at should basically solve this, at least for the case of GNOME
Software and DNF used in combination, right?
--
Matthew Miller
Fedora Project Leader
___
Yum-devel mailing list
Yum-devel@lists.baseurl.org
http://lists.baseurl.org/mailman/listinfo/yum-devel
ses when 'mark' is needed.
--
Matthew Miller
Fedora Project Leader
___
Yum-devel mailing list
Yum-devel@lists.baseurl.org
http://lists.baseurl.org/mailman/listinfo/yum-devel
ile in %post or make a symlink (ew either way, and also, fragile), or
else rethink the subpackage aspect.
--
Matthew Miller
Fedora Project Leader
___
Yum-devel mailing list
Yum-devel@lists.baseurl.org
http://lists.baseurl.org/mailman/listinfo/yum-devel
GNOME Software is using too.)
--
Matthew Millermat...@mattdm.org <http://mattdm.org/>
Fedora Project Leader mat...@fedoraproject.org <http://fedoraproject.org/>
___
Yum-devel mailing list
Yum-devel@lists
vant.
> > > Were there corresponding reasons for moving it back?
> > Anyone?
> My guess is that it was an accident ... Matthew?
Yes, I think it's by accident. Somewhat relevant from the git log:
commit 2592de5f05c060b906ee8089b841d0fe40f3ed5d
Author: Matthew Miller
Date
At Marko Myllynen's suggestion, remove a redundant rm -f. Also, check
that PIDFILE gets set to _something_.
---
yum-cron/yum-cron.sh | 14 +-
1 files changed, 9 insertions(+), 5 deletions(-)
diff --git a/yum-cron/yum-cron.sh b/yum-cron/yum-cron.sh
index dcf8620..e300fa7 100755
--- a
This now uses noclobber on a file rather than mkdir plus a pid file under
that, to avoid a race condition pointed out by Marko Myllynen.
It also removes the separate TSLOCK/tslock file, since the sleep is now
separated into the cron script -- leaving the main pid lock effectively the
same as the o
---
yum-cron/yum-cron.sh | 18 +-
1 files changed, 9 insertions(+), 9 deletions(-)
diff --git a/yum-cron/yum-cron.sh b/yum-cron/yum-cron.sh
index 2806538..dcf8620 100755
--- a/yum-cron/yum-cron.sh
+++ b/yum-cron/yum-cron.sh
@@ -45,13 +45,13 @@ fi
# This holds the output from
If the script dies or we don't send successfully, leave it around for
possible diagnostics.
---
yum-cron/yum-cron.sh |9 +
1 files changed, 5 insertions(+), 4 deletions(-)
diff --git a/yum-cron/yum-cron.sh b/yum-cron/yum-cron.sh
index c90c6e9..2806538 100755
--- a/yum-cron/yum-cron.sh
The first patch reworks the lock handling to both avoid a race condition and
to simplify things down to one single lock file. Then there's some minor
cleanup.
Marko Myllynen says "Otherwise looking good, feel free to add Acked-by or
Reviewed-by to the patch if you like :-)", but I'm too timid.
On Fri, Jul 29, 2011 at 12:59:41PM -0400, James Antill wrote:
> No need to unless you are desperate to get it upstream ASAP. Shouldn't
> take too long to work out what to do with this.
I'm going to do it so I don't forget later. :)
--
Matthew Miller mat...@matt
This sets MAILTO to root by default, instead of empty. The script still
checks for the presence of the mail program, so it doesn't add an additional
dependency, and basic behavior is the same (since cron sends mail to root by
default anyway), but the meaningful subject line is just _better_.
---
y
This sets MAILTO to root by default, instead of empty. The script still
checks for the presence of the mail program, so it doesn't add an additional
dependency, and basic behavior is the same (since cron sends mail to root by
default anyway), but the meaningful subject line is just _better_.
---
y
On Fri, Jul 29, 2011 at 12:45:36PM -0400, James Antill wrote:
> ...I think it'd be great to call whatever solution we go with from
> yum-cron ... but apart from that I don't think yum-cron should do
> anything special.
Well, that's easy, then. I'll just leave it
pit out to
stdout/stderr by the rpm library, right?
--
Matthew Miller mat...@mattdm.org <http://mattdm.org/>
___
Yum-devel mailing list
Yum-devel@lists.baseurl.org
http://lists.baseurl.org/mailman/listinfo/yum-devel
se you want each report to be.
In the meantime, I'm fine with the default being 0.
--
Matthew Miller mat...@mattdm.org <http://mattdm.org/>
___
Yum-devel mailing list
Yum-devel@lists.baseurl.org
http://lists.baseurl.org/mailman/listinfo/yum-devel
re could be a "yum-cron groupupdate" action which
would read a list of groups from /etc/yum/yum-cron/groupupdate.
--
Matthew Miller mat...@mattdm.org <http://mattdm.org/>
___
Yum-devel mailing list
Yum-devel@lists.baseurl.org
http://lists.baseurl.org/mailman/listinfo/yum-devel
onable (TM) thing a sysadmin might want into the main script, and then
local admins can change the cron config to call the actions they want.
