Bug#291699: Support for ibook xkeymap
Package: xlibs
Version: 4.3.0.dfsg.1-10
Severity: wishlist
Tags: patch upstream
Hi,
I recently installed Debian testing on my iBook G4. This keyboard is
different from the Apple Desktop Keyboard. (cfr.
http://maconlinux.org/images/adbkeys.gif vs.
http://developer.apple.com/documentation/Hardware/Developer_Notes/Macintosh_CPUs-G4/iBookG4/3_Input-Output/chapter_4_section_13.html#//apple_ref/doc/uid/TP40001480-CH207-BCIDDBEG)
The most annyoing thing esp. for German users is the missing AltGr
functionality.
The included patch alows this setup:
# setxkbdmap -v 10
Setting verbose level to 10
locale is C
Applied rules from xfree86:
model: ibook
layout: de
variant:nodeadkeys
Trying to build keymap using the following components:
keycodes: xfree86(ibook)+aliases(qwertz)
types: complete
compat: complete
symbols:pc/pc(pc105)+pc/de(nodeadkeys)
geometry: macintosh(ibook)
For clarification, since kernel 2.6 the ppc keyboard code uses the input
driver and supplied PC-keycodes, hence the pc-keycodes and symbols are
correct.
This patch does:
- define an iBook keyboard geometry file
- map the Apple key to what AltGr does on a PC keyboard (Originally,
this was mapped to windows menu keys, where I never heard of anyone
using them, so IMHO the loss is minor. The other "Windows Key"
emulation codes are already "taken" by the mouse button emulation
kernel driver, so on an ibook, really noone misses them.)
- if you specify "ibook" as your keyboard model.
Bye,
Joerg
--- ./geometry/macintosh.orig 2005-01-22 14:44:45.169101088 +0100
+++ ./geometry/macintosh2005-01-22 15:02:19.259854936 +0100
@@ -153,3 +153,119 @@
}; // End of "Keypad" section
}; // End of "default" geometry
+
+xkb_geometry "ibook" {
+
+description= "iBook G4 Keyboard";
+width= 325;
+height= 200;
+
+shape.cornerRadius= 1;
+shape "NORM" { { [ 18,18] }, { [2,1], [ 16,17] } };
+shape "HALF" { { [ 18, 9] }, { [2,1], [ 16, 8] } };
+shape "BKSP" { { [ 28,18] }, { [2,1], [ 26,17] } };
+shape "TAB" { { [ 28,18] }, { [2,1], [ 26,17] } };
+shape "RTRN" {
+ { [ 0,0], [ 18, 0], [18,37], [4,37], [4,18], [0,18] },
+ { [ 2,1], [ 16, 1], [16,36], [6,36], [6,17], [2,17] } };
+shape "CAPS" { { [ 32,18] }, { [2,1], [ 30,17] } };
+shape "LMTA" { { [ 28,18] }, { [2,1], [ 26,17] } };
+shape "LFSH" { { [ 23,18] }, { [2,1], [ 21,17] } };
+shape "RTSH" { { [ 42,18] }, { [2,1], [ 40,17] } };
+shape "RALT" { { [ 23,18] }, { [2,1], [ 21,17] } };
+shape "SPCE" { { [ 94,18] }, { [2,1], [ 92,17] } };
+
+shape "POWR" { cornerRadius= 9, { [ 18, 18 ] } };
+
+shape "LED" { cornerRadius= 2, { [ 4, 4 ] } };
+
+
+indicator.onColor= "green";
+indicator.offColor= "green30";
+indicator.top= 67;
+indicator.shape= "LED";
+indicator "NumLock" { shape = "LED"; top= 57; left= 152; };
+
+text.color= "black";
+
+section.left= 23;
+row.left= 1;
+key.shape= "NORM";
+key.gap= 1;
+
+section "Power" {
+top= 16;
+ left= 240;
+ key.shape="POWR";
+ row {
+ top= 1;
+ keys { };
+ };
+}; // End of "Power" Section
+
+section "Function" {
+ top= 52;
+ key.shape="HALF";
+ row {
+ top= 1;
+ keys { ,
+ { , 11 }, , , , ,
+ { , 10 }, , , , , ,
+ { , 11 }
+ };
+ };
+}; // End of "Function" section
+
+section "Alpha" {
+ top= 62;
+ row {
+ top= 1;
+ keys { , , , , ,
+ , , , , ,
+ , , , { , "BKSP" }
+ };
+ };
+ row {
+ top= 20;
+ keys { { , "TAB" }, , , ,
+ , , , , , ,
+ , , , { , "RTRN" }
+ };
+ };
+ row {
+ top= 39;
+ keys { { , "CAPS" }, , , ,
