- CRUX - lists.crux.nu

3.8 rc4
by Daryl F 13 Mar '25

13 Mar '25

Good job Devs! This one is is working well. I upgraded a 3.7 VM to 3.8 rc4. Then I rebuilt all the ports in the df repo. Here's the results from that. python3 new version caused footprint mismatch in 40 ports. perl new version caused footprint mismatch in 19 ports. The defaults in gcc have changed a bit. Some warnings have become errors. glibc has rearranged something. Some function prototypes are no longer available as they were before. So some older source code had to have #include's added. That's all for the tl;dr. What follows is notes on problems and the resolution for each. Thanks again Devs! Tested CRUX 3.8-rc4 upgrade on a CRUX 3.7 VM. Worked well. Spent some time fiddling with a few error messages before I remembered to rejmerge :-O cyclic dependencies: python3-tomli python3-flit-core Haven't sorted this out yet but it is not serious. These seem to be in the df repo from CRUX 3.5/3.6 days. Since they are in opt repo they will be removed from df repo. #chmlib (df) chm_lib.c:727:15: error: implicit declaration of function 'pread64'; did you mean 'pread'? [-Wimplicit-function-declaration] 727 | readLen = pread64(h->fd, buf, (long)len, os); | ^~~~~~~ | pread NOTE: contrib version of chmlib will compile. This is a manifestation of the glibc change. #volume-key (df) libtool: install: (cd /var/tmp/crux/volume_key/src/volume_key-0.3.12/python/python3; /bin/sh "/var/tmp/crux/volume_key/src/volume_key-0.3.12/libtool" --tag CC --mode=relink gcc -O2 -march=x86-64 -pipe -Wall -W -Wcast-align -Wmissing-noreturn -Wnested-externs -Wpointer-arith -Wshadow -Wundef -Wwrite-strings -module -avoid-version -lglib-2.0 -L/usr/lib -ldl -lm -o _volume_key.la -rpath /usr/lib/python3.12/site-packages _volume_key_la-volume_key_wrap.lo ../../lib/libvolume_key.la -ldl -lm -lglib-2.0 -lssl3 -lsmime3 -lnss3 -lnssutil3 -lplds4 -lplc4 -lnspr4 -lpthread -ldl -inst-prefix-dir /var/tmp/crux/volume_key/pkg) Traceback (most recent call last): File "<string>", line 2, in <module> ModuleNotFoundError: No module named 'imp' /usr/lib/python3.10/imp.py in 3.7, not exist in 3.8 "This module is no longer part of the Python standard library. It was removed in Python 3.12 after being deprecated in Python 3.4.. The removal notice includes guidance for migrating code from imp to importlib.. The last version of Python that provided the imp module was Python 3.11." -- https://docs.python.org/3/library/imp.html Needed some extra work on the Pkgfile to use autoreconf. And still footprint due to python3 bump. Changes impl on 3.7 since it can co-exist there. #lynx (opt) In file included from LYMain.c:17: LYMain.c: In function 'main': LYMain.c:1454:24: error: implicit declaration of function 'cuserid' [-Wimplicit-function-declaration] 1454 | STREQ((char *) cuserid((char *) NULL), ANONYMOUS_USER) | ^~~~~~~ Upgrade to lynx 2.9.2-1 fixed it. #xmlto (contrib) xmlif/xmlif.l:46:8: error: type defaults to 'int' in declaration of 'ifsense' [-Wimplicit-int] 46 | static ifsense; /* sense of last `if' or unless seen */ | ^~~~~~~ xmlif/xmlif.l:243:1: error: return type defaults to 'int' [-Wimplicit-int] 243 | { | ^ fix: Upgrade to 0.0.29 #avidemux (df) src/libavcodec/x86/mathops.h:125: Error: operand type mismatch for `shr' gas changed assembler rules again. fix: added patch #ntp (df) configure: error: could not locate pthread_detach() fix: -Wno-error=int-conversion -Wno-error=incompatible-pointer-types first gets it through configure, second gets it through compile #vde2 (df) wirefilter.c:136:13: error: type of 'i' defaults to 'int' [-Wimplicit-int\] fix: -Wno-error=implicit-int #pmount (df) luks.c:147:7: error: implicit declaration of function 'creat' [-Wimplicit-function-declaration] 147 | f = creat( path, 0600); there seems to be some breakage caused by glibc. Suddenly basic functions like strtok_r and creat are not available to old source code. fix: patched to include fcntl.h and unistd.h in luks.h #gconf (df) basic-gconf-app.c:458:60: error: passing argument 1 of 'gtk_dialog_get_content_area' from incompatible pointer type [-Wincompatible-pointer-types] fix: CFLAGS+= -Wno-error=incompatible-pointer-types #alpine (df) osdepbas.c not found...try make clean and new make osdepckp.c not found...try make clean and new make osdeplog.c not found...try make clean and new make Building with SSL and plaintext passwords disabled unless SSL/TLS make[4]: *** [Makefile:955: osdepbas.c] Error 1 fix: change to latest snapshot #libicns (df) png2icns.c:328:25: error: implicit declaration of function 'unlink' [-Wimplicit-function-declaration] fix: patched png2icns to include unistd.h A total of 337 ports were built to build all df repo ports and it's dependencies. -Daryl

