https://www.tomshardware.com/news/linux-removes-486-cpu-support

486 will be removed from the kernel and especially from 6.1 onwards rust will be integrated with the kernel…

+ exec setarch i486 systemd-nspawn -q -D /var/lib/archbuild/staging-i486/root -E PATH=/usr/local/sbin:/usr/local/bin:/usr/bin –register=no –keep-unit –as-pid2 –bind=/var/cache/archbuild32 pacman -Syuu –noconfirm
error: failed to initialize alpm library

This seems to be a problem of either update-archlinux32-package in devops or of the (commented) code path in the PKGBUILD which gets executed by this script.

I tried to install archlinux32 using the archinstall installer, which is shipped with the installation medium.
Installation is working fine until pacstrap, where it will crash because all mirrors will return 404. I checked
my connection and it is working and dns also works. When I viewed the /etc/pacman.d/mirrorlist file, I saw that the script
wrote the default archlinux mirrors for the region specified in the installer to the file and the don’t contain the 32bit packages.

08.12.2017 -  FS#18  - [android-tools] missing a -latomic when linking adb (and maybe other utilities). Didn’t find the place where to report this upstream!?

`/bin/vim` in version 8.2.5046-2.0 has an rpath on `libperl.so` version 5.36, which is currently only available in staging (i686/core is at version 5.34). vim does not start (cannot open shared object file).

```
$ readelf -d /bin/vim

Dynamic section at offset 0×486090 contains 43 entries:

Tag        Type                         Name/Value

[…]
0x000000000000000f (RPATH) Library rpath: [/usr/lib/perl5/5.36/core_perl/CORE]
```

/build/binutils/src/binutils-gdb/libiberty/fibheap.c: In function ‘fibheap_replace_key_data’:
/build/binutils/src/binutils-gdb/libiberty/fibheap.c:38:25: error: ‘LONG_MIN’ undeclared (first use in th
is function)
   38 | #define FIBHEAPKEY_MIN  LONG_MIN
      |                         ^~~~~~~~
/build/binutils/src/binutils-gdb/libiberty/fibheap.c:220:30: note: in expansion of macro ‘FIBHEAPKEY_MIN’
  220 |   if (okey == key && okey != FIBHEAPKEY_MIN)
      |                              ^~~~~~~~~~~~~~
/build/binutils/src/binutils-gdb/libiberty/fibheap.c:36:1: note: ‘LONG_MIN’ is defined in header ‘<limits.h>’; did you forget to ‘#include <limits.h>’?
   35 | #include "fibheap.h"
  +++ |+#include <limits.h>

Amazing code quality

-r--r--r-- 1 root root 0 May 12 07:00 /etc/ca-certificates/extracted/tls-ca-bundle.pem

curl -v https://www.openssl.org
*   Trying 104.89.5.252:443...
* Connected to www.openssl.org (104.89.5.252) port 443 (#0)
* ALPN, offering h2
* ALPN, offering http/1.1
* error setting certificate verify locations:  CAfile: /etc/ssl/certs/ca-certificates.crt CApath: none
* Closing connection 0
curl: (77) error setting certificate verify locations:  CAfile: /etc/ssl/certs/ca-certificates.crt CApath: none

Attached to Project: Archlinux32
Opened by Andreas Baumann - 13.12.2017
Last edited by Erich Eckner - 03.01.2018
 FS#21  - [ceph] unit tests failing or segfault

5/142 Test   #3: test_objectstore_memstore.sh ............***Failed    1.13 sec
7/142 Test   #1: run-rbd-unit-tests.sh ...................***Failed    1.49 sec