+ , , , , , ,
+ , ,
+ };
+ };
+ row {
+ top= 58;
+ keys { { , "LFSH" }, , , , ,
+ , , , , , ,
+ , { , "RTSH" }
+ };
+ };
+ row {
+ top= 77;
+ keys { , { , "LCTL" }, { , "LALT" },
+ { , "RALT" },
+ { , "SPCE" },
+ { , "RALT" },
+ };
+ };
+}; // End of "Alpha" section
+section "Cursor" {
+ key.shape="HALF";
+ top= 138;
+ left= 242;
+ row {
+ top= 1;
+ keys { { , 20} };
+ };
+row {
+ top= 10;
+ keys { , , };
+ };
+}; // End of "Cursor section"
+};
--- ./keycodes/xfree86.orig 2005-01-20 08:25:10.0 +0100
+++ ./keycodes/xfree86 2005-01-22 15:07:03.359665200 +0100
@@ -12,6 +12,11 @@
include "xfree86(basic)"
= 51;
= 94;
+ = 113;
+// Microsoft keyboard extra keys
+ = 115;
+ = 116;
+
Bug#290259: Update regarding the keyboard
Hi, for completeness, I found a solution for my X-keyboard problem, and submitted it as bug #291699 As I have sent a Console keymap already, I supply only the files for X in this mail Bye, Joerg signature.asc Description: Digital signature
Bug#290259: Update regarding the keyboard
On Mon, Jan 24, 2005 at 03:22:18PM +0100, Holger Levsen wrote: > On Saturday 22 January 2005 16:12, Joerg Dorchain wrote: > > As I have sent a Console keymap already, I supply only the files for X > > in this mail > > you didn't :-) I am sorry, I was unclear, they are in bug#291699 I cannot attach them ATM, as I am on holidays and this internet cafe has a [EMAIL PROTECTED] keyboard and nothing but webmail. Bye, Joerg signature.asc Description: Digital signature
Bug#341541: Allow for simple username in pam_member_attribute
Package: libpam-ldap
Version: 178-1sarge1
Severity: wishlist
Tags: patch, upstream
libpam-ldap allow only for distinguished names in the
pam_member_attribute when checking groups for allowed logins.
With the attached patch, simple user as as they appear in the posixGroup
schema are possible as well, configurable via a config option.
The defaults are the used behaviour, of course.
Please consider integrating it.
Thanks,
Joerg
Index: pam_ldap.c
===
--- pam_ldap.c (revision 1)
+++ pam_ldap.c (working copy)
@@ -636,6 +636,7 @@
result->userattr = NULL;
result->groupattr = NULL;
result->groupdn = NULL;
+ result->groupattr_dn = 1;
result->getpolicy = 0;
result->checkhostattr = 0;
result->checkserviceattr = 0;
@@ -1040,6 +1041,10 @@
{
CHECKPOINTER (result->groupattr = strdup (v));
}
+ else if (!strcasecmp (k, "pam_member_attribute_is_dn"))
+ {
+ result->groupattr_dn = !!strcasecmp (v, "no");
+ }
else if (!strcasecmp (k, "pam_min_uid"))
{
result->min_uid = (uid_t) atol (v);
@@ -3898,7 +3903,8 @@
{
rc = ldap_compare_s (session->ld,
session->conf->groupdn,
- session->conf->groupattr, session->info->userdn);
+ session->conf->groupattr,
+
(session->conf->groupattr_dn)?session->info->userdn:session->info->username);
if (rc != LDAP_COMPARE_TRUE)
{
snprintf (buf, sizeof buf, "You must be a %s of %s to login.",
Index: pam_ldap.h
===
--- pam_ldap.h (revision 1)
+++ pam_ldap.h (working copy)
@@ -95,6 +95,8 @@
char *groupdn;
/* group membership attribute; defaults to uniquemember */
char *groupattr;
+/* groupattr checking for distinguished name */
+int groupattr_dn;
/* LDAP protocol version */
int version;
/* search timelimit */
Index: pam_ldap.5
===
--- pam_ldap.5 (revision 1)
+++ pam_ldap.5 (working copy)
@@ -327,12 +327,18 @@
.B pam_groupdn
Specifies the distinguished name of a group to which a user must belong
for logon authorization to succeed.