1 0

ports/xorg (3.7): [notify] xorg-server: update to version 21.1.16
by crux＠crux.nu 26 Feb '25

26 Feb '25

Commit hash: fb482fe82366ef65623f3054d1fe59010c51218f Author: Juergen Daubert <jue(a)jue.li> Date: Wed Feb 26 10:27:15 2025 UTC+01:00 URL: https://git.crux.nu/ports/xorg/commit/fb482fe82366ef65623f3054d1fe59010c512… Commit message: [notify] xorg-server: update to version 21.1.16 includes several security fixes, see https://lists.x.org/archives/xorg-announce/2025-February/003584.html https://lists.x.org/archives/xorg-announce/2025-February/003585.html

1 0

libbsd/libmd and CRUX 3.8
by Daryl F 14 Feb '25

14 Feb '25

The TODO list for CRUX 3.8 says that libbsd and libmd were committed to the core repo add implied it was due to shadow enabling libbsd by default. Since libmd is required by libmd it also must be put in core. My curiousity (and caution) led me to build CRUX 3.8-rc3 on a qemu kvm. I dislike libbsd because it worms its way into so many autoconf built packages after it is installed. And when libbsd gets updated the host can be difficult or impossible to work with before the revdep. #crux-devel/2022-09-27.log:[15:33:28]: <farkuhar> looks like switching from autotools to cmake prevented the libarchive build from linking against libbsd. I just ran 'prt-get update -fr libarchive' on a system with libbsd installed, and ldd /usr/bin/bsdtar does not report any linking to files owned by libbsd. Hooray for cmake builds. Too bad so many packages still use autoconf. I found https://github.com/shadow-maint/shadow/issues/779 interesting reading. There was some pushback in August of 2023 and the devs added --with-libbsd to deal with it. In autoconf --with-libbsd implies --without-libbsd too. Two of the functions that libbsd added that were once sought after for code safety, strlcpy(3) and strlcat(3), were added to glibc 2.38. The shadow devs decided to copy the BSD readpassphrase(3) code into shadow since shadow uses BSD-3-Clause license and libbsd does too. libbsd and libmd are not required by shadow and libmd is not required by openssh. openssh builds fine without it. shadow builds without it if --without-libbsd is added to ./configure. There is no explanation I can find for the libmd dep in openssh. It appears in https://git.crux.nu/ports/core/commit/30541007d30acf30169ad8c18c6e06a6590e1… with the description "Merge remote-tracking branch 'origin/3.7' into 3.8" and seems inadvertent. To see what effect libbsd and libmd had on a fresh install of 3.8-rc3 I started with only the core packages installed. I used ldd in /bin, /sbin, /usr/bin, and /usr/sbin to see which binaries where already linked to libbsd or libmd. Of course if libbsd is linked by a binary it also links libmd. These core tools will not be available after an update to libmd and/or libbsd and before revdep is run to rebuild their owners: Linked to libbsd in /usr/bin: chage chfn chsh expiry faillog getsubids newgidmap newgrp newuidmap passwd sg su Linked to libmd in /usr/bin: scp sftp ssh ssh-add ssh-agent ssh-keygen ssh-keyscan Linked to libbsd in /bin: login Linked to libbsd in /sbin nologin Linked to libbsd in /usr/sbin chgpasswd groupadd groupdel groupmems groupmod grpck pwck useradd userdel usermod vigr vipw Linked to libmd in /usr/sbin sshd After the install of only core and examining what was using libmd or libbsd I installed all the opt and xorg ports. Of the 99 ISO ports in opt only one used libbsd: talloc (used by samba). Of the 98 ISO ports in xorg only one used libbsd: xorg-server, make these binaries vulnerable: Xorg, Xnest, and Xvfb. By now you can see I *really* don't like libbsd in a Linux system. It belongs in a BSD system. I'm sorry if this generates more work for 3.8 but libbsd with autoconf is like cancer for a Linux system; the sooner it is cut out the easier it is on the patient. And as always: thanks for your work on CRUX. Still the best Linux distro. Regards, Daryl P.S. As near as I can tell libbsd was only updated three times during the lifetime of CRUX 3.7 so future breakage possibilities might be minimal.