46/142 Test #57: unittest_util ………………………*Failed 0.05 sec
51/142 Test #59: unittest_lru ……………………….*Failed 0.57 sec
92/142 Test #100: unittest_erasure_code_shec_arguments ….*Failed 0.63 sec
96/142 Test #102: unittest_journal ……………………*Exception: SegFault 0.90 sec
98/142 Test #106: unittest_mds_sessionfilter …………..*Exception: SegFault 1.02 sec
106/142 Test #116: unittest_bluefs …………………….*Exception: SegFault 0.66 sec
107/142 Test #117: unittest_bluestore_types …………….*Exception: SegFault 1.25 sec
108/142 Test #119: unittest_memstore_clone ……………..*Exception: SegFault 2.51 sec
114/142 Test #124: unittest_osdscrub …………………..*Exception: SegFault 0.63 sec
115/142 Test #125: unittest_pglog ……………………..*Exception: SegFault 1.23 sec
116/142 Test #126: unittest_hitset …………………….*Failed 0.52 sec
125/142 Test #134: test_ceph_argparse.py ……………….*Failed 1.64 sec
130/142 Test #6: run-tox-ceph-disk …………………..*Failed 94.75 sec
132/142 Test #142: unittest_rbd_mirror …………………*Exception: SegFault 0.51 sec
140/142 Test #2: run-cli-tests ………………………*Failed 167.72 sec
141/142 Test #110: mgr-dashboard-smoke.sh ………………*Failed 301.60 sec

Closed by Erich Eckner
03.01.2018 21:15
Reason for closing: Won’t fix
Additional comments about closing:

blacklisted ceph

Comments (9)

Admin
Erich Eckner commented on 17.12.2017 19:11

I’ll build it w/o check(), but we should definitely look into this (later)
Admin
Andreas Baumann commented on 19.12.2017 12:09

ceph goes heavy non-i686:

shell#> cmake ….

Error at cmake/modules/BuildDPDK.cmake:61 (message):

not able to build DPDK support: unsupported target.
"i686-native-linuxapp-gcc" not listed in

Call Stack (most recent call first):

cmake/modules/BuildDPDK.cmake:83 (do_build_dpdk)
cmake/modules/BuildSPDK.cmake:4 (build_dpdk)
CMakeLists.txt:239 (build_spdk)

I suspect libvirt and the other packages can use ceph, but do not require it
really. I would make ceph an optdepend.
Admin
Andreas Baumann commented on 19.12.2017 12:10

Using it without check is no option IMHO.
Admin
Andreas Baumann commented on 03.01.2018 19:29

New try, same bugs:

        1 - run-rbd-unit-tests.sh (Failed)
        2 - run-cli-tests (Failed)
        3 - test_objectstore_memstore.sh (Failed)
       59 - unittest_lru (Failed)
      100 - unittest_erasure_code_shec_arguments (Failed)
      102 - unittest_journal (SEGFAULT)
      106 - unittest_mds_sessionfilter (SEGFAULT)
      116 - unittest_bluefs (SEGFAULT)
      117 - unittest_bluestore_types (SEGFAULT)
      119 - unittest_memstore_clone (SEGFAULT)
      124 - unittest_osdscrub (SEGFAULT)
      125 - unittest_pglog (SEGFAULT)
      126 - unittest_hitset (Failed)
      134 - test_ceph_argparse.py (Failed)
      142 - unittest_rbd_mirror (SEGFAULT)

Decision: remove ceph dependency on libvirt (more?), then blacklist the package.

Reason: we cannot maintain all software for companies upstream.
Admin
Andreas Baumann commented on 03.01.2018 19:45

mmh, a (most likely incomplete) list of software using ceph:

- libvirt: seems to be an optional storage method
- qemu: obvious, if used in combination with libvirt to store the disk image with ceph
- pifpaf: “Suite of tools and fixtures to manage daemons for testing”, ceph used in tests only
- fio: “Scriptable I/O tool for storage benchmarks and drive testing”, ceph seems to be optional

So, all seem to work fine without ceph.
Admin
Andreas Baumann commented on 03.01.2018 19:53

ceph on gihub: a project with pull requests only and no bug reports.. ok then.
Admin
Erich Eckner commented on 03.01.2018 20:59

agreed: I’ll blacklist it, once I compiled a list of ceph dependent packages.
Admin
Erich Eckner commented on 03.01.2018 21:04

