ach function it appears in.)
make[1]: *** [init.o] Error 1
make[1]: Leaving directory `/home/pere/src/sysvinitsvn/src'
make: *** [all] Error 2
%
Happy hacking,
--
Petter Reinholdtsen
[Petter Reinholdtsen]
> The current svn source fail to build with SE linux enabled.
I investigated some more, and the MNT_DETACH was introduced after
glibc 2.7, which is the version in Debian Lenny on my laptop. It is
present in 2.10 in the Debian unstable repository. This patch might
solve
Anyone here know /bin/mountpoint well enough to comment on
http://bugs.debian.org/460898 >?
I suspect the patch is good, but have not spent time on the source to
figure it out.
Bringing it up here in case we want it in the new sysvinit release.
Happy hacking,
--
Petter Reinholdtsen
cking,
--
Petter Reinholdtsen
PWR reason
(OK/BAD) */
#define PWRSTAT"/var/run/powerstatus" /* COMPAT: SIGPWR
reason (OK/BAD) */
#if 0
Happy hacking,
--
Petter Reinholdtsen
ete.
Please commit the Debian patches you believe make sense for this
release, and let me know which were commited. I doubt I will have
time to do much testing this week, but would love a new release. If
you got time to work on it, please push out a new release quickly. :)
Happy hacking,
--
Petter Reinholdtsen
Looking at the sysvinit bugs in Debian, I came across
http://bugs.debian.org/68621 > again, asking for PAM support in
sulogin. This sound like a good idea to me. Werner, what is your
view on this?
I do not know if the patch still apply.
Happy hacking,
--
Petter Reinholdtsen
On Tue, Mar 23, 2010 at 01:45:05PM +0100, Dr. Werner Fink wrote:
> On Mon, Mar 22, 2010 at 07:37:09PM +0100, Petter Reinholdtsen wrote:
> > Looking at the sysvinit bugs in Debian, I came across
> > http://bugs.debian.org/68621 > again, asking for PAM support in
> > sulogin.
gt; otherwise the stuff makes no sense.
I would expect it to use /etc/pam.d/login.
Happy hacking,
--
Petter Reinholdtsen
Werner, if you give me your freshmeat user name, I can try to add you
to the freshmeat project. I've taken over the
http://freshmeat.net/projects/sysvinit > management, but would
like to give you access too.
Happy hacking,
--
Petter Reinholdtsen
n other sides.
I have no input on this. :)
Happy hacking,
--
Petter Reinholdtsen
w, why is the pam stuff in init.c, and not in sulogin.c?
I am staring to suspect that we should postpone the PAM stuff to a
later release, while we figure out all these unknowns.
Happy hacking,
--
Petter Reinholdtsen
d look at the PAM stuff for the next release.
I have not yet had time to wrap my head around why PAM is useful in
init, and thus do not really have an opinion. If you believe it is
useful, and the Solaris version already got it, that is two good
arguments for including it. But not now
[Werner Fink]
> OK, just do it.
I've stamped and tagged the new release, but seem to lack the
privilege on https://savannah.nongnu.org/projects/sysvinit> to do
releases. Can you fix that?
This is the final changelog:
sysvinit (2.88dsf) world; urgency=low
[ Petter Reinholdtsen ]
[Petter Reinholdtsen]
> I've stamped and tagged the new release,
I made my tarball for this available from
http://people.skolelinux.org/pere/sysvinit-2.88dsf.tar.gz > for
those that want to have a look.
Happy hacking,
--
Petter Reinholdtsen
e this
savannah stuff later?
Happy hacking,
--
Petter Reinholdtsen
ppy hacking,
--
Petter Reinholdtsen
y hacking,
--
Petter Reinholdtsen
rner, you applied it. Do I misunderstand something?
Happy hacking,
--
Petter Reinholdtsen
sed variable 'path'
killall5.c:737: warning: unused variable 'exe'
killall5.c:690: warning: unused variable 'real'
Werner, any preferences on how to fix this?
Happy hacking,
--
Petter Reinholdtsen
According to http://bugs.debian.org/580272 >, the sysvinit code
to enable SELinux is broken. Werner, you implemented the current
version. Do you have any idea how it should be fixed?