+.TP
.B pam_member_attribute
Specifies the attribute to use when testing a user's membership of a
group specified in the
.B pam_groupdn
option.
.TP
+.B pam_member_attribute_is_dn
+Specifies wether attribute specified by
+.B pam_member_attribut
+is a distinguished name. The default is "yes".
+.TP
.B pam_min_uid
If specified, a user must have a POSIX user ID of at least
.B uid
signature.asc
Description: Digital signature
Bug#341541: Acknowledgement (Allow for simple username in pam_member_attribute)
Hi all, I just noticed that this bug might be similiar to #292030, although I consider my patch more general ;-) Sorry for duplicates, Joerg signature.asc Description: Digital signature
Bug#341541: Typo in patch
Hi,
I am sorry for being unorganized.
AS I have no problems with non-double negations and bouncing keyboards,
here is the proper patch.
Bye,
Joerg
Index: pam_ldap.c
===
--- pam_ldap.c (revision 1)
+++ pam_ldap.c (working copy)
@@ -636,6 +636,7 @@
result->userattr = NULL;
result->groupattr = NULL;
result->groupdn = NULL;
+ result->groupattr_dn = 1;
result->getpolicy = 0;
result->checkhostattr = 0;
result->checkserviceattr = 0;
@@ -1040,6 +1041,10 @@
{
CHECKPOINTER (result->groupattr = strdup (v));
}
+ else if (!strcasecmp (k, "pam_member_attribute_is_dn"))
+ {
+ result->groupattr_dn = !strcasecmp (v, "no");
+ }
else if (!strcasecmp (k, "pam_min_uid"))
{
result->min_uid = (uid_t) atol (v);
@@ -3898,7 +3903,8 @@
{
rc = ldap_compare_s (session->ld,
session->conf->groupdn,
- session->conf->groupattr, session->info->userdn);
+ session->conf->groupattr,
+
(session->conf->groupattr_dn)?session->info->userdn:session->info->username);
if (rc != LDAP_COMPARE_TRUE)
{
snprintf (buf, sizeof buf, "You must be a %s of %s to login.",
Index: pam_ldap.h
===
--- pam_ldap.h (revision 1)
+++ pam_ldap.h (working copy)
@@ -95,6 +95,8 @@
char *groupdn;
/* group membership attribute; defaults to uniquemember */
char *groupattr;
+/* groupattr checking for distinguished name */
+int groupattr_dn;
/* LDAP protocol version */
int version;
/* search timelimit */
Index: pam_ldap.5
===
--- pam_ldap.5 (revision 1)
+++ pam_ldap.5 (working copy)
@@ -327,12 +327,18 @@
.B pam_groupdn
Specifies the distinguished name of a group to which a user must belong
for logon authorization to succeed.
+.TP
.B pam_member_attribute
Specifies the attribute to use when testing a user's membership of a
group specified in the
.B pam_groupdn
option.
.TP
+.B pam_member_attribute_is_dn
+Specifies wether attribute specified by
+.B pam_member_attribut
+is a distinguished name. The default is "yes".
