[systemd-devel] Shedding some legacy naming: syslog "priority"

2013-01-31 Thread David Strauss
Before I spend time rolling a patch, what are the thoughts on renaming the "priority" arguments to something like "verbosity"? This change would not alter the function signature, only the naming and documentation. We would continue to use syslog's constants. "Priority" is confusing while developin

[systemd-devel] journal use case -- advice requested

2013-01-31 Thread David Strauss
We're using a wrapper for sd_journal_print() that adds a few extra items before the message: the thread ID and the build ID. This is okay for now, but I'd really like to get advice on how to inject those as proper fields without adding them manually to every logging call. I'd also like to maintain

Re: [systemd-devel] Feature request: schedule jobs for last day of month

2013-01-31 Thread David Strauss
On Wed, Jan 30, 2013 at 9:47 PM, Andrey Borzenkov wrote: > That's pretty common for running backups, performing accounting tasks > etc. I'm not too opposed, but maybe we should consider matching the feature set of an established standard like iCal repeating events? It'd be kind of neat to get an

Re: [systemd-devel] build failures in latest master from d848b9cbfa0ba72381363accce481600169df2eb

2013-01-31 Thread Dave Reisner
On Thu, Jan 31, 2013 at 07:21:27PM -0500, Colin Walters wrote: > On Thu, 2013-01-31 at 13:09 -0500, Dave Reisner wrote: > > > http://pkgbuild.com/~dreisner/build.log > > So I distilled a few relevant linker lines to compare, but basically the > only difference is -O0 vs -O2, and the gnome-ostree

Re: [systemd-devel] build failures in latest master from d848b9cbfa0ba72381363accce481600169df2eb

2013-01-31 Thread Colin Walters
On Thu, 2013-01-31 at 13:09 -0500, Dave Reisner wrote: > http://pkgbuild.com/~dreisner/build.log So I distilled a few relevant linker lines to compare, but basically the only difference is -O0 vs -O2, and the gnome-ostree linker lines have -L/usr/lib before -lgpg-error -ldl -lrt. Nothing that sh

Re: [systemd-devel] Bad handling of "/" in instance names?

2013-01-31 Thread Lars Kellogg-Stedman
> I'm assuming this happens because systemd is trying to create a > symlink > with a "/" in the name and is falling over. I can think of a number > of > ways of dealing with this: So thanks to grawity on #systemd I took a closer look at the systemd.unit man page, where I discovered that the '/' -

Re: [systemd-devel] [Patch] NumLock setting from vconsole.conf

2013-01-31 Thread Matthias Berndt
Hi, The attached version of the patch contains updates to the man pages. Cheers Matthias diff --git a/man/vconsole.conf.xml b/man/vconsole.conf.xml index 45156b7..3707279 100644 --- a/man/vconsole.conf.xml +++ b/man/vconsole.conf.xml @@ -74,7 +74,10 @@ vconsole.keymap.toggle=,

[systemd-devel] Bad handling of "/" in instance names?

2013-01-31 Thread Lars Kellogg-Stedman
I have a templated service for which each instance needs to receive arguments of the form "foo/bar". This works fine when starting the service: # systemctl start myservice@foo/bar.service But trying to *enable* this service results in an error: # systemctl enable myservice@foo/bar Failed

Re: [systemd-devel] build failures in latest master from d848b9cbfa0ba72381363accce481600169df2eb

2013-01-31 Thread Dave Reisner
On Thu, Jan 31, 2013 at 12:55:23PM -0500, Colin Walters wrote: > On Thu, 2013-01-31 at 10:52 -0500, Dave Reisner wrote: > > > You previously added libsystemd-daemon as a dependency of > > libsystemd_shared_la as well, which means that libudev links to > > libsystemd-daemon. I suspect you were alon

Re: [systemd-devel] [Patch] NumLock setting from vconsole.conf

2013-01-31 Thread Tomasz Torcz
On Thu, Jan 31, 2013 at 01:26:58PM +0100, Matthias Berndt wrote: > Hi, > > It has bothered me for some time that vconsole.conf doesn't offer a setting > to > boot with NumLock enabled. With the patch in the attachment you can put > NUM_LOCK=y in vconsole.conf (or vconsole.num_lock=y on the kern

Re: [systemd-devel] build failures in latest master from d848b9cbfa0ba72381363accce481600169df2eb

2013-01-31 Thread Colin Walters
On Thu, 2013-01-31 at 10:52 -0500, Dave Reisner wrote: > You previously added libsystemd-daemon as a dependency of > libsystemd_shared_la as well, which means that libudev links to > libsystemd-daemon. I suspect you were alone, or at least the vocal > minority, on this. The commit message is pret

[systemd-devel] [Patch] NumLock setting from vconsole.conf

2013-01-31 Thread Matthias Berndt
Hi, It has bothered me for some time that vconsole.conf doesn't offer a setting to boot with NumLock enabled. With the patch in the attachment you can put NUM_LOCK=y in vconsole.conf (or vconsole.num_lock=y on the kernel command line) and it'll do the right thing. SCROLL_LOCK, CAPS_LOCK, vcons

[systemd-devel] Random "busy" while renaming net interface on KVM

2013-01-31 Thread Gerardo Exequiel Pozzi
Hello During testing Arch Linux ISO under KVM (1.2.0) (using systemd 197), looks like when rename rules must be applied network interface is still busy. You can see on the journal, when things goes bad: Jan 31 16:27:53 archiso kernel: pci :00:03.0: [10ec:8139] type 00 class 0x02 Jan 31

Re: [systemd-devel] build failures in latest master from d848b9cbfa0ba72381363accce481600169df2eb

2013-01-31 Thread Dave Reisner
On Thu, Jan 31, 2013 at 10:29:40AM -0500, Colin Walters wrote: > Ok, here's a first patch that helps my build: > > > From 4908cae79b86b1e427bb2001784f51c93148a4a1 Mon Sep 17 00:00:00 2001 > From: Colin Walters > Date: Thu, 31 Jan 2013 10:28:35 -0500 > Subject: [PATCH] build-sys: Link libudev to

Re: [systemd-devel] build failures in latest master from d848b9cbfa0ba72381363accce481600169df2eb

2013-01-31 Thread Colin Walters
Ok, here's a first patch that helps my build: >From 4908cae79b86b1e427bb2001784f51c93148a4a1 Mon Sep 17 00:00:00 2001 From: Colin Walters Date: Thu, 31 Jan 2013 10:28:35 -0500 Subject: [PATCH] build-sys: Link libudev to libsystemd-id128-internal.la It needs it, and not having it breaks my build

Re: [systemd-devel] tests

2013-01-31 Thread Zbigniew Jędrzejewski-Szmek
On Thu, Jan 31, 2013 at 08:35:56AM +0100, Thomas H.P. Andersen wrote: > Hi Zbigniew, Hi, > _cleanup_free_ only cleans up when the variable goes out of scope > right? So just using it on "r" here would leak all but the last use of > it? Yes. > Using individual variables for each test case like r1

Re: [systemd-devel] [PATCH] util: *DO NOT* loop for EINTR handling with close_nointr()

2013-01-31 Thread Colin Walters
On Wed, 2013-01-30 at 11:44 -0300, Cristian Rodríguez wrote: > El 25/01/13 14:27, Colin Walters escribió: > > Lightly tested in gnome-ostree qemu - review carefully please! > > > > See the linked references for why we should not do this. > > > > Signed-off-by: Colin Walters > > --- > > src/share