I only see libvirt, python-pifpaf and qemu depending on ceph
Admin
Erich Eckner commented on 03.01.2018 21:07

ah, that’s because you already removed the dependencies

Google Cache

Bing Cache

/usr/bin/c++ -D_FILE_OFFSET_BITS=64 -D_GNU_SOURCE -D_LARGEFILE_SOURCE -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS -I/build/clang/src/clang-13.0.1.src/build/unittests/Tooling -I/build/clang/src/clang-13.0.1.src/unittests/Tooling -I/build/clang/src/clang-13.0.1.src/include -I/build/clang/src/clang-13.0.1.src/build/include -I/build/clang/src/llvm-13.0.1.src/utils/unittest/googletest/include -I/build/clang/src/llvm-13.0.1.src/utils/unittest/googlemock/include -march=i486 -mtune=generic -O2 -pipe -fno-plt -fexceptions         -Wp,-D_FORTIFY_SOURCE=2 -Wformat -Werror=format-security         -fstack-clash-protection  -Wp,-D_GLIBCXX_ASSERTIONS -fPIC -fno-semantic-interposition -fvisibility-inlines-hidden -Werror=date-time -Wall -Wextra -Wno-unused-parameter -Wwrite-strings -Wcast-qual -Wno-missing-field-initializers -Wimplicit-fallthrough -Wno-class-memaccess -Wno-redundant-move -Wno-pessimizing-move -Wno-noexcept-type -Wdelete-non-virtual-dtor -Wsuggest-override -Wno-comment -Wmisleading-indentation -fdiagnostics-color -ffunction-sections -fdata-sections -fno-common -Woverloaded-virtual -fno-strict-aliasing -pedantic -Wno-long-long -O3 -DNDEBUG  -fno-exceptions -Wno-suggest-override -std=c++14 -MD -MT unittests/Tooling/CMakeFiles/ToolingTests.dir/RecursiveASTVisitorTests/CallbacksLeaf.cpp.o -MF unittests/Tooling/CMakeFiles/ToolingTests.dir/RecursiveASTVisitorTests/CallbacksLeaf.cpp.o.d -o unittests/Tooling/CMakeFiles/ToolingTests.dir/RecursiveASTVisitorTests/CallbacksLeaf.cpp.o -c /build/clang/src/clang-13.0.1.src/unittests/Tooling/RecursiveASTVisitorTests/CallbacksLeaf.cpp
{standard input}: Assembler messages:
{standard input}: Error: open CFI at the end of file; missing .cfi_endproc directive
c++: fatal error: Killed signal terminated program cc1plus
compilation terminated.

We can ommit the tests..

Also lto is not good (breaks early) and you need lots of memory and swap to build it.

The dart sdk files are still availble for ia32:-
https://storage.googleapis.com/dart-archive/channels/stable/release/2.18.2/sdk/dartsdk-linux-ia32-release.zip

can you send the old PKGBUILD which was used

no task description

# pacman -Syu dino
$ dino
main.vala:25: Fatal error: /usr/lib/libgupnp-1.2.so.1: undefined symbol: gssdp_client_can_reach

/usr/src/debug/emacs-28.1/src/image.c:10009: undefined reference to `rsvg_handle_set_stylesheet’ # /usr/bin/ld: /usr/src/debug/emacs-28.1/src/image.c:10196: undefined reference to `rsvg_handle_set_stylesheet’

I scheduled a rebuild of viewnior, let me know if viewnior 1.6-3.1 works.

Attached to Project: Archlinux32
Opened by Luke Shumaker - 13.06.2018
Last edited by Erich Eckner - 28.06.2018
 FS#41  - [extra/vlc] is uninstallable because [extra/ffmpeg2.8] was removed

[extra/ffmpeg2.8] (which provided ffmpeg=2.8; [extra/ffmpeg] is 3.4) was removed, despite that it was still needed by [extra/vlc].

[extra/vlc] is an old version (2.2); the current version (3.0), which no longer depends on ffmpeg2.8, is currently in [testing].