+.TP
.B pam_min_uid
If specified, a user must have a POSIX user ID of at least
.B uid
signature.asc
Description: Digital signature
Bug#336876: at calculates date off by 1 year
Package: at Version: 3.1.9 Severity: important the following transcript show the bug [EMAIL PROTECTED]:/tmp$ date Tue Nov 1 20:04:50 CET 2005 [EMAIL PROTECTED]:/tmp$ at -v now + 3 month Thu Feb 1 20:05:00 2007 warning: commands will be executed using /bin/sh at> at time should be Feb 1 2006 at> job 10 at Thu Feb 1 20:05:00 2007 [EMAIL PROTECTED]:/tmp$ Clearly, it should be 2006. Kind regards, Joerg signature.asc Description: Digital signature
Bug#337389: Please include tcpwrappers support (e.g with this patch)
Package: sysklogd
Version: 1.4.1-17
Severity: wishlist
Tags: patch, upstream
Hi,
IMHO it would be a good idea for syslogd to support tcpwrappers, as it
helps DoS via the internet. Most other daemons allow some sort of
control from where they allow access, and tcpwrappers is one of the
easiest to apply. The control feature of tcpwrappers is not so important
in this case, as syslogd intentionally logs all received messages with
sender ;-)
The patch below is in works-for-me state. In real life the feature
should probably be bracketed by options, as there are platforms without
libwrap, and probably reflected in build-depends as well.
Kind regards,
Joerg
diff -urN sysklogd-jd/Makefile sysklogd-1.4.1/Makefile
--- sysklogd-jd/Makefile1998-10-12 22:25:15.0 +0200
+++ sysklogd-1.4.1/Makefile 2005-11-04 09:59:21.0 +0100
@@ -23,6 +23,9 @@
# uncommenting the following define.
# LIBS = ${LIBS} -linux
+# Use tcpwrappers
+LIBS = -lwrap
+
# Define the following to impart start-up delay in klogd. This is
# useful if klogd is started simultaneously or in close-proximity to syslogd.
# KLOGD_START_DELAY = -DKLOGD_DELAY=5
diff -urN sysklogd-jd/syslogd.c sysklogd-1.4.1/syslogd.c
--- sysklogd-jd/syslogd.c 2005-11-04 09:27:32.0 +0100
+++ sysklogd-1.4.1/syslogd.c2005-11-04 09:56:06.0 +0100
@@ -491,6 +491,7 @@
#include
#include
#include
+#include
#ifndef TESTING
#include "pidfile.h"
#endif
@@ -1139,8 +1140,15 @@
* letters so we could match them against
whatever.
* -Joey
*/
- printchopped(from, line, \
+ /* When using tcpwrapper be careful to
+* avoid loops, as it uses syslog by
+* itself. (Hopefully unix sockets) -JD
+*/
+ if (hosts_ctl("syslogd", STRING_UNKNOWN,
+ inet_ntoa(frominet.sin_addr),
STRING_UNKNOWN)) {
+ printchopped(from, line, \
i + 2, finet);
+ }
} else if (i < 0 && errno != EINTR && errno != EAGAIN) {
dprintf("INET socket error: %d = %s.\n", \
errno, strerror(errno));
signature.asc
Description: Digital signature
Bug#340310: echoping in https mode sends request in clear
Package: echoping Version: 5.2.0-2 Hi, I use echoping in conjunction with smokeping. I found a problem that is easiest reproduceable this way: - set up an https-server on localhost - Use the follwoing command line /usr/bin/echoping -v -t 10 -w 1 -4 -C -h /smokepingtest.txt -n 5 127.0.0.1:443 - watch e.g. with ngrep: ngrep -d lo port 443 - and see clear text: T 127.0.0.1:36378 -> 127.0.0.1:443 [AP] GET /smokepingtest.txt HTTP/1.1..User-Agent: Echoping/5.2.0..Host: 127.0.0.1:443..Connection: close Smokeping itself complains with: "TTP error "0 OK Clearly, when you specify -C also for localhost the request should be encrypted. Bye, Joerg signature.asc Description: Digital signature
Bug#304909: Please add rtp mp3 capaility to xmms
Package: xmms Version: 1.2.10-2 Severity: wishlist Hi, IMHO it would be a great when xmms would be able to play rtp streams containg mpeg audio data (aka "Internetradio") A patch is available at http://studwww.ira.uka.de/~s_oden/xmms-rtp.patch which makes handling of rtp streams similiar to http streams. As rtp can be multicasted, it is friendlier to the network than http. TIA for implementing this feature. Bye, Joerg signature.asc Description: Digital signature
Bug#306261: pppd persist eats up all cpu after reconnect
Package: ppp
Version: 2.4.3-20050321+1
Hi,
My setup is pppoe with the rp-pppoe-plugin. I am "alwalys on" with the
persist option, although forcibly disconnected by my provider every 24h.
