, then opensmtpd starts successfully on
boot.
I suspect the "real" issue is that opensmtpd is not designed to be
started before the interfaces in its configuration file exist.
--
Mark Lawrence
ded to prevent is, even after reading the commit.
For me, running
/sbin/logsave -s /tmp/nonexistent.log id
creates the logfile with the content I would expect.
Man logsave(8) only refers to logsave holding output in memory if the parent
directory doesn't exist.
Can you elucidate?
Thanks
Mark
I read the conclusion here as deprecate forcefsck in favour of tune2fs(8) -E
force_fsck.
Suggested patch attached.
Mark
commit 3251825f8070d89fe0872107bc51562ddc7f6292
Author: Mark Hindley
Date: Mon Sep 15 14:02:30 2025 +0100
Deprecate forcefsck.
Closes: #686895
diff --git a
The current version of the patch is attached.
It includes a few stray formatting changes as I have updated init-d-script to
comply with our recent style guidance.
Best wishes
Mark
diff --git a/debian/init-d-script b/debian/init-d-script
index 31ded62f..56aa9b42 100755
--- a/debian/init-d-script
. Now I look back at the email, it was only
sent to me directly, not the list. I am happy to just use /run.
Mark
Control: tags -1 patch
I agree using both an entry in /etc/fstab and /etc/default/tmpfs makes little
sense.
I propose to emphasise that /etc/fstab overrides /etc/default/tmpfs. Patch
attached.
Mark
commit d22a0e1869718a16b9b97621c13c01317c453c1c
Author: Mark Hindley
Date: Mon Sep 15 10:57
This is valgrind upsteam bug https://bugs.kde.org/show_bug.cgi?id=508145
Fix now also committed on the VALGRIND_3_25_BRANCH:
https://sourceware.org/cgit/valgrind/commit/?h=VALGRIND_3_25_BRANCH
The VALGRIND_3_25_BRANCH contains a couple of backports that didn't make
it to the last stable release y
This is valgrind upsteam bug https://bugs.kde.org/show_bug.cgi?id=508145
Fix now also committed on the VALGRIND_3_25_BRANCH:
https://sourceware.org/cgit/valgrind/commit/?h=VALGRIND_3_25_BRANCH
The VALGRIND_3_25_BRANCH contains a couple backports that didn't make
it to the last 3.25.1 release yet.
Hi,
I think this is better to avoid an error when stopping an already stopped
oneshot service.
Mark
commit 5125062c9653b9a3be0462f889d62b9ef929d05a
Author: Mark Hindley
Date: Mon Sep 15 08:19:19 2025 +0100
Avoid error when stopping already stopped oneshot which is no-op.
diff --git a
Control: tags -1 patch
I have a proposed generic framework to support mounting additional kernel
filesystems. I have added support for debugfs and cgroup2 but omitted cgroup.
None are enabled by default.
Comments welcome. Thanks.
Mark
diff --git a/debian/src/initscripts/etc/default/rcS
b
If the diff is easier, see below. But I have whitespace changes which might make
it a fiddly merge.
Mark
diff --git a/debian/init-d-script b/debian/init-d-script
index 31ded62f..bd4f2b04 100755
--- a/debian/init-d-script
+++ b/debian/init-d-script
@@ -39,6 +39,29 @@ call() {
fi
Control: tags -1 patch
Andrew and Lorenzo,
Thanks for this.
Attached is an initial implementation. I decided hooking into init-d-script at
the correct place was better and it sitll leaves the do_*_override functions for
the sysadmin, if they require.
Mark
#!/bin/sh
# See init-d-script(5) for
Control: tags -1 patch
Jesse,
Does something like this help?
Mark
commit 059cdd2838a912abe9e915601d44be99d4913879
Author: Mark Hindley
Date: Mon Aug 25 14:39:10 2025 +0100
Use stdbool.h, fixes FTBFS with gcc-15.
Closes: #1097938
diff --git a/startpar.c b/startpar.c
index
Control: tags -1 patch
Jesse,
Would something like this work?
Mark
commit e16d596a0b53256f89193063ee980c13c81f9368
Author: Mark Hindley
Date: Mon Aug 25 14:51:17 2025 +0100
Use stdbool.h, fixes FTBFS with gcc-15.