(I am unsure if [testing/vlc] depends on extra/ffmpeg=3.4 or testing/ffmpeg=4.0)
Closed by Erich Eckner
28.06.2018 20:25
Reason for closing: Fixed

  Comments (3)
  Related Tasks (0/0)

Admin
Tyzoid commented on 13.06.2018 21:39

Checking https://packages.archlinux32.org/extra/i686/vlc/ vs https://packages.archlinux32.org/testing/i686/vlc/, it appears that the extra/ffmpeg satisfies the testing/vlc requirements.

Here’s from vlc in testing:
libavcodec.so.58 (ffmpeg)
libavformat.so.58 (ffmpeg)
libavutil.so.56 (ffmpeg)

vs vlc in extra:
not satisfiable dependency: “libavcodec.so.56” not satisfiable dependency: “libavformat.so.56” not satisfiable dependency: “libavutil.so.54” Admin
Erich Eckner commented on 18.06.2018 09:37

vlc 3.0.3 is now in extra
Admin
Andreas Baumann commented on 21.06.2018 11:36

vlc 3.0.3 works fine on testing.

Google Cache

libswscale/x86/rgb2rgb_template.c: In function ‘interleaveBytes_sse2’:
libswscale/x86/rgb2rgb_template.c:1838:9: error: the register ‘xmm2’ cannot be clobbered in ‘asm’ for the current target
 1838 |         __asm__(
      |         ^

Also on i486 symbols are missing if not using LTO and probing of libraries like x264/x265 fail.

This is pentium4 related problem. Firefox crashes on startup with following messages:

ExceptionHandler::GenerateDump cloned child 814
ExceptionHandler::SendContinueSignalToChild sent continue signal to child
ExceptionHandler::WaitForContinueSignal waiting for continue signal...

Screenshot. Firefox folder after crash.

Temporary solution is to change in /etc/pacman/conf

Architecture=i686

, delete firefox for pentium4 from pacman cache and reinstall firefox

firefox-106.0.3.1.1-i686

This is just the sad state of what people do to working software:

if(CMAKE_SYSTEM_PROCESSOR STREQUAL "i686" AND HAVE_STACKREALIGN_FLAG)
    add_compile_options(-mstackrealign)
endif()

which results in nasm to fail with “unknown option -m”

a) they rewrite half of the build system without apparently understanding fully the autoconf part
b) they don’t bother to delete to autoconf part
c) they write code for porting the software apparenty without testing (why bother then in the first place?)

16.11.2017 - It’s the development version of the FOX toolkit. Breaks in some int/FXival/void * abstractions: FXWSQueue.cpp: In member function ‘FX::FXbool …

This could be bad because our librsvg (the Rust one) lags behind and librsvg-og might not
implement new stuff (for i486).

Attached to Project: Archlinux32
Opened by bill auger - 13.05.2019
Last edited by Andreas Baumann - 02.10.2019
 FS#73  - [gajim] fails to launch when (optional) gupnp-igd is installed

$ gajim
No translations found
Dirs searched: [PosixPath(’/usr/local/share’), PosixPath(’/usr/share’), PosixPath(’/usr/share/gdm’), PosixPath(’/var/lib/menu-xdg’)]

(gajim:1588): dbind-WARNING : 15:14:29.481: Couldn’t connect to accessibility bus: Failed to connect to socket /tmp/dbus-Nv1LTXmb8O: Connection refused (gajim:1588): WARNING **: 15:14:29.534: Failed to load shared library ‘libgupnp-igd-1.0.so.4’ referenced by the typelib: libgupnp-1.2.so.0: cannot open shared object file: No such file or directory
/usr/lib/python3.7/site-packages/gajim/common/app.py:281: Warning: cannot retrieve class for invalid (unclassed) type ‘void’

gupnp_igd = GUPnPIgd.SimpleIgd()

Traceback (most recent call last):

File "/usr/lib/python3.7/site-packages/gajim/application.py", line 185, in _startup
  app.detect_dependencies()
File "/usr/lib/python3.7/site-packages/gajim/common/app.py", line 281, in detect_dependencies
  gupnp_igd = GUPnPIgd.SimpleIgd()