(Groupupdate is an interesting one, and I can see the appeal in having a
simple file to edit rather than trying to set a list of groups in a shell
confi
nd, similarly, plugins like download-updates add command-line options, but
do not appear to add anything to yum shell. How should that be handled?
--
Matthew Miller mat...@mattdm.org <http://mattdm.org/>
___
Yum-devel mailin
de
whether it's errors or warnings.
This is used to generate a helpful subject line and (perhaps more
importantly) to generate meaningful log messages.
This all actually works very well, but makes me wince. Is there a better way
in these modern times?
--
Matthew Miller ma
is a language construct, while [ is a
_command_ (even though a built-in one) with weird syntax to make it _appear_
like a language construct.
--
Matthew Miller mat...@mattdm.org <http://mattdm.org/>
___
Yum-devel ma
No changes to keys or values. File is reordered to group parameters by which
part of the system they affect.
---
yum-cron/yum-cron.sysconfig | 124 ++-
1 files changed, 76 insertions(+), 48 deletions(-)
diff --git a/yum-cron/yum-cron.sysconfig b/yum-cron/y
No code changes -- just comments.
---
yum-cron/yum-cron.sh | 84 +++--
1 files changed, 60 insertions(+), 24 deletions(-)
diff --git a/yum-cron/yum-cron.sh b/yum-cron/yum-cron.sh
index d1a32ec..7238416 100755
--- a/yum-cron/yum-cron.sh
+++ b/yum-cron/
In the unlikely but possible case where LOCKDIR contained files _other_ than
the PIDFILE being tested for, rmdir wouldn't remove it, causing the lock to
persist and then rapidly be tested over and over as the script respawns.
Now uses rm -rf to zap the LOCKDIR (as is done at other points in the co
No code changes. Added description in comments, changed "nightly" to
"scheduled" in echoed statements.
I also removed the "Author" line, out of no disrepect to Seth, but just
'cause it's not in the Fedora spec for sysvinit files.
---
yum-cron/yum-cron.sysvinit | 19 ++-
1 files
No code changes -- everything is behind a "#".
---
yum-cron/yum-cleanup.cron.sh | 11 +++
yum-cron/yum-update.cron.sh | 13 -
2 files changed, 15 insertions(+), 9 deletions(-)
diff --git a/yum-cron/yum-cleanup.cron.sh b/yum-cron/yum-cleanup.cron.sh
index 0842135..e38e80f
This is for consistency, and because the script is bash-specific anyway.
---
yum-cron/yum-cron.sh | 30 ++
1 files changed, 14 insertions(+), 16 deletions(-)
diff --git a/yum-cron/yum-cron.sh b/yum-cron/yum-cron.sh
index 83dc7f6..5ba3be4 100755
--- a/yum-cron/yum-cro
No code changes -- just whitespace.
---
yum-cron/yum-cron.sh | 40
1 files changed, 20 insertions(+), 20 deletions(-)
diff --git a/yum-cron/yum-cron.sh b/yum-cron/yum-cron.sh
index 7238416..83dc7f6 100755
--- a/yum-cron/yum-cron.sh
+++ b/yum-cron/yum-cro
---
yum-cron/yum-cron.sh | 12 ++--
1 files changed, 6 insertions(+), 6 deletions(-)
diff --git a/yum-cron/yum-cron.sh b/yum-cron/yum-cron.sh
index 5ba3be4..de1608e 100755
--- a/yum-cron/yum-cron.sh
+++ b/yum-cron/yum-cron.sh
@@ -10,7 +10,7 @@
# abruptly shutting down mid-transaction.
This sets MAILTO to root by default, instead of empty. The script still
checks for the presence of the mail program, so it doesn't add an additional
dependency, and basic behavior is the same (since cron sends mail to root by
default anyway), but the meaningful subject line is just _better_.
Secon
---
yum-cron/TODO | 19 +++
1 files changed, 19 insertions(+), 0 deletions(-)
create mode 100644 yum-cron/TODO
diff --git a/yum-cron/TODO b/yum-cron/TODO
new file mode 100644
index 000..28e1964
--- /dev/null
+++ b/yum-cron/TODO
@@ -0,0 +1,19 @@
+Logging:
+- Add basic syslog
---
yum-cron/TODO | 19 +++
1 files changed, 19 insertions(+), 0 deletions(-)
create mode 100644 yum-cron/TODO
diff --git a/yum-cron/TODO b/yum-cron/TODO
new file mode 100644
index 000..28e1964
--- /dev/null
+++ b/yum-cron/TODO
@@ -0,0 +1,19 @@
+Logging:
+- Add basic syslog
This is for consistency, and because the script is bash-specific anyway.
---
yum-cron/yum-cron.sh | 30 ++
1 files changed, 14 insertions(+), 16 deletions(-)
diff --git a/yum-cron/yum-cron.sh b/yum-cron/yum-cron.sh
index 83dc7f6..5ba3be4 100755
--- a/yum-cron/yum-cro
In the unlikely but possible case where LOCKDIR contained files _other_ than
the PIDFILE being tested for, rmdir wouldn't remove it, causing the lock to
persist and then rapidly be tested over and over as the script respawns.