Closes: #1096840
diff --git a/insserv.c b/insserv.c
index
es because it is ignoring the errors.
But the warning/error is also in that build log.
Cheers,
Mark
Package: bugs.debian.org
Severity: important
I just upgraded using apt to trixie and now I can only log into my system if I
select the rescue boot in grub and then wait for the wireless network to come
up.
If I just boot without any intervention it sits on the screen that appears
after Grub. Swi
Package: src:gdb
Version: 16.3-4
debugedit testsuite on hurd-amd64 shows:
gdb: warning: A handler for the OS ABI "GNU/Hurd" is not built into this
configuration of GDB. Attempting to continue with the default i386 settings.
https://buildd.debian.org/status/fetch.php?pkg=debugedit&arch=hurd-amd6
On Thu, 21 Aug 2025 03:24:15 -0300 Leandro Cunha
wrote:
> Hi,
>
> Can you get version 0.29 of experimental and test it so we can see if
> the problem still occurs?
>
> Download of
>
https://ftp.debian.org/debian/pool/main/c/celluloid/celluloid_0.29-1_amd64.deb
>
> After: dpkg -i celluloid_0.29-
o-s-s triggers on the systemd unit but
kexec-tools has no unit.
I wonder if bin:initscripts might be a better option.
Matthew, what do you think?
But I hope this is moot and the kexec-tools maintainer is more enlightened.
Mark
Hi,
A similar request was made to Devuan a while ago[1].
My reading is that an initscript is no longer required. Based on
the changelog entry[2], the non-systemd way of doing this is now
kexec-load-kernel; reboot
Mark
[1] https://bugs.devuan.org/871
[2]
https://sources.debian.org/src
ripts/, but the actual script is in /etc/openvpn/.
I think it would be a good idea to patch these documentation files to reflect
the correct location.
Thanks,
Mark
$ aptitude search askpass
p ksshaskpass
- interactively prompt users for a passphrase for
ssh-add
i A lxqt-openssh-askpass
- OpenSSH user/password GUI dialog for LXQt
i A lxqt-openssh-askpass-l10n
- Language package for
If more reason is needed for switching the default, this is from a
freshly deployed server in Dubai:
root@ae-dubai:~# speedtest
Retrieving speedtest.net configuration...
Cannot retrieve speedtest configuration
ERROR: HTTP Error 403: Forbidden
root@ae-dubai:~# speedtest --secure
Retrieving spee
Package: speedtest-cli
Version: 2.1.3-2
Severity: important
When using the --secure flag, servers with closer proximity are used
and other connection problems seem not to happen.
-- System Information:
Debian Release: 12.11
APT prefers stable-updates
APT policy: (500, 'stable-updates'), (500,
le dependency files.
If I have understood what you are trying to do, I think you need to specify
--insserv-dir as well.
Mark
t, the file could
be moved under /var/lib/?
Thanks for your time with this.
Mark
Paul,
Thanks for this.
On Sat, Jul 19, 2025 at 10:58:02AM +0200, Paul Gevers wrote:
> On Sat, 12 Jul 2025 15:38:41 +0100 Mark Hindley wrote:
> > + # Remove any timestamp to force regeneration of all scripts.
> > + rm -f /var/tmp/${DPKG_MAINTSCRIPT_PACKAGE}.stamp
>
ha's reproducer in #1108944)
Thanks. I suggest the following patch. It fixes the mmdebstrap reroducer for
me.
Mark
commit 27a1dee33fef7659c121eb7da3f180988351f5ed
Author: Mark Hindley
Date: Sat Jul 12 12:54:21 2025 +0100
utrans-rc.postinst: use manual trigger to regenerate scripts so it h
y
utrans-rc producing /etc/init.d/exim4 from /usr/lib/systemd/system/exim4.service
before exim4-base is installed. But exim4-daemon-light depends on exim4-base, so
I don't currently see how that can happen.
Do you have a reliable reproducer?
Thanks
Mark
Thanks. I am away until Thursday, but will look then.
Maybe the postinst should use the trigger so it happens after all
packages are installed and configured.
Mark
headache of fitting a fast-moving program like
Kodi into the Debian release system, but is there any way of this patch to
kodi-inputstream-adaptive making it into Trixie?