Thank to the persist option pppd reconnects automatically after a
disconnect.
Now the problem: The first connects succeeds and all behaves as
exspected. After the the first dis- and reconnect (i.e. after 24h) pppd
eats up all cpu time. The reconnect itself works as expected, but the
systems becomes notably slow with pppd playing cpu-hog.
An strace looks like this:
read(14, 0x808f7a2, 1502) = -1 EAGAIN (Resource
temporarily unavailable)
read(13, 0x808f7a2, 1502) = -1 EAGAIN (Resource
temporarily unavailable)
rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0
rt_sigprocmask(SIG_BLOCK, [HUP INT USR2 TERM CHLD], NULL, 8) = 0
rt_sigprocmask(SIG_UNBLOCK, [HUP INT USR2 TERM CHLD], NULL, 8) = 0
gettimeofday({1114424467, 453717}, NULL) = 0
select(15, [12 13 14], NULL, [12 13 14], {6, 267221}) = 1 (in [12], left
{6, 268000})
gettimeofday({1114424467, 454119}, NULL) = 0
read(14, 0x808f7a2, 1502) = -1 EAGAIN (Resource
temporarily unavailable)
read(13, 0x808f7a2, 1502) = -1 EAGAIN (Resource
temporarily unavailable)
with fd 13 and 14 being connected to /dev/ppp.
Bye,
Joerg
signature.asc
Description: Digital signature
Bug#306261: pppd persist eats up all cpu after reconnect
On Mon, Apr 25, 2005 at 09:37:55PM +0200, Marco d'Itri wrote: > On Apr 25, Joerg Dorchain <[EMAIL PROTECTED]> wrote: > > > My setup is pppoe with the rp-pppoe-plugin. I am "alwalys on" with the > > persist option, although forcibly disconnected by my provider every 24h. > Which kernel version are you running? Linus' 2.6.11. Would the .config help as well? Bye, Joerg signature.asc Description: Digital signature
Bug#299577: Support for ibook xkeymap
On Tue, Mar 15, 2005 at 03:57:10PM +0100, Thorsten Gunkel wrote: > Just two small comments: > > - Apple printed an @ sign on the l key on my keyboard so maybe one should be > able get an @ from that key (I personally prefer the q key but for new > users this might be unexpected if l doesn't work). Well, should be easy to add. Just do and update the debian bug record with it (#291699 ;-) BTW, adding altgr to the console keymap was so easy that I forgot a report for it. I'll check wether it is still necessary. > - I'm used to have a delete and a remove key. Is that supported with your > current layout? I just suggested to use the enter key (left to the cursor > keys) for that for the console keymap (see Bug #299577). What is you > opinion about that? If I read the kernel driver correctly, delete is emulated by the driver as fn-backspace. (cfr. linux/drivers/macintosh/adbhid.c) BTW, I also have a patch waiting to be accepted that gives sysreq by fn-esc. Besides, the enter key has a second layer within the emulated numeric block (see the urls given in bug #291699), so IMHO there is only one meaning left, KP_ENTER. Thank you for pointing me to this. Bye, Joerg signature.asc Description: Digital signature
Bug#302709: pppoe persits does not work