Now uses rm -rf to zap the LOCKDIR (as is done at other points in the co
No code changes -- just comments.
---
yum-cron/yum-cron.sh | 84 +++--
1 files changed, 60 insertions(+), 24 deletions(-)
diff --git a/yum-cron/yum-cron.sh b/yum-cron/yum-cron.sh
index d1a32ec..7238416 100755
--- a/yum-cron/yum-cron.sh
+++ b/yum-cron/
---
yum-cron/yum-cron.sh | 12 ++--
1 files changed, 6 insertions(+), 6 deletions(-)
diff --git a/yum-cron/yum-cron.sh b/yum-cron/yum-cron.sh
index 5ba3be4..de1608e 100755
--- a/yum-cron/yum-cron.sh
+++ b/yum-cron/yum-cron.sh
@@ -10,7 +10,7 @@
# abruptly shutting down mid-transaction.
This sets MAILTO to root by default, instead of empty. The script still
checks for the presence of the mail program, so it doesn't add an additional
dependency, and basic behavior is the same (since cron sends mail to root by
default anyway), but the meaningful subject line is just _better_.
Secon
No code changes -- just whitespace.
---
yum-cron/yum-cron.sh | 40
1 files changed, 20 insertions(+), 20 deletions(-)
diff --git a/yum-cron/yum-cron.sh b/yum-cron/yum-cron.sh
index 7238416..83dc7f6 100755
--- a/yum-cron/yum-cron.sh
+++ b/yum-cron/yum-cro
No changes to keys or values. File is reordered to group parameters by which
part of the system they affect.
---
yum-cron/yum-cron.sysconfig | 124 ++-
1 files changed, 76 insertions(+), 48 deletions(-)
diff --git a/yum-cron/yum-cron.sysconfig b/yum-cron/y
No code changes. Added description in comments, changed "nightly" to
"scheduled" in echoed statements.
I also removed the "Author" line, out of no disrepect to Seth, but just
'cause it's not in the Fedora spec for sysvinit files.
---
yum-cron/yum-cron.sysvinit | 19 ++-
1 files
No code changes -- everything is behind a "#".
---
yum-cron/yum-cleanup.cron.sh | 11 +++
yum-cron/yum-update.cron.sh | 13 -
2 files changed, 15 insertions(+), 9 deletions(-)
diff --git a/yum-cron/yum-cleanup.cron.sh b/yum-cron/yum-cleanup.cron.sh
index 0842135..e38e80f
elated to https://bugzilla.redhat.com/show_bug.cgi?id=640918)
--
Matthew Miller mat...@mattdm.org <http://mattdm.org/>
___
Yum-devel mailing list
Yum-devel@lists.baseurl.org
http://lists.baseurl.org/mailman/listinfo/yum-devel
the file into different patches is nice, and I do appreciate
> it ... but the main thing is if you can split the "doesn't do anything"
> patches from the ones that do. That way I can skim 100+ lines of comment
> changes, and not have to think about them.
Yup, will
by going straight through
each file rather than sensibly breaking each change into a separate, sane
commit. I'll try to be more restrained in the future. :)
I can also split the config file comments from the code comments; the former
is really end-user documentation while the latter is jus
inutes. So it may have some accidentally-introduced bugs,
which in any case I promise to smooth out quickly. (It's now running live on
a couple of test systems, so I should find out.) If you want to wait to push
this one I understand. :)
--
Matthew Miller mat...@mattdm.
e yum-cron script. (And I haven't even considered
systemd yet.)
--
Matthew Miller mat...@mattdm.org <http://mattdm.org/>
>From 6292154491fabac4f96a679474d0c81d4ee8f058 Mon Sep 17 00:00:00 2001
From: Matthew Miller
Date: Thu, 28 Jul 2011 16:49:11 -0400
Subject: [PA
reload)
+ restart
+ ;;
+ reload)
+ ;;
+ condrestart)
+ [ -f "$lockfile" ] && restart
+ ;;
+ status)
+ if [ -f $lockfile ]; then
+ echo $"Nightly yum update is enabled."
+ RETVAL=0
+ else
+
On Thu, Jul 28, 2011 at 03:36:51PM -0400, seth vidal wrote:
> > Hmmm; did the patch not come through? First time using git-email
> I was hoping it was one of those fancy 'no change' patches. ;)
No such luck, man. :) Attempted resend happening... now.
--
Matthew M
On Thu, Jul 28, 2011 at 03:00:44PM -0400, Matthew Miller wrote:
> So what's all this then? :)
Hmmm; did the patch not come through? First time using git-email....
--
Matthew Miller mat...@mattdm.org <http://mattdm.org/>
_
On Thu, Jul 28, 2011 at 02:48:50PM -0400, Matthew Miller wrote:
> Previously, the entire yum-cron shell script lived in /etc/cron.daily, using
> options set in /etc/sysconfig/yum-cron to further restrict the days when
> updates and cleanup are done. This moves the main functionality to
50 matches
Mail list logo