Thanks,
Mark
-- System Information:
Debian Release: 13.0
APT prefers testing-security
APT policy: (500, 'testing-sec
unusable after making a USSD request and the fix is
simple, it would be worth including in trixie.
Thanks
Mark
[1] https://gitlab.gnome.org/GNOME/calls/-/issues/699
[2] https://gitlab.gnome.org/GNOME/calls/-/merge_requests/776
your software tends to be excellent) then please let me know or send
me a link to any useful info as ChatGPT can give less than perfect info
on this.
Thanks and Best Regards,
Mark
On Thu, 19 Jun 2025 08:21:22 +0200 Salvatore Bonaccorso wrote:
> Control: reassign -1 src:linux 6.12.3
Package: linux-image-6.12.32-amd64
Version: 6.12.32-1
Severity: normal
X-Debbugs-Cc: markasmith...@gmail.com
Dear Maintainer,
*** Reporter, please consider answering these questions, where appropriate ***
* What led up to the situation?
* What exactly did you do (or not do) that was eff
chime in).
utrans[1] can produce cron or xinetd fragments from timer units. It can also
produce LSB initscripts or openrc runscripts from service and socket units.
[1] https://tracker.debian.org/utrans
Mark
On Tue, Mar 11, 2025 at 08:06:52AM +, Sid T wrote:
> > When I select "Power off" from the power menu at the top right of the
> > screen if there any pending package updates then instead of shutting
> > down GNOME will reboot into a single user mode, apply the pending
> > package updates and the
Okay, we see that patch was applied upstream resolving this. Thanks much!
Best -
--
Mark
smime.p7s
Description: S/MIME cryptographic signature
ou will need to reboot with init=/bin/sh or
similar first.
[1]
https://salsa.debian.org/systemd-team/systemd/-/raw/debian/master/debian/systemd.prerm?ref_type=heads
Mark
Package: yamagi-quake2
Version: 8.20+ctf1.09+dfsg-1
Severity: normal
X-Debbugs-Cc: mcorn...@mcornick.com
Dear Maintainer,
When running yamagi-quake2 version 8.50+dfsg-1 on trixie, with the full
Quake 2 data set installed with game-data-packager, the Q2DMx set of
maps (maps intended for deathmatch
for confirming so quickly!
--
Mark
smime.p7s
Description: S/MIME cryptographic signature
Package: passwd
Version: 1:4.17.4-1
Severity: normal
Tags: patch
Dear Maintainer,
In useradd(8), we have the following:
-d, --home-dir HOME_DIR
The new user will be created using HOME_DIR as the value for the user's
login directory. The default is to append the LOGIN name to BASE_DIR
r option -M, there is already some text about CREATE_HOME.
> I guess it was an overseen that -d would also need some text.
>
Yes, I'm guessing same.
Best -
--
Mark
smime.p7s
Description: S/MIME cryptographic signature
Here's the patch
--- useradd.8.old 2025-04-18 14:16:25.705328604 +
+++ useradd.8 2025-04-18 15:12:17.842060165 +
@@ -93,11 +93,7 @@
\fILOGIN\fR
name to
\fIBASE_DIR\fR
-and use that as the login directory name\&. If the directory
-\fIHOME_DIR\fR
-does not exist, then it will be create
From: "mark.yang"
* fix build failure with GCC-15
netcat.c:1443:41: error: passing argument 2 of 'signal' from incompatible
pointer type [-Wincompatible-pointer-types]
1443 | signal(SIGALRM, quit);
| ^~~~
|
Package: chirp
Version: 1:20250221-1
Severity: important
X-Debbugs-Cc: mcorn...@mcornick.com
Dear Maintainer,
On trixie, the current version of chirpw crashes on launch with an
ImportError that appears related to wxWidgets:
mcornick@x230:~$ chirpw
Traceback (most recent call last):
File "/usr/
bian.org/debian/ucf or just attach a patch here.
Thanks
Mark
Control: tags -1 patch
The specific src:snmp issue was addressed in #932775.
I think the best solution is to document that /etc/default/$NAME variables are
not available at top-level.
