Opened 4 years ago

Closed 4 years ago

#12984 closed enhancement (fixed)

elogind-243.4

Reported by: Bruce Dubbs Owned by: Pierre Labastie
Priority: normal Milestone: 9.1
Component: BOOK Version: SVN
Severity: normal Keywords:
Cc:

Description

New version.

Change History (19)

comment:1 by Pierre Labastie, 4 years ago

Owner: changed from blfs-book to Pierre Labastie
Status: newassigned

comment:2 by Pierre Labastie, 4 years ago

his is the first release of the new version 243 series.
We start with 243.4 to be in sync with upstream systemd tagging.

The latest upstream commit this version is synced to is
systemd-stable/v243-stable:70e8c19
Changes and Additions

    Support "reboot to firmware" (loginctl option --firmware-setup)
    Support booting into boot menu (loginctl option --boot-loader-menu)
    Support booting into boot entry (loginctl option --boot-loader-entry)
    loginctl: Show json output if requested even if zero entries
    add SetBrightness() bus call for setting brightness of leds/backlight devices associated with a seat
    consider "greeter" sessions suitable as "display" sessions of a user
    sleep: replace rtc wakealarm with CLOCK_BOOTTIME_ALARM s2h
    busctl: always prefix bus error message with local string
    don't unlink session fifo when exiting elogind, so they survive daemon restarts.
    hibernate: use swaps in priority order

Fixed Bugs

    Break off session creation if session_set_leader() fails
    A few potetial memory leaks have been fixed
    Fix current virtualbox detection
    Fix build with compilers with default stack-protector enabled
    Identify smb3 and ceph as network filesystems
    Fix emission of PropertiesChanged on seats
    Fix emission of PropertiesChanged for users
    cg_path_get_owner_uid: fix parsing UID from session data

comment:3 by Pierre Labastie, 4 years ago

Looks like there are tests for elogind (ninja test). One test fails because there is no /etc/os-release file (for SysV). Maybe we could add one (that is, in LFS have the same instructions for systemd and sysV in 9.1 The End).

comment:4 by Pierre Labastie, 4 years ago

Well, another possibility is just to change the name of the file in the test, because the actual content of the file is irrelevant.

comment:5 by Pierre Labastie, 4 years ago

Resolution: fixed
Status: assignedclosed

Fixed at r22548

[edit: renodr - corrected revision number (change 2nd digit to 2 from 5)]

Last edited 4 years ago by Douglas R. Reno (previous) (diff)

comment:6 by Pierre Labastie, 4 years ago

Resolution: fixed
Status: closedreopened

Upstream have withdrawn the release see https://github.com/elogind/elogind/issues/144

comment:7 by Pierre Labastie, 4 years ago

Living open for now. Hopefully this version will eventually get released... Note that the absence of the NEWS file was not a problem for our build. It must be needed for some doc format, but not man pages.

comment:8 by Pierre Labastie, 4 years ago

Reverted at r22549 for now

comment:9 by Douglas R. Reno, 4 years ago

I have noticed that some DEs are reading /etc/os-release over /etc/lsb-release. I think it should be added to the SysV version for consistency

in reply to:  9 comment:10 by Bruce Dubbs, 4 years ago

Replying to renodr:

I have noticed that some DEs are reading /etc/os-release over /etc/lsb-release. I think it should be added to the SysV version for consistency

That is easy enough, but what values are relevant? Here are the definitions:

https://www.freedesktop.org/software/systemd/man/os-release.html

comment:11 by Douglas R. Reno, 4 years ago

The only values that I think are relevant are:

NAME= VERSION= ID= PRETTY_NAME= VERSION_CODENAME=

That's what GNOME and Plasma will look at if I recall correctly.

Here's an example (from my systemd system):

renodr [ /sources ]$ cat /etc/os-release 
NAME="Linux From Scratch"
VERSION="20190714-systemd"
ID=lfs
PRETTY_NAME="Linux From Scratch 20190714-systemd"
VERSION_CODENAME="Douglas R. Reno"

comment:12 by Pierre Labastie, 4 years ago

Looks like it just amounts to remove the revision="systemd" tags in chapter09/theend.xml. Will do.

Back to the ticket: Upstream hopes to make a release today: see one of the last comments in https://github.com/elogind/elogind/issues/141

in reply to:  12 comment:13 by Pierre Labastie, 4 years ago

Replying to pierre.labastie:

Looks like it just amounts to remove the revision="systemd" tags in chapter09/theend.xml. Will do.

Was slightly more complicated, because of the two different version entities in LFS, but done at LFS revision 11721.

comment:14 by Pierre Labastie, 4 years ago

Hmm, the maintainer was optimistic. In order to not lose all the work I had done for the new version, I have committed the instructions for testing, and some comments for things which seemed to be needed for next version at r22558

comment:15 by Pierre Labastie, 4 years ago

The maintainer has had another issue, so delaying again the release.

comment:16 by Douglas R. Reno, 4 years ago

Ouch, that's a really nasty one... a segfault isn't a nice way to go out if a command line option fails!

comment:17 by Bruce Dubbs, 4 years ago

It looks like upstream has this fixed now. Version number is still 243.4. md5sum:

fb5ba9da1115d68d2b6cdbcd10167b7d elogind-243.4.tar.gz

comment:18 by Pierre Labastie, 4 years ago

New release notes:

Changes and Additions

    Support "reboot to firmware" (loginctl option --firmware-setup)
    loginctl: Show json output if requested even if zero entries
    Add SetBrightness() bus call for setting brightness of leds/backlight devices associated with a seat
    Consider "greeter" sessions suitable as "display" sessions of a user
    sleep: replace rtc wakealarm with CLOCK_BOOTTIME_ALARM s2h
    busctl: always prefix bus error message with local string
    Don't unlink session fifo when exiting elogind, so they survive daemon restarts.
    hibernate: use swaps in priority order
    Remove __cpu_mask testing from test-sizeof.c; elogind does not use any cpu_set functions this is needed for.
    sd_is_mq() now returns -ENOSYS (#146) ; We don't actually know (or care!) if the argument is a message queue.
    Add note about build directory placement in README.md (#145)

Fixed Bugs

    Break off session creation if session_set_leader() fails
    A few potetial memory leaks have been fixed
    Fix current virtualbox detection
    Fix build with compilers with default stack-protector enabled
    Identify smb3 and ceph as network filesystems
    Fix emission of PropertiesChanged on seats
    Fix emission of PropertiesChanged for users
    Add missing includes for umask() in uaccess-command.c (#141)
    Added a better 'broken AF_ALG' detection (#147)

comment:19 by Pierre Labastie, 4 years ago

Resolution: fixed
Status: reopenedclosed

Fixed at r22618

Note: See TracTickets for help on using tickets.