Happy hacking,
--
Petter Reinholdtsen
Hi, Werner. Take a look at http://bugs.debian.org/580564 >.
I've verified this with our test suite, and insserv no longer
understand the X-Interactive: true header. This test suite function
fail:
test_x_interactive() {
echo
echo "info: Check if X-Interactive header work"
echo "BTS #458224"
echo
The current insserv version generate an empty directory every time
make is called. The cause is this line:
override TMP :=$(shell mktemp -d $(PACKAGE)-$(VERSION).)
Any ideas how to rewrite insserv to avoid this?
Happy hacking,
--
Petter Reinholdtsen
< shl->rm_eo) {
+ *(pbuf+shl->rm_eo) = '\0';
+ interactive = xstrdup(pbuf+shl->rm_so);
} else
interactive = empty;
}
Happy hacking,
--
Petter Reinholdtsen
yname, argv);
} else {
if (enforce > 0) {
@@ -2884,7 +2878,6 @@
}
}
}
- if (rc == 0) umount2("/proc", MNT_DETACH);
}
#endif
/* Start booting. */
Happy hacking,
--
Petter Reinholdtsen
iting the timeout. This now make
> the system not usable.
>
> If ctrl+c is not supported anymore, it should blocked to avoid this
> strange behaviour.
I tested this, and pressing ^c while rcS.d/ scripts are executed break
the entire boot.
Happy hacking,
--
Petter Reinholdtsen
ve run alone with direct connection to the
tty, if I have understood how startpar work. Not sure how that affect
^c behaviour.
Happy hacking,
--
Petter Reinholdtsen
Hi, Werner. Not sure where to send startpar issues, so I try here as
I know I can reach you here. :)
I proposed a patch for http://bugs.debian.org/584102 >, basicly
allowing startpar to continue even when tcgetattr(0, &tio) by running
scripts sequencially. Can you have a look at let me know if i
a how to get parallel booting working also when stdin and
stdout is a pipe, like it is with OpenVZ at the moment?
Happy hacking,
--
Petter Reinholdtsen
strcmp(arg, "start"))
+ if (arg && !strcmp(arg, "start"))
{
int s, t, len;
pid_t child;
Happy hacking,
--
Petter Reinholdtsen
o networked file systems are listed
in /etc/fstab. :) Not sure how hard that would be. Lets continue this
part of the discussion without CC-ing the BTS.
Happy hacking,
--
Petter Reinholdtsen
# Provides: dnsmasq $named
# ...
## END INIT INFO
Werner, any opinion on this?
I suspect the LSB should be updated to document that this is an
allowed setup, if we go this way.
Happy hacking,
--
Petter Reinholdtsen
nick names for a Provide.
This seem to me to be equivalent to
# Provides: dnsmasq $named
How would the expressed relation be different by inventing a new
header type for virtual facility provides?
Happy hacking,
--
Petter Reinholdtsen
[Florian Hülsmann]
> In file included from /usr/include/linux/fs.h:26,
> from hddown.c:39:
> /usr/include/asm/bitops.h:585: error: syntax error before "int"
> make[1]: *** [hddown.o] Error 1
> make[1]: Leaving directory `/usr/local/sysvinit-2.88dsf/src'
> make: *** [all] Error 2
>
should do before I unsubscribe from the mailing list.
I hope some of the current maintainer will join Werner Fink in
maintaining the upstream version and take over for me there too.
I'm sorry to let this project go, but I just do not have spare time to
spend on it any more. :(
--
Happy hacking
P
r releases to
keep everyone in sync?
--
Happy hacking
Petter Reinholdtsen
d killall5 should stick together, as they are the
same binary, so it do not make sense to move pidof but not killall5.
--
Happy hacking
Petter Reinholdtsen
me and at work to do much on sysvinit. :(
--
Happy hacking
Petter Reinholdtsen
wo functions being
> inadvertently exported.
Thank you. I've commited your patches to the upstream svn repository.