Mark
commit eff6c194b8afabc3c5ba449c0399b9301fa7736f
Author: Mark Hindley
Date: Tue Mar 18 10:55:24 2025
Control: tags -1 pending
Carles,
On Mon, Mar 31, 2025 at 08:18:41AM +0100, Mark Hindley wrote:
> This is no longer the correct salsa repo. Please use
> https://salsa.debian.org/debian/ucf or just attach a patch here.
Actually, I have downloaded your patch from the MR above, but I don&
stemd-sysv and its
>dependencies depends on systemd-sysv.
Yes. But why do you want to install sysvinit-core *and* systemd-sysv?
What are you trying to do/achieve?
Mark
e to testing.
Yes, reassigning.
Mark
in kmod, in this case /etc/init.d/kmod.
The kmod tools (modprobe etc) are still in bin:kmod.
> kmod version in trixie is 34-3 (lower), so it is just getting removed.
I suspect you are seeing the same issue here as #1100955, in particular
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1100955#13
Reassigning to kmod.
Mark
tive)?
>
> ~#apt-get dist-upgrade
Don't try to force it because it has the undesired effects you observe.
Once a suitable version of kmod migrates, I think this will resolve. But since
the kmod autopkgtest regressions are causing the failure to migrate, I will
reassign.
Mark
[1] h
le side-effects of this fix in other system
> configurations.
I am not certain either. I suggest we postpone the change until early in the
forky cycle.
Mark
)
Thanks
Mark
Control: forcemerge 922423 -1
Cristian,
This patch no longer applies.
The whitespace and style cleanups will be addressed as part of #922423. So I
will merge with that.
Thanks
Mark
-d-script based for a while.
This question is really about the scripts packaged in bin:initscripts. I am not
taking on rewriting them all for init-d-script!
Mark
Package: src:linux
Version: 6.12.12-1
Severity: important
X-Debbugs-Cc: mpearson-len...@squebb.ca
Dear Maintainer,
Kernel config PINCTRL_INTEL_PLATFORM is needed for touchpads to work on
Lunarlake and newer platforms.
Confirmed and tested on Lenovo X1 Carbon G13 LNL
-- Package-specific info:
I am going to suggest a few more.
- avoid obsolete x prefixes such as [ x"$VARIABLE" = xvalue ]
- avoid [ .. -a .. ] and [ .. -o .. ] in POSIX scripts (undefined)
Mark
wline between conditional test and 'then'
- no space before redirection operators
In addition:-
- prefer '[ ]' to test(1) or builtin
Comments? Changes?
Mark
Control: reassign -1 partconf
Control: forcemerge 220282 -1
On Mon, Jan 09, 2023 at 06:22:25PM +0100, Thorsten Glaser wrote:
> On Mon, 9 Jan 2023, Mark Hindley wrote:
>
> >Yes, That is the debootstrap one. However on a new installation, partconf
> >then
> >fills in t
m2 > cryptdisks > checkfs > mount
>
> My configuration would require
>
> cryptdisks-early > lvm2 > checkroot > mountroot > cryptdisks > lvm2 >
> checkfs > mount
If so, I propose to reassign to cryptsetup as that package is responsible for
the ordering/misordering you have indentified.
Mark
0 0
>
> maybe one of those default options require portmap?
Yes, very possibly.
Reading nfs(5) I think it might be mountport and/or mountproto.
On your system, where are those options configured? /etc/nfs.conf? /etc/exports
on the server?
Are you able to test or confirm?
Mark
Package: ejabberd
Version: 24.12-2
Severity: important
My apparmor service was breaking just now after an unstable
upgrade due to ejabberd it seems:
---
# systemctl status apparmor.service
× apparmor.service - Load AppArmor profiles
Loaded: loaded (/usr/lib/systemd/system/apparmor.ser
On Wed, Mar 12, 2025 at 01:41:41PM +, Mark Hindley wrote:
> Yes, at the moment "|| is redundant.
Sorry, that got garbled
Yes, at the moment "|| true" is redundant
Mark
On Wed, Mar 12, 2025 at 02:23:06PM +0100, Marco d'Itri wrote:
> On Mar 12, Mark Hindley wrote:
>
> > + update-rc.d kmod remove || true
> Why the "|| true"?
>
> update-rc.d does not fail even the symlinks have already been removed.