Package: ppp Version: 2.4.3-20050321 Severity: important Hi, I have a flat rate pppoe connection which I want to be up all time. For obscure (and non-changeable) reasons, I session is disconnected by the remote side every 24h (=1440minutes). I can reconnect immediately after, so IMHO persist with 5 second holdoff ist close enough to "always on". This used to work quite well. Following bugs #298657 and #301018 I have installed ppp-2.4.3-20050321. Unfortunately, for me, it does not help. This is the log from a remote side terminated session. I am mainly wondering where the signal 15 (SIGTERM) comes from. (A SIGHUP triggers a reconnect as expected.) Old session: Mar 31 17:57:41 Redstar pppd[2734]: LCP terminated by peer Mar 31 17:57:41 Redstar pppd[2734]: Connect time 1440.0 minutes. Mar 31 17:57:41 Redstar pppd[2734]: Sent 8031016 bytes, received 41791108 bytes. Mar 31 17:57:41 Redstar pppd[2734]: Couldn't increase MTU to 1500 Mar 31 17:57:41 Redstar pppd[2734]: Couldn't increase MRU to 1500 Mar 31 17:57:44 Redstar pppd[2734]: Connection terminated. Mar 31 17:57:44 Redstar pppd[2734]: PPP session is 6304 Mar 31 17:57:44 Redstar pppd[2734]: Using interface ppp0 Mar 31 17:57:44 Redstar pppd[2734]: Connect: ppp0 <--> eth0 Mar 31 17:57:44 Redstar pppd[2734]: Couldn't increase MTU to 1500 Mar 31 17:57:44 Redstar pppd[2734]: Couldn't increase MRU to 1500 Mar 31 17:57:44 Redstar pppd[2734]: Terminating on signal 15 Mar 31 17:57:44 Redstar pppd[2734]: Modem hangup Mar 31 17:57:44 Redstar pppd[2734]: Exit. Here is the log from the manually started session. New Session: (ifup ppp0) Mar 31 18:22:12 Redstar pppd[6261]: Plugin rp-pppoe.so loaded. Mar 31 18:22:12 Redstar pppd[6262]: pppd 2.4.3 started by root, uid 0 Mar 31 18:22:13 Redstar pppd[6262]: PPP session is 5099 Mar 31 18:22:13 Redstar pppd[6262]: Using interface ppp0 Mar 31 18:22:13 Redstar pppd[6262]: Connect: ppp0 <--> eth0 Mar 31 18:22:13 Redstar pppd[6262]: Couldn't increase MTU to 1500 Mar 31 18:22:13 Redstar pppd[6262]: Couldn't increase MRU to 1500 Mar 31 18:22:13 Redstar pppd[6262]: Couldn't increase MRU to 1500 Mar 31 18:22:15 Redstar pppd[6262]: Couldn't increase MTU to 1500 Mar 31 18:22:15 Redstar pppd[6262]: Couldn't increase MRU to 1500 Mar 31 18:22:15 Redstar pppd[6262]: Couldn't increase MRU to 1500 Mar 31 18:22:22 Redstar pppd[6262]: Couldn't increase MTU to 1500 Mar 31 18:22:22 Redstar pppd[6262]: Couldn't increase MRU to 1500 Mar 31 18:22:22 Redstar pppd[6262]: Couldn't increase MTU to 1500 Mar 31 18:22:22 Redstar pppd[6262]: Couldn't increase MRU to 1500 Mar 31 18:22:23 Redstar pppd[6262]: PAP authentication succeeded Mar 31 18:22:23 Redstar pppd[6262]: peer from calling number 00:90:1A:40:04:F6 authorized Mar 31 18:22:23 Redstar pppd[6262]: Cannot determine ethernet address for proxy ARP Mar 31 18:22:23 Redstar pppd[6262]: local IP address 212.88.133.153 Mar 31 18:22:23 Redstar pppd[6262]: remote IP address 217.24.235.26 How it should be, i.e. my last working automatic reconnect: Mar 28 07:29:18 Redstar pppd[790]: LCP terminated by peer Mar 28 07:29:18 Redstar pppd[790]: Connect time 1439.8 minutes. Mar 28 07:29:18 Redstar pppd[790]: Sent 22386274 bytes, received 129721323 bytes. Mar 28 07:29:18 Redstar pppd[790]: Couldn't increase MTU to 1500 Mar 28 07:29:18 Redstar pppd[790]: Couldn't increase MRU to 1500 Mar 28 07:29:21 Redstar pppd[790]: Connection terminated. Mar 28 07:29:26 Redstar pppd[790]: PPP session is 6451 Mar 28 07:29:26 Redstar pppd[790]: Using interface ppp0 Mar 28 07:29:26 Redstar pppd[790]: Connect: ppp0 <--> eth0 Mar 28 07:29:26 Redstar pppd[790]: Couldn't increase MTU to 1500 Mar 28 07:29:26 Redstar pppd[790]: Couldn't increase MRU to 1500 Mar 28 07:29:26 Redstar pppd[790]: Couldn't increase MRU to 1500 Mar 28 07:29:28 Redstar pppd[790]: Couldn't increase