It will be included in the next release. :)
--
Happy hacking
Petter Reinholdtsen
Make sense. I extended it a bit to also work with the Debian
multiarch system, and commited it upstream. Should make it into the
next release.
--
Happy hacking
Petter Reinholdtsen
r
reject requests as soon as I notice them.
--
Happy hacking
Petter Reinholdtsen
f it work for you. :)
--
Happy hacking
Petter Reinholdtsen
[Matias A. Fonzo]
> Is working under musl-0.9.15 and glibc-2.13.
>
> I have another diff as the complement of the previous patch
> (attached).
Thank you. Commited. :)
--
Happy hacking
Petter Reinholdtsen
[Matias A. Fonzo]
> Hi again,
>
> The following patch (diff) adds support in the Makefile to obtain
> binaries statically linked (optionally).
Thank you. Commited.
--
Happy hacking
Petter Reinholdtsen
It would be a lot easier to release a new version of sysvinit if we
had a test suite testing that the system work. At the moment, I am
unsure if the system work at all until I try to boot with it, and that
is a some work to set up.
If we had a simple 'make check' target running a test suite, I wo
[Matias A. Fonzo]
> GCC produces warnings for some unused variables in shutdown.c and
> sulogin.c:
I fixed it in subversion, using a slightly different patch.
--
Happy hacking
Petter Reinholdtsen
test the current subversion code and let us know if
there are problems with it.
--
Happy hacking
Petter Reinholdtsen
In Debian, the location of the initctl socket used by sysvinit is
moved from /dev/initctl to /run/initctl. Should we make this the
default for sysvinit?
--
Happy hacking
Petter Reinholdtsen
t; /dev/initcl to /run?.
http://bugs.debian.org/638019 > contain the request to change
it. Getting it to work on more platforms seem to be the key part. :)
--
Happy hacking
Petter Reinholdtsen
to hope to reach the same conclusion as init did when
it booted when selecting location of the pipe.
--
Happy hacking
Petter Reinholdtsen
contain
headers. Change based on patch from Elias Pipping.
* Add fix for Debian bug #580272: use return value 1 of
is_selinux_enabled() to determine if SELinux is enabled,
otherwise initialize SELinux and load the policy. Patch from
Petter Reinholdtsen.
* Make quotes visible in
o could help fix all the issues discovered by Coverity?
--
Happy hacking
Petter Reinholdtsen
way it is.
> I'll present my vision in next dwo weeks with working code.
Thanks.
> Sure, i can try.
>
> Where i can find a report?
Check out
https://lists.gnu.org/archive/html/sysvinit-devel/2014-01/msg5.html >,
register and ask for access. Next, I can grant it.
--
Vennlig hilsen
Petter Reinholdtsen
Hi.
I just tried to update the Debian package of insserver from 1.14.0 to
1.16.0, and discovered that the script ordering produced is completely
bogus. The bug is in the unpatch package too, but not detected by the
test suite. Here is the last output from 'make check':
info: test if loop invol
[Petter Reinholdtsen]
> Anyone got a clue what is going on?
I got a clue. I had forgotten the SUSE setup, where every script have
both stop and start symlinks. The real problem I had was that the
code to insert default settings in the non-SUSE case was broken. This
patch solved the issue.
so I am unsure exactly what is the correct behaviour here.
Are there other SUSE patches floating around?
--
Happy hacking
Petter Reinholdtsen
nit/ >,
but have not been able to find any other releases. I've copied that
file to savannah. Werner, do you know where startpar tarballs can be
found? Would be good to have more of the startpar history in the
download directory.
--
Happy hacking
Petter Reinholdtsen
oops_error = -1;
> + return NULL;
> + }
> + break;
Perhaps the if(fscanf)...} part should be made into a function or
macro, to avoid duplicate code? It is repeated quite a lot of times.
--
Happy hacking
Petter Reinholdtsen
evel == 0))
Did you figure out what the point of checking wrote_utmp_rlevel is?
It seem like a no-brainer, but I did not dare to do this change myself
without understanding why the code was like it was and exactly what it
did. Can you explain?
--
Happy hacking
Petter Reinholdtsen
[Michał Kulling]
> In current version spawn() function using dummy int f for handling
> opened TTY - i've changed it to int ftty variable.