1 0

ports/core (3.7): [notify] rsync: updated to version 3.4.0, security fixes
by crux＠crux.nu 15 Jan '25

15 Jan '25

Commit hash: f8a02eeab40bd9d921cdb3d3872132fd924999e3 Author: Juergen Daubert <jue(a)jue.li> Date: Wed Jan 15 02:04:34 2025 UTC+01:00 URL: https://git.crux.nu/ports/core/commit/f8a02eeab40bd9d921cdb3d3872132fd92499… Commit message: [notify] rsync: updated to version 3.4.0, security fixes see https://ftp.samba.org/pub/rsync/NEWS.html

1 0

CRUX 3.8 release candidate/testing ISO available
by Matt Housh 08 Jan '25

08 Jan '25

Greetings, all, There is a test ISO available for the upcoming CRUX version 3.8. It is *NOT* released or finalized yet but please feel free to test it and report any ISO- or setup-specific issues. It can be found here: https://crux.nu/files/rc/crux-3.8-rc3.iso https://crux.nu/files/rc/crux-3.8-rc3.md5 https://crux.nu/files/rc/crux-3.8-rc3.sha256 Regards, Matt

4 6

Humble contribution to the diffusion of Crux Linux
by Entropía Binaria 04 Jan '25

04 Jan '25

Good morning. I hope you are having a great start to this new year. I apologize for not being familiar with mailing lists: that is the reason why I am writing this email in a "direct" way (although I am actually subscribed to the lists). I am a computer science teacher of various levels, from Uruguay, founder of a blog and a YouTube educative channel, both called "Entropía binaria" and which respond to the same thing: emancipation through the use of free (libre) software. Recently, I have made 6 videos about Crux and one about CruxEX, (having contacted Mr. Arne Exton for direct information about his system). All the "Entropía binaria" videos are in Spanish, but I have subtitled the CruxEX video in Swedish and also in English, in order to link it with the Swedish and English-speaking communities, as well as the Spanish-speaking one. At the moment, I am also subtitling the 6 Crux videos, starting with Swedish and ending with English. The idea of these words is to let you know that we are taking this great system seriously, both on its own and as a basis for building other systems, and that we intend to continue this saga in the future. We also want to tell that if you find the content of the videos useful, you can take them and do whatever you want with them, since they are under Creative Commons license on YouTube. The same goes for the entry about Crux and CruxEX on our blog. If you find this material useful enough to reference it on the Crux website, it would be an honor for us. Without obligation, we leave you all the corresponding information and the links in question: Link to the first video of the Crux saga: https://www.youtube.com/watch?v=6TB5B9LDFeA Link to our only video about CruxEX: https://www.youtube.com/watch?v=EYV59WBWXps Blog post for Crux: https://entropiabinaria.blogspot.com/2024/12/guia-de-instalacion-paso-paso-… Blog post for CruxEX: https://entropiabinaria.blogspot.com/2024/12/guia-rapida-de-instalacion-pas… Greetings and thanks! *Hugo Napoli (Entropía binaria)* --