MTU to 1500 Mar 28 07:29:28 Redstar pppd[790]: Couldn't increase MRU to 1500 Mar 28 07:29:28 Redstar pppd[790]: Couldn't increase MRU to 1500 Mar 28 07:29:35 Redstar pppd[790]: Couldn't increase MTU to 1500 Mar 28 07:29:35 Redstar pppd[790]: Couldn't increase MRU to 1500 Mar 28 07:29:36 Redstar pppd[790]: Couldn't increase MTU to 1500 Mar 28 07:29:36 Redstar pppd[790]: Couldn't increase MRU to 1500 Mar 28 07:29:37 Redstar pppd[790]: PAP authentication succeeded Mar 28 07:29:37 Redstar pppd[790]: peer from calling number 00:90:1A:40:04:F6 authorized Mar 28 07:29:37 Redstar pppd[790]: local IP address 212.88.133.153 Mar 28 07:29:37 Redstar pppd[790]: remote IP address 217.24.235.26 My config (if more excerpts are useful, just tell): /etc/network/interfaces: auto ppp0 iface ppp0 inet ppp provider dsl-provider Redstar:~# pppd dryrun call dsl-provider Plugin rp-pppoe.so loaded. pppd options in effect: holdoff 5 # (from /etc/ppp/peers/dsl-provider) persist # (from /etc/ppp/peers/dsl-provider) maxfail 0 # (from /etc/ppp/peers/dsl-provider) dryrun # (from command line) plugin rp-pppoe.so # (from /etc/pp
Bug#302709: pppoe persits does not work
On Sat, Apr 02, 2005 at 04:25:36PM +0200, Marco d'Itri wrote: > On Apr 02, Joerg Dorchain <[EMAIL PROTECTED]> wrote: > > > This is the log from a remote side terminated session. I am mainly > > wondering where the signal 15 (SIGTERM) comes from. (A SIGHUP triggers a > Me too. I suggest you add logging statements to pppd to determine this. Can you give give any hint what you mean by this? E.g. debug statement in the pppd source code? If so, where to start? Althoug I am quite familiar with ppp as a protocol, I never had a look at the sources before. Thank you for the quick reply, Joerg signature.asc Description: Digital signature
Bug#302709: More infos on SIGTERM
Hi, after some tests (each of which took 24h) I have a surprising result. When the ppp interface is started via ifup (regradless wether automitically at boot time or by hand from the command line), pppd gets a SIGTERM when the remote side closes the session. When I start the interface by hand via pon dsl-provider, pppd keeps on living when the remote side closes the connection and reconnects 5 seconds later (as configured). HTH, Joerg signature.asc Description: Digital signature
Bug#302709: Try experimental ifupdown
On Wed, Apr 06, 2005 at 05:54:54PM +0200, Thomas Hood wrote: > On Wed, 2005-04-06 at 16:57 +0200, Joerg Dorchain wrote: > > Maybe ifup and ifdown can use the environment variables to see > > whether they are run from hotplug scripts or not and accordingly leave > > pppd alone if they are. > > > It is easier to modify hotplug's net.agent script so that it avoids > calling ifdown if the interface is a PPP one. An even cleaner solution > would be to implement #303383; then PPP interfaces would be neither > upped nor downed by hotplug by default (as with all other kinds of > interfaces) and the administrator would enable hotplug control of PPP > interfaces by adding > > allow-hotplug ppp0 > > to /etc/network/interfaces. This is a generalization of the auto keyword. Great! Go for it, if my voice counts. Bye, Joerg signature.asc Description: Digital signature
Bug#302709: Try experimental ifupdown