TypeError: could not get a reference to type class
Traceback (most recent call last):

File "/usr/lib/python3.7/site-packages/gajim/application.py", line 220, in _activate
  self.interface = Interface()
File "/usr/lib/python3.7/site-packages/gajim/gui_interface.py", line 2615, in __init__
  cfg_was_read = parser.read()
File "/usr/lib/python3.7/site-packages/gajim/common/optparser.py", line 86, in read
  self.update_config(old_version, new_version)
File "/usr/lib/python3.7/site-packages/gajim/common/optparser.py", line 158, in update_config
  caps_cache.capscache.initialize_from_db()

AttributeError: ‘NoneType’ object has no attribute ‘initialize_from_db’ Traceback (most recent call last):

File "/usr/lib/python3.7/site-packages/gajim/application.py", line 269, in do_shutdown
  app.logger.commit()

AttributeError: ‘NoneType’ object has no attribute ‘commit’

Closed by Andreas Baumann
02.10.2019 19:36
Reason for closing: Fixed
Additional comments about closing:

Seems to work on pentium4/stable. Feel free to reopen if you it crashes
for you..

Google Cache

gcompris-qt 3.0-1.0 is missing a dependency on llvm14 (or maybe it needs a rebuild for llvm) and thus it doesn’t show any game in the menu.

#0  0xb79a0c6d in raise () from /usr/lib/libc.so.6
#1  0xb79892c6 in abort () from /usr/lib/libc.so.6
#2  0xb79e449c in __libc_message () from /usr/lib/libc.so.6
#3  0xb7a6ff4c in __fortify_fail () from /usr/lib/libc.so.6
#4  0xb7a6ff19 in __stack_chk_fail () from /usr/lib/libc.so.6
#5  0xb7182004 in ?? ()
   from /usr/lib/python3.9/site-packages/numpy/core/_multiarray_umath.cpython-39-i386-linux-gnu.so
#6  0xb7179c87 in ?? ()
   from /usr/lib/python3.9/site-packages/numpy/core/_multiarray_umath.cpython-39-i386-linux-gnu.so
#7  0xb6fd589a in PyInit__multiarray_umath ()
   from /usr/lib/python3.9/site-packages/numpy/core/_multiarray_umath.cpython-39-i386-linux-gnu.so
#8  0xb7d07f35 in ?? () from /usr/lib/libpython3.9.so.1.0
#9  0xb7c37301 in ?? () from /usr/lib/libpython3.9.so.1.0
#10 0xb7c46897 in PyObject_Call () from /usr/lib/libpython3.9.so.1.0
#11 0xb7c2c557 in _PyEval_EvalFrameDefault () from /usr/lib/libpython3.9.so.1.0
#12 0xb7c260b9 in ?? () from /usr/lib/libpython3.9.so.1.0
#13 0xb7c36b05 in _PyFunction_Vectorcall () from /usr/lib/libpython3.9.so.1.0
#14 0xb7c2b4e5 in _PyEval_EvalFrameDefault () from /usr/lib/libpython3.9.so.1.0
#15 0xb7c36938 in _PyFunction_Vectorcall () from /usr/lib/libpython3.9.so.1.0
#16 0xb7c2794c in _PyEval_EvalFrameDefault () from /usr/lib/libpython3.9.so.1.0
#17 0xb7c36938 in _PyFunction_Vectorcall () from /usr/lib/libpython3.9.so.1.0
#18 0xb7c2774a in _PyEval_EvalFrameDefault () from /usr/lib/libpython3.9.so.1.0
#19 0xb7c36938 in _PyFunction_Vectorcall () from /usr/lib/libpython3.9.so.1.0
#20 0xb7c2774a in _PyEval_EvalFrameDefault () from /usr/lib/libpython3.9.so.1.0

Builds fine on i686 and pentium4, so I suspect a i486 issue in numpy.