Yes, at the moment &quo
Control: tags -1 patch
Marco,
On Tue, Mar 11, 2025 at 10:39:48AM +0100, Marco d'Itri wrote:
> so if you want this change then send a tested patch.
This patch works for me and is similar in functionality to udev.postinst.
Mark
commit 070ef4aba05e3c85bd28a6e726223002c4a34b7e
Auth
On Tue, Mar 11, 2025 at 07:53:36PM +0100, Paul Gevers wrote:
> If the upload happens in the near future, yes.
Many thanks for the quick response. Uploaded.
Mark
Thanks very much
Mark
On Mon, Mar 10, 2025 at 05:15:12PM +0100, Marco d'Itri wrote:
> On Mar 10, Mark Hindley wrote:
>
> > Would you please revert this change? To remove the initscript at this stage
> > in
> No, I will not dedicate any more resources to sysvinit support.
It isn't jus
On Mon, Mar 10, 2025 at 05:36:29PM +0100, Marco d'Itri wrote:
> On Mar 10, Mark Hindley wrote:
>
> > At the moment sysadmin changes to /etc/init.d/kmod will be lost because
> > kmod.maintscript has rm_conffile /etc/init.d/kmod 34-1~
> >
> > That would need to
ed in https://bugs.debian.org/1100072. Kmod shouldn't be using rm
on the symlinks. At least use update-rc.d and check if /etc/init.d/kmod doesn't
still exist (i.e hasn't been taken over). If you are unsure, you might want to
look at Michael Biebl's way of dealing with this situation in udev.postinst.
Mark
Thanks for your consideration.
Mark
dpkg-source: warning: extracting unsigned source package
(/home/mark/src/devuan/build/sysvinit_3.14-4.dsc)
diff -Nru sysvinit-3.14/debian/changelog sysvinit-3.14/debian/changelog
--- sysvinit-3.14/debian/changelog 2025-02-24 09:27:01.0 +
/kmod the kmod postinst will remove it and the symlinks regardless.
To avoid further spread, the broken version in sid should not migrate to
trixie. Hence, raising to RC severity.
Mark
[1] https://bugs.debian.org/1100072
https://salsa.debian.org/md/kmod/-/commit/c2ae42ca3a338fba0f190e3a61a635d386b1e8fa
This will need resolving in src:kmod unless we rename the initscript. But that
will lose sysadmin modifications.
Mark
[1] https://bugs.debian.org/1100035
there is a reliable way to
> > identify a Devuan system in maintainerscript, so that I could make
> > the copying of the logrotate file a no-op in Devuan systems..
The obvious way to identify a Devuan system is testing if the file
/etc/devuan_version exists.
HTH
Mark
Thorsten,
On Wed, Mar 05, 2025 at 08:38:12AM +, Mark Hindley wrote:
> This is the only outstanding request I see here. Reassigning to initscripts.
> But
> this is not actionable so close to freeze.
Actually, I wonder if it would ever be wise at all: mounting of network
filesystems
ing to initscripts. But
this is not actionable so close to freeze.
Mark
we already have X-Interactive, which, strictly speaking,
> is not in basis (it can be expressed in terms of Should-Start and
> X-Start-Before).
>
> Jess, how much complexity would it be to implement X-Exclusive (or reusing
> X-Interactive for stop sequence)?
What are your thoughts here?
Mark
Control: tags -1 upstream
Jesse,
I have just tried this with insserv 1.26.0 and the self-referential virtual
facility appears to be silently ignored. Is that the intended behaviour? If so,
is it documented anywhere (I can't immediately see it).
Thanks
Mark
used in dependency-based boot. So I will reassign and close.
If you feel the documentation is still not clear, feel free to reopen.
Thanks
Mark
[1]
https://sources.debian.org/src/sysvinit/3.14-3/debian/src/sysv-rc/doc/README.Debian/
erfaces.
If you have allow-hotplug, then ifup is asynchronous and the networking
initscripts can't know when the interface is actually ready.
If you want synchronous behaviour, use auto rather than hotplug.
Is this still an issue for you?
Thanks
Mark
Thorsten,
I don't see the suggestion having been made of removing the executable bit from
the initscript? It should ensure the script is never run and the K and S
symlinks are removed.
Does that achieve what you wanted?
Best wishes
Mark
but they are neither in the log file
> "/var/log/boot" or syslog.