On Wed, Apr 06, 2005 at 05:48:22PM +0200, Thomas Hood wrote: > On Wed, 2005-04-06 at 17:05 +0200, Marco d'Itri wrote: > > If persist is not used then starting pppd at boot time is not much > > useful, and this is the only sensible reason for starting it with > > ifupdown. > > > I think you are right: the majority of people who control PPP interfaces > with ifup/ifdown (rather than with pon/poff) are doing it because they > want PPP on boot. In that case they probably also add the "persist" > option by hand and in that case they don't want hotplug to ifdown the > PPP interface on loss of connection. I had the impression that pon/poff were going to be deprecated in favour of using ifup/ifdown as with all other interfaces. Bye, Joerg signature.asc Description: Digital signature
Bug#299577: Support for ibook xkeymap
On Fri, Apr 08, 2005 at 03:44:15PM +0200, Helge Kreutzmann wrote: > > Well, should be easy to add. Just do and update the debian bug record > > with it (#291699 ;-) > > Actually, 121297 already contains some info as well. Can you check > that your keyboard is the same as on the photo attached later in this bug? Looks like. You can convince yourself as well with xkbprint and the geometry file I did ;-) > > BTW, adding altgr to the console keymap was so easy that I forgot a > > report for it. I'll check wether it is still necessary. > > Please do. Getting it working "out of the box" is quite important. My > ibook works now as described in #250259. My keymap is the same as yours except two things: - I replaced KP_Comma with KP_Period (Matter of tast IMHO) - I have keycode 100 as Alt, where you have AltGr > > If I read the kernel driver correctly, delete is emulated by the driver > > as fn-backspace. (cfr. linux/drivers/macintosh/adbhid.c) > > Correct, so standard shutdown/reboot is fn-ctrl-alt-backspace on > ibook. (Took me quite a while myself to find it). Yes, this could be mentioned some explicitly ;-) Bye, Joerg signature.asc Description: Digital signature
Bug#321068: ip6tables-restore segfaults
Package: iptables Version: 1.3.1-2 I just updated to the mentioned version of iptables. The file below causes ip6tables-restore < to segfault. I'd expect a more descriptive error message like '..error on line...' The kernel is 2.6.12. The last known working version of iptables was 1.2.9. Bye, Joerg Rules file that segfaults. Mac-addresses are replaced by XXs. # Generated by ip6tables-save v1.2.9 on Mon Feb 9 15:58:00 2004 *filter :INPUT DROP [0:0] :FORWARD DROP [0:0] :OUTPUT ACCEPT [20:2080] :blocknlog - [0:0] :checkblock - [0:0] :checkmac - [0:0] -A INPUT -s ::/0 -d ::/0 -i lo -j ACCEPT -A INPUT -s ::/0 -d ::/0 -i sit1 -p tcp -m multiport --dports 22,25,53,80,443,993 -j ACCEPT -A INPUT -s ::/0 -d ::/0 -i sit1 -p udp -m multiport --dports 53,123 -j ACCEPT -A INPUT -s ::/0 -d ::/0 -j checkblock -A FORWARD -s ::/0 -d ::/0 -j checkblock -A blocknlog -s ::/0 -d ::/0 -m limit --limit 1/sec -j LOG --log-prefix "Bad v6 Packet:" --log-level 5 -A blocknlog -s ::/0 -d ::/0 -j DROP -A checkblock -s ::/0 -d ::/0 -j checkmac -A checkblock -s ::/0 -d ::/0 -j ACCEPT -A checkblock -s ::/0 -d ::/0 -j blocknlog -A checkmac -s ::/0 -d ::/0 -i eth2 -m mac --mac-source XX:XX:XX:XX:XX:XX -j RETURN -A checkmac -s ::/0 -d ::/0 -i eth2 -m mac --mac-source XX:XX:XX:XX:XX:XX -j RETURN -A checkmac -s ::/0 -d ::/0 -i eth2 -m mac --mac-source XX:XX:XX:XX:XX:XX -j RETURN -A checkmac -s ::/0 -d ::/0 -i eth1 -j RETURN -A checