2010/1/11 Ohad Levy ohadl...@gmail.com:
+1
is there a reason why not to strip all facts?
I'm a little cautious about doing that in general in case someone has
a dependancy on a fact with a space in or valid but odd setups
(directories with trailing space). Although all the examples I can
Paul Nasrat wrote:
2010/1/11 Ohad Levy ohadl...@gmail.com:
is there a reason why not to strip all facts?
I'm a little cautious about doing that in general in case someone has
a dependancy on a fact with a space in or valid but odd setups
(directories with trailing space). Although all the
Markus Roberts mar...@reality.com writes:
This last bears some elaboration; if the following steps occur:
* Puppet execs child process (apt-get, for example) and attempts to
capture the output
* The child process launches a grandchild process (the buggy version
of the osirid daemon
Markus Roberts mar...@reductivelabs.com writes:
This patch incorporates Ricky Zhou's idea for establishing a default
SIGPIPE trap of IGNORE in child, which protects really dumb daemons
from the effects of writing to a closed stdout. As even moderately
smart daemons ought not do this, these
just a note: it's quite essential that SIGPIPE is set to SIG_DFL in the
child process. luckily, execve(2) will reset to SIG_DFL unless it's set
to SIG_IGN. a child with SIGPIPE set unexpectedly to SIG_IGN can easily
get stuck in endless loops. the grandchild SHOULD die when it tries to
Markus Roberts mar...@reductivelabs.com writes:
just a note: it's quite essential that SIGPIPE is set to SIG_DFL in the
child process. luckily, execve(2) will reset to SIG_DFL unless it's set
to SIG_IGN. a child with SIGPIPE set unexpectedly to SIG_IGN can easily
get stuck in endless loops.
Kjetil --
this isn't really a very unique problem to solve, a lot of applications
have done this before, efficiently. (e.g., reading large blocks is
guaranteed safe when reading from slow devices, this includes pipes
and sockets. a fast device (ie. hard disk) will hang until the whole
On Jan 11, 2010, at 9:58 AM, Markus Roberts wrote:
Not quite. Cargo cult is doing something that looks like what
worked in the past, in the hopes of recreating that past success, and
continuing to do so in the face of repeated failure. Change
management is refraining from changing something
All --
We have a confirmed regression (see #3025) of #1563 (apt-get hangs
puppet) which was introduced in 0.25.2 with the patch to use pipes
instead of temp files to communicate with sub-processes for SELinux
(#2731). This problem persists despite an heroic effort by all the
kings horses to
On Jan 11, 2010, at 1:42 PM, Markus Roberts wrote:
All --
We have a confirmed regression (see #3025) of #1563 (apt-get hangs
puppet) which was introduced in 0.25.2 with the patch to use pipes
instead of temp files to communicate with sub-processes for SELinux
(#2731). This problem persists
In the coin flip I got parsimony and that leaves James stuck as the
voice for catholicon. His proposed list of possible patches to
include in 0.25.3 as long as we're at it:
•James's packaging changes
•Bug #1464: Mount resource complains about missing options field
•Bug
On 12/01/2010 at 9:20 am, Markus Roberts mar...@reductivelabs.com
wrote:
In the coin flip I got parsimony and that leaves James stuck as the
voice for catholicon. His proposed list of possible patches to
include in 0.25.3 as long as we're at it:
●James's packaging changes
●
As the author of the patches for #1464 , #2845 and #2887 , I'm going to vote
for inclusion. They're all things that would have gone into 0.25.2 if had
gotten them in time. (but I'd be willing to listen to an argument that they
haven't been tested sufficiently, if someone wants to make that
I'd rather leave #1842 unpatched, since it doesn't do any harm, and removing
the option (even if it's doing nothing) could cause breakage.
That's basically where James and I were. So the include list stands at:
•Reversion of #2713 to fix #3013 / #3023 / #3025
• James's
I'm good with all of these.
On Jan 11, 2010, at 3:01 PM, Jesse A Wolfe wrote:
As the author of the patches for #1464 , #2845 and #2887 , I'm going
to vote for inclusion. They're all things that would have gone into
0.25.2 if had gotten them in time. (but I'd be willing to listen to
an
I'm not hearing any objections, so we're going to go for it, omitting
#1842 on vague collective unease but letting the rest ride.
-- Markus
--
You received this message because you are subscribed to the Google Groups
Puppet Developers group.
To post to this group, send email to
2010/1/11 Thomas Bellman bell...@nsc.liu.se:
Paul Nasrat wrote:
2010/1/11 Ohad Levy ohadl...@gmail.com:
is there a reason why not to strip all facts?
I'm a little cautious about doing that in general in case someone has
a dependancy on a fact with a space in or valid but odd setups
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Not meaning to butt in on my own patch thread, but is there anything
particularly wrong with the code below?
This patch was generated from the latest head of the day.
I'll have a couple of hours coming up to work on this so would like to
get some
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Puppet 0.25.3 - code-named Clifford
The 0.25.3 release is a maintenance release in the 0.25.x branch. The
release addresses a regression introduced in 0.25.2 that caused issues
with command execution.
The release is available at:
19 matches
Mail list logo