>
> Below the contents of the file "/var/log/boot",
> from what I see on the screen there are missing
> messages from udev and other rc scripts.
I wonder if the missing messages are on stderr and are being missed by
bootlogd.
Jesse,
Do you have any thoughts here?
Thanks
Mark
Control: tags -1 patch forky
I suggest the following patch for forky-cycle.
Mark
commit 2de6da195d49da7d83d93da1dd6fddf5562c764b
Author: Mark Hindley
Date: Tue Feb 25 10:11:49 2025 +
Update inittab on new installations to use inittab(5) consistent id names.
Closes: #538257
es. I am less certain that changing serial lines and modems is a good
idea.
Also, only changing linux; patches for non-linux are welcome.
Updated patch for forky below.
Mark
commit d25505a79f52948ea8c309ca5fa4c26bb8eb4a15
Author: Mark Hindley
Date: Tue Feb 25 09:59:12 2025 +
Update ini
this is a usage that would be good to support.
> I tried to just implement what Hristo proposed and didn't find it too much
> hacky:
I am testing your patch. Thanks.
Could you provide a testsuite case as well?
Mark
ke sense to be
> installed in parallel.
I agree. Proposed patch attached.
Mark
commit ec0ca80027fac26041e0079e0dfec2a485604b68
Author: Mark Hindley
Date: Sun Feb 23 12:28:01 2025 +
d/control: bootlogd Conflicts plymouth; they both log to /var/log/boot.log.
Closes: #669017
di
Control: reassign -1 insserv
Control: done -1 1.14.0-3
These messages came from update-rc.d-insserv which was removed in version
1.14.0-3.
Reassigning and closing.
Mark
Control: tags -1 patch
Michael,
Thanks.
I suppose we can only print the warning message if VERBOSE mode.
Is the behaviour better with this patch?
Mark
commit 0ef23f5b322b83f9be8cc466009c3ca99039025f
Author: Mark Hindley
Date: Fri Feb 21 16:21:55 2025 +
if-up.d/mountnfs: only
I am no lxc expert, but there is a (fairly recent) suggested fix here[1].
I found this line in /etc/fstab in the container’s rootfs: devpts /dev/pts
devpts nosuid,noexec,gid=tty,mode=620 0 0 After I added ptmxmode=666 to mount
options, /dev/ptmx became 666.
Does that help?
Mark
[1
d I propose to downgrade this to wishlist. Using the IP address in
/etc/fstab is probably better IMHO.
Mark
0 0
Are you using an IP address here?
Does /etc/network/interfaces use auto or allow-hotplug?
Mark
Control: tags -1 moreinfo
Patrik,
I know you submitted this a long time ago. Apologies for the unreasonable delay
in replying.
Is this still an issue? Do you have auto or allow-hotplug for the problematic
interfaces in /etc/network/interfaces?
Mark
in time.
Or am I missing something? Is this still an unresolved issue?
Thanks
Mark
Control: tags -1 moreinfo
Vincent,
Thanks for this and apologies for the extreme delay
I am unclear what you are really wanting here. How is if-up.d/mountnfs supposed
to determine which auto interfaces are important and which it can ignore?
Or have I misunderstood?
Mark
Dirk,
On Wed, Feb 19, 2025 at 12:13:00PM +, Mark Hindley wrote:
> > You could create the folder on the fly in the run script, see patch below.
Actually that doesn't always work either. I think diff --unidirectional-new-file
is a better solution.
M
script, see patch below.
Yes, I prefer that to spraying .gitignore all over the place.
Thanks.
Does the testsuite pass for you then?
Mark
ng.
I can't reproduce this. How are you running the tests or the build? What is your
build environment?
Thanks
Mark
r, the
detail may now be lost in the mists of time, in which case I will probably close
this.
Thanks
Mark
s DEB_BUILD_OPTIONS="nocheck
> noautodbgsym noddebs"
Also removed Makefile check from all recipe.
Does these patches help?
Mark
commit 6376a97644b791b2e66fe7f2afdd87f4df54a3bb
Author: Mark Hindley
Date: Wed Feb 19 09:12:58 2025 +
Makefile: don't run check in all recipe.
d
1 - 100 of 2774 matches
Mail list logo