Thank you. Applied after moving the 'int ftty;' part to the scope where
it is used.
--
Happy hacking
Petter Reinholdtsen
[Michał Kulling]
> Currently in spawn() function is using dup() without return code
> validation, i added simple check with additional output to stderr.
Thank you. Applied.
--
Happy hacking
Petter Reinholdtsen
thing to fix in sysvinit too?
--
Happy hacking
Petter Reinholdtsen
t tried to close it. But I have not
investigated either.
The previous patch made it easier to see in Coverity exactly where
things go wrong, which is why I applied them. Did you expect them to
get rid of the leak?
--
Happy hacking
Petter Reinholdtsen
I've just registered https://scan.coverity.com/projects/1408 >
to insserv for bugs too. The current svn repo is build and uploaded,
and I am waiting for the result. Should be available in a day or two,
I guess.
--
Happy hacking
Petter Reinholdtsen
the behaviour that
shutdown have had the last 20 years or so? Can you provide a table
showing the behaviour of different shutdown implementations?
--
Happy hacking
Petter Reinholdtsen
this behavior really needs to be
> retained. 20 years of tradition doesn't mean it makes sense ;).
> the sysvinit man page doesn't even make it clear that the -h is
> required.
Aha. :)
--
Happy hacking
Petter Reinholdtsen
[Petter Reinholdtsen]
> I've just registered https://scan.coverity.com/projects/1408 >
> to insserv for bugs too. The current svn repo is build and
> uploaded, and I am waiting for the result. Should be available in a
> day or two, I guess.
The info is available from
https
://anonscm.debian.org/viewvc/initscripts-ng/trunk/src/insserv/debian/patches/
>.
Is this a change to keep and commit upstream, or should I drop it?
--
Happy hacking
Petter Reinholdtsen
to use my github account.
I added the variable. :)
--
Happy hacking
Petter Reinholdtsen
ow, so I suspect it
should be fixed. Any details on how to reproduce it?
--
Vennlig hilsen
Petter Reinholdtsen
essary.
The fact that it only work when / is writable is not really unsafe, it
is just a requirement that the sysadmin need to be aware of if / is
mounted read only by default.
--
Happy hacking
Petter Reinholdtsen
TOU, SIG_DFL);
- TEMP_FAILURE_RETRY(dup2(p->fd, 0));
- TEMP_FAILURE_RETRY(dup2(2, 1));
+ (void)TEMP_FAILURE_RETRY(dup2(p->fd, 0));
+ (void)TEMP_FAILURE_RETRY(dup2(2, 1));
closeall();
execlp(myname, myname, "-f", "--", p->name, NULL);
--
Happy hacking
Petter Reinholdtsen
= (long int) (expression); \
while (__result == -1L && errno == EINTR); \
__result; }))
--
Happy hacking
Petter Reinholdtsen
nge seem compelling to me, as it is a good thing to
have consistent options across all command line tools. Any
objections?
I counted using commands similar to this:
zgrep -- --dryrun /usr/share/man/man*/*|wc -l
zgrep -- --dry-run /usr/share/man/man*/*|wc -l
--
Happy hacking
Petter Reinholdtsen
why init exits?
I'm trimming the list of receivers to the sysvinit mailing list, to
avoid flodding too many people.
--
Happy hacking
Petter Reinholdtsen
init-devel@), and let us know what you plan to do with thew
coverity information?
--
Happy hacking
Petter Reinholdtsen
ted.
> However, we sure the memory is enough at boot time. But apps maybe
> use a lot of memory after boot, then reboot, and catch the error.
No idea.
--
Happy hacking
Petter Reinholdtsen
us know what you plan to do with thew
coverity information?
--
Happy hacking
Petter Reinholdtsen
se API or whatnot. :)
--
Happy hacking
Petter Reinholdtsen
" of the coverity
project get the emails.
--
Happy hacking
Petter Reinholdtsen
sh eyes soon
show up to move the project forward.
--
Happy hacking
Petter Reinholdtsen
[Petter Reinholdtsen
> Please present yourself on the mailing list.