1 0

Heya all 👋
by James Mills 25 Dec '24

25 Dec '24

Some of you on this list may or may not remember me :) How is CRUX doing these days? cheers James -- James Mills / prologic Join Yarn.social today! The only decentralised social media that respects your privacy and freedoms! E: prologic(a)shortcircuit.net.au W: prologic.shortcircuit.net.au Blog: Read my Blog Yarn: @prologic@twtxt.net

3 2

pango 1.54.0 fails to update
by Steve Volumetric 01 Dec '24

01 Dec '24

Hello All, it looks like the link that prt-get tries to grab the pango source from is a dead link. I replaced that link with the one shown here: https://www.linuxfromscratch.org/blfs/view/svn/x/pango.html and now it downloads, (had to ignore signatures too, just for testing) and tries to compile, but I also get a lot of footprint mismatches. I'm guessing that's because it's expecting 1.54 and I'm feeding it 1.55. Just thought I'd mention the dead link.

1 0

An issue with connmand in init
by Wawrzek Niewodniczanski 30 Nov '24

30 Nov '24

Hello, I have a problem with connmand not properly starting during OS init. - The connmand script quietly dies during init. I cannot find anything in system logs. - The same script works perfectly fine, when it's started from interactive shell - I tried various approaches to the script, with start-stop-daemon or without it; trying a loop waiting for IP to be added. Nothing worked. - You can check the current version (minimalistic) below. - I think it used to work, but I cannot remember what changed and make it bad. I appreciate any suggestion. Where should I look? What can I change to enforce logs or any form of feedback. Regards, Wawrzek PS. Current state of the script. ``` #!/bin/sh # # /etc/rc.d/connamnd: start/stop connman daemon # NAME="/usr/sbin/connmand" case $1 in start) $NAME ;; stop) /usr/bin/killall -e $NAME ;; restart) $0 stop sleep 2 $0 start ;; *) echo "usage: $0 [start|stop|restart]" ;; esac -- Dr Wawrzyniec Niewodniczański or Wawrzek for short PhD in Quantum Chemistry & MSc in Molecular Engineering WWW: http://wawrzek.name E-MAIL: jobs(a)wawrzek.name Linux User #177124

1 0

qt6-webengine, qt6-base, and cups interdependency issue
by Daryl F 18 Nov '24

18 Nov '24

Hi all, I use a docker snapshot environment to create accurate footprints for the ports in the df repo. Ports are built in isolation to ensure there is no parasitic interaction in dependencies. This has worked well for several years. It catches missing dependencies and generates footprints that may end up reporting NEW items but never MISSING items. This methodology is broken by this test in the qt6-webengine Pkgfile: if [[ ! -e /usr/lib/qt6/plugins/printsupport/libcupsprintersupport.so ]]; then printf '\e[31m%s\e[m\n' " opt/qt6-base is not build with opt/cups support, and opt/qt6-webeninge depends on it! Please make sure opt/cups is installed, then rebuild opt/qt6-base. After that, you can attempt to build opt/qt6-webengine again. " exit 1 fi Because qt6-base is never built when cups is installed since it is not a dependency listed in qt6-base the build of qt6-webengine always fails. This is not a problem in my desktop environment. I use cups so it is already installed when I build qt6-base. This kind of hidden requirement in qt6-webengine breaks my quality control methods. For now all I can do is maintain my own port of qt6-base with cups listed in the dependencies. However, I suggest that this kind of test in qt6-webengine violates the tried-and-true intent of the dependency list in a Pkgfile in subtle ways and precludes some automation tools. I understand that not everyone that wants qt6-base wants cups but almost everthing that uses qt6-webengine does else the Print function in apps like qutebrowser can print but can only save to disk. I don't think including cups in qt6-base is much of a hit. Anyone using qt6-*anything* is already taking a huge hit in resources. Just because cups is installed with qt6-base no one has to start the cups daemon unless they specifically choose to in the their rc.conf. Again, thanks to the dev team for their hard work providing the best distro around.o Any suggestions? Comments? -Daryl

4 17