can upload testsuites logs if needed (2 logs, first is stdout and second is stderr). Alain -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page