Package: sendmail
Version: 8.14.4-4
Severity: minor
Sendmail on Wheezy sometimes has interoperability problems with other
SSL implementations. Some of these can be fixed by disabling TLS 1.1
and TLS 1.2. Sendmail 8.14.8 supports SSL options to do this, but
Sendmail 8.14.4-4 does not. Could we b
Package: sendmail
Version: 8.14.4-4
Severity: important
Tags: patch
Hello,
Sendmail 8.14.4-4 fails with an error message under rare conditions:
http://serverfault.com/questions/482035/sendmail-issue-with-ipv6
If a domain has an MX record that is a CNAME that has both A and
records, sendmail
Package: libmilter1.0.1
Version: 8.14.3-9.4
Severity: normal
libmilter in Sendmail 8.14.3 leaks memory when smfi_setsymlist is called.
This bug was fixed in Sendmail 8.14.4 as per this release note from Sendmail:
LIBMILTER: Fix memory leak that occurred when smfi_setsymlist()
On Tue, 05 Apr 2011 22:56:05 -0400
"Kurt B. Kaiser" wrote:
> If you look at remind's manpage, it appears that the INCLUDE statement
> can be also used with a path relative to the directory of the
> reminders file in which it is found. See the example in the RUN
> COMMAND section that follows the
Hi,
This is not an important bug. See my reply at
https://rt.cpan.org/Ticket/Attachment/879055/455342/
I do not plan on fixing it in upstream.
Regards,
David.
(Sorry for sending my earlier reply to the wrong Debian bug #.)
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.
This is not an important bug. See my reply at
https://rt.cpan.org/Ticket/Attachment/879055/455342/
I do not plan on fixing it in upstream.
Regards,
David.
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists
This is not an important bug. See my response
at https://rt.cpan.org/Ticket/Attachment/879051/455339/
I do not plan on fixing it upstream.
Regards,
David.
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists
Hi,
The problem appears to be the img-reorder filter. It's not
sophisticated enough to handle the mixed HTML and JavaScript.
Regards,
David.
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Package: privoxy
Version: 3.0.9-1
Severity: important
When I visit http://www.bmo.com/home through Privoxy, it mangles
the file http://www.bmo.com/scripts/bmo_common.js so that the Web site no
longer works.
The file itself is rather large, but here's a unified diff:
--- GOOD-bmo_common.js 2010-
Maurice Massar wrote:
> So this check matches on the default sendmail hello message, which makes
> it quite contraproductive. (And just to nitpick: why exclude 127.0.0.1 but
> not ::1? (-;)
Indeed...
Here is a patch against our development version. It should be pretty obvious
how to apply it ag
Any word on the status of this bug?
Regards,
David.
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Package: sysklogd
Version: 1.5-5
Severity: important
syslogd in Lenny can't create files > 2GiB. It looks
like the bug fix for #120574 was reverted. Any reason
why?
Regards,
David.
-- System Information:
Debian Release: 5.0.2
APT prefers stable
APT policy: (500, 'stable')
Architecture: i
This is almost certainly *not* a bug in MIMEDefang, but
a bug in the way you're using it.
Read mimedefang.pl and look at the source code that detects
a verification server loop.
-- David.
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Tr
Package: perl
Version: 5.10.0-19
Severity: normal
*** Please type your report below this line ***
Perl 5.10.0 seems not to be able to create, destroy, then recreate an
embedded Perl interpreter the way Perl 5.8.8 could under etch.
Please see attached test case. Untar, then:
./configur
This patch is against "git master", but should apply cleanly;
embperl.c has not changed in ages. I'd appreciate it if
people could test this and let me know how it goes.
Regards,
David.
diff --git a/embperl.c b/embperl.c
index f89779a..d9f1837 100644
--- a/embperl.c
+++ b/embperl.c
@@ -38,6 +3
I wrote:
panic: MUTEX_LOCK (22) [op.c:453] at /usr/bin/mimedefang.pl line 21.
BEGIN failed--compilation aborted at /usr/bin/mimedefang.pl line 21.
This appears to be a bug in Perl. If you look at the source code
in perl.c:
static void
S_init_tls_and_interp(PerlInterpreter *my_perl)
{
dVAR
I'm the author of MIMEDefang. I can reproduce this on lenny.
Some stracing shows that perl_parse fails when you try to
do a reload:
panic: MUTEX_LOCK (22) [op.c:453] at /usr/bin/mimedefang.pl line 21.
BEGIN failed--compilation aborted at /usr/bin/mimedefang.pl line 21.
This looks like a bizarre
Steve McIntyre wrote:
Maybe I'm being dense here, but I don't think that
m/\/.*\/(Packages|Sources)$/o is going to match a file called
Packages.gz.
Here's the scenario: Suppose you have the following files that
are found in the order given:
Packages.gz
Packages
Then the original code will:
And here is a proposed patch: The idea is we do all the re-gzipping in one
pass and the md5summing in another.
Regards,
David.