Pacman recently upgraded python from 3.8 to 3.9, which breaks the gdb package:

% gdb
gdb: error while loading shared libraries: libpython3.8.so.1.0: cannot open shared object file: No such file or directory

$ gedit
gedit: error while loading shared libraries: libtepl-5.so.0: cannot open shared object file: No such file or directory

Exist extra/tepl 6.00.0-1.0 in the pacman.
I’m not sure if gedit is old or can only be used with an older version of tepl, which is not possible because i get tepl from 3 days.

Attached to Project: Archlinux32
Opened by Luke Shumaker - 21.11.2017
Last edited by Erich Eckner - 25.11.2017
 FS#14  - [ghc-mod] Needs rebuilt

Yesterday, I mentioned on IRC that many haskell packages need rebuilt. In response, deep42thought moved a bunch of packages from staging to stable, and told me to open a bug report if the issue persisted.

For the most part, this seems resolved. Some pacman -Qo/-Ql/ldd/grep magic tells me that all of the haskell packages I have are fine, except for [ghc-mod].
Closed by Erich Eckner
25.11.2017 14:58
Reason for closing: Fixed
Additional comments about closing:

removed - upstream removed it, too

  Comments (1)
  Related Tasks (0/0)

Luke Shumaker commented on 23.11.2017 04:47

deep42thought has removed ghc-mod, reflecting its removal in Arch. Requesting closure.

Google Cache

Real machines (i686) break apart in mutliple ways: systemd-journald fails to start, also udev seems to have trouble. This leaves one with a rescue console. And at least in my rescue console, my keyboard did not work.
I also have a pentium4 vm which won’t boot currently (I might give some details, when I have a little more time).

26.09.2018 - using openbox (not gnome) $ gnome-terminal # Couldn’t register with accessibility bus: Did not receive a reply. Possible causes include: the …

Installing yay from AUR

git clone https://aur.archlinux.org/yay.git
cd yay
makepkg -sri

gives following error

go: updates to go.sum needed, disabled by -mod=readonly

Reason - outdated version of go in i686 architecture

Temporary solution - use go for pentium4 (change temporarily architecture to pentium4 in /etc/pacman.conf)

*** work/functions/guile.base   Wed Nov 16 17:55:19 2022
--- work/functions/guile.log    Wed Nov 16 17:55:19 2022
***************
*** 1,3 ****
--- 1,7 ----
+ JIT failed to initialize
+ disabling automatic JIT compilation
+ JIT failed due to resource exhaustion
+ disabling automatic JIT compilation

24.09.2018 - The current extra/imagemagick depends on ‘perl<5.27’, but core/perl is 5.28. $ pacman -Si imagemagick perl Repository : extra Name …

/init: line 477: mount: Value too large for defined data type

At least ISO 2022.04.03 is affected..

start leafpad
everything looks fine
type or paste anything and it segfaults with this error in dmesg

leafpad[550]: segfault at 1 ip b77adf13 sp bfd0eaec error 4 in libgobject-2.0.so.0.6400.3[b7787000+2f000]

libarchive has a dependency upon the zstd libzstd with a version recent enough to provide the symbol ZSTD_minCLevel.

PKGBUILD:
depends=(... 'zstd>=1.3.6')

This is an issue when upgrading pacman to support the zstd packaging format.
pacman must depend upon libarchive>=3.3.3 for zstd compression support.
libarchive will then depend upon zstd>=1.3.6 for the symbol ZSTD_minCLevel.

Without these, upgrading an old Arch Linux system gives the incomplete - so unhelpful - error message:

error: could not open file /var/cache/pacman/pkg/something.pkg.tar.zst: Unrecognized archive format

How does pacman recognize an archive format? If the user does not already know the answer, they are in trouble. And, mistakenly circumventing the package compression to force an upgrade of pacman alone will lead to even more trouble, since pacman itself is not the source of the problem.

make and guile made it half (and broken) to stable.
breaking the cycle required a shim package libffi33 in build-support.
packages depending of libffi are rescheduled and need to be rebuilt first, pushing libffi.so.8 to stable now would make
‘make/guile’ work again, but break a lot of other stuff!