[Mathieu Therrien]
> Hi. I'm software engineer (baccalaureate of University of Quebec at
> Montreal, Canada). Since 2007, I work for Bell Canada and Desjardins
> Bank (Quebec credit union) as cyber security
ded to the
> savannah upstream group too.)
I was a bit involved with migrating Debian Edu from subversion to git.
The scripts involved are available from
https://anonscm.debian.org/cgit/debian-edu/d-e-gitscripts.git/tree/ >,
but I no longer remember any details.
--
Happy hacking
Petter Reinholdtsen
t speak for the others, but I am lurking myself, but lack the spare
time needed to give sysvinit much priority. I also lurk in
#pkg-sysvinit on irc.debian.org.
What about yourself, what is your interest in sysvinit?
--
Happy hacking
Petter Reinholdtsen
sysvinit
Very good. Since no-one objected in this thread, I've approved your
membership in the Savannah project.
--
Vennlig hilsen
Petter Reinholdtsen
ate sysvinit upstream to git?
--
Happy hacking
Petter Reinholdtsen
on should be included when the svn repo is converted to
git. I suspect git-svn might do a good job with the migration, but have
not tested.
--
Happy hacking
Petter Reinholdtsen
in three, one for sysvinit,
startpar and insserv.
--
Happy hacking
Petter Reinholdtsen
dea.
Do you lack the skills required to migrate a Subversion repository to
git? The procedure is well documented on several web pages, and the
tools available are quite good. If you need more information, I am
happy to give guidance via IRC, the #pkg-sysvinit IRC channel on
irc.debian.org.
--
Happy hacking
Petter Reinholdtsen
repo from savannah, so no-one misunderstands and belive it
is the authorative source? Those that want to use git can still use the
git-svn translator, so there should be no negative effect on usability.
--
Happy hacking
Petter Reinholdtsen
. Is it possible to add three different git repositories to
a Savannah project? I could not find any answers in
https://savannah.gnu.org/maintenance/UsingGit/ >.
--
Happy hacking
Petter Reinholdtsen
[Petter Reinholdtsen]
> Anyone got any idea why the tag is missing?
I figured out a way to migrate the tags. Please have a look at the
script and the resulting git repository, to see if it look good to you.
If it work well for sysvinit, we should be able to migrate insserv and
startpar
[Petter Reinholdtsen]
>> The current subversion directory contain three subdirectories,
>> sysvinit, startpar and insserv. I suspect these should go into
>> separate git repositories. Is it possible to add three different git
>> repositories to a Savannah project? I co
[Petter Reinholdtsen]
> I created http://savannah.gnu.org/support/?109462 > to see if the
> Savannah admins could help me, but a similar request from October is
> still not answered, so I fear it will not get a reply soon. :(
This time the reply was quick, and the three git repos are
[Petter Reinholdtsen]
> This time the reply was quick, and the three git repos are now created
> and ready to be filled with the migrated content. Are the rest of you
> happy with the result from the migration script?
No-one objected, so I created three git repositories with tags
[Petter Reinholdtsen]
> Unfortunately I was not able to replace the 'master' branch in
> sysvinit.git with the migrated history, so that history is in the
> 'newmaster' branch. I've asked the savannah admins to replace master
> with newmaster in
> https://s
sure my preferences should matter much. I would prefer the
mailing list or the ticket tracking system to avoid forgetting the
patch.
--
Happy hacking
Petter Reinholdtsen
[Jesse Smith]
> I will check it out and, if all goes well, re-commit the
> patches/changes from the bug tracker and downstream.
I look forward to it.
--
Happy hacking
Petter Reinholdtsen
ot; branch around the end of the month. (Assuming no one
> objects or finds a flaw in the beta code.)
I'll see what I can do next month. I've spent so much time on the git
migration that my time budget for this is spent I have to focus my
effort elsewhere for the next few weeks.
--
Happy hacking
Petter Reinholdtsen
The migration was also briefly discussed on this mailing list
2011-06-21.
--
Happy hacking
Petter Reinholdtsen
1 - 100 of 113 matches
Mail list logo