--- make_disc_trees.pl.ORIG 2007-03-05 11:09:46.0 -0500
+++ make_disc_trees.pl 2008-10-27 16:49:08.0 -0400
@@ -323,16 +323,20 @@
return ($m
Package: debian-cd
Version: 3.0.2
Severity: important
In the file tools/make_disc_trees.pl in debian-cd-3.0.2, we see this:
sub md5_files_for_release {
my ($md5, $size, $filename);
$filename = $File::Find::name;
# Recompress the Packages and Sources files; workaround for bug
# #
Package: apt-cacher
Version: 1.5.3
Severity: minor
As per bug #479799, I went ahead and set up separate configuration files
for Ubuntu vs. Debian. It works nicely, except that
apt-cacher-report.pl hard-codes the configuration file and doesn't let
you override it, so I can't generate separate repo
Package: apt-cacher
Version: 1.5.3
Severity: important
This single, flat cache directory causes problems if you
try to use apt-cacher for Ubuntu as well as Debian. For
example, both Debian Etch and Ubuntu Gutsy ship a package
called "xfonts-knickers". Both of them call the file
xfonts-knickers_0
Noah Meyerhans wrote:
Though it's still possible that this affects packages other than
MIMEDefang, so I suppose we can't really consider it resolved generally.
The SA developers have stated that SA is not designed to run in a
multi-threaded environment, so I think it *is* resolved. Fixing thi
As of MIMEDefang 2.64, we properly set "$$" when forking a new
process, even with embedded Perl. So this issue is now moot.
Regards,
David.
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Please *do not* apply the 01-optional-filename.diff patch.
The upstream Remind code has been modified so that if Remind
is invoked as "rem", it assumes a default filename. Please do
not fork Debian's version of Remind from upstream unnecessarily.
Regards,
David (Remind author)
--
To UNSUBSC
MIMEDefang's "embedded Perl" is unrelated to the Debian "eperl" package.
Also, although most MIMEDefang users do in fact use it with Sendmail,
modern versions of Postfix have a complete-enough Milter implementation
that MIMEDefang works with Postfix also.
Regards,
David.
--
To UNSUBSCRIBE, e
[EMAIL PROTECTED] wrote:
I visited http://www.mimedefang.org/ but could not find a Submit Bug
button. Never mind.
The proper place to submit bug reports for CPAN modules is
http://rt.cpan.org/Public/Dist/Display.html?Name=mime-tools
MIME::Entity has got it wrong about *digest* boundaries.
This patch is against our development tree. May not apply cleanly
to 2.61.
--
David.
diff --git a/mimedefang.c b/mimedefang.c
index 8430c2c..e42ec7e 100644
--- a/mimedefang.c
+++ b/mimedefang.c
@@ -520,6 +520,9 @@ mfconnect(SMFICTX *ctx, char *hostname, _SOCK_ADDR *sa)
#endif
if (sa->
Package: sendmail
Version: 8.13.4-3sarge3
Severity: important
The default for the MTA Timeout.iconnect setting as set by the debian-mta
config file is 5 seconds. This is absurdly aggressive and causes ridiculous
amounts of unnecessary queuing. This timeout should be set to at least
30 seconds an
Package: sendmail
Version: 8.13.4-3sarge3
Severity: important
The Sendmail init script includes a "clean_queues" shell function that
makes several passes throught the Sendmail queue directory. This is fine
on a dinky toy system, but if you have a heavily loaded system talking
to a Microsoft Excha
Package: sendmail
Version: 8.13.4-3sarge3
Severity: normal
The /usr/share/sendmail/sendmail script uses ad-hoc locking techniques
when invoked with the argument cron-msp. Here's the code:
if [ ! -f $STAMP_DIR/cron_msp ]; then
touch $STAMP_DIR/cron_msp;
#clean_queues /var/spool/mq
Setting $Features{'Virus:CLAMAV'} unconditionally will NOT work
and will cause your filter to fail.
--
David (author of MIMEDefang)
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Richard,
Thanks for your reply.
> And which of the other distros (with simple makefiles) support
> * changing of databases or other .mc settings based upon the connection
> profile (dhcp domain, ppp connection, etc.)
> * allow you to run split receive/deliver daemons trivially (want
Package: sendmail
Version: 8.13.4-3
Severity: important
If the clock on the computer is reset because of battery loss or some
other reason so that the m4 files have a date far in the future, bootup
hangs as the Sendmail startup script creates, recreates, re-recreates, ...
the Makefile. I guess th
Dan Jacobson wrote:
> Inflexibility in pppd parsing of interface names strikes again:
> # pppd plugin rp-pppoe.so usb0
> Plugin rp-pppoe.so loaded.
> pppd: unrecognized option 'usb0'
Use nic-usb0 instead. This is fixed in 3.6; anything
beginning nic-* is assumed to be an interface name.
--
Davi
>>But I don't like that you now need to supply a filename to query
>>another date than today.
Honestly, I'm inclined just to get rid of "rem" and absolutely
require a filename for "remind". I don't think it's a terrible
hardship to require the filename; if people want to use a default
filename,
Hi, René
> I have implemented this behavior, a patch is appended. Could you both
> please test it?
Took a quick glance. My only criticism is:
> + dot_reminders = (char*)malloc(len);
> + snprintf(dot_reminders, len, "%s/%s", home, DOT_REMINDERS);
We should check for malloc failing. I w
Hi, Rene.
I plan on obsoleting "rem" by having Remind use a default
filename. However, if you do supply a date, time and/or rep argument,
I'm going to require the filename. So the syntax will be:
remind [options] [filename [date] [time] [*rep] ]
It's just too messy/complicated to guess whether
38 matches
Mail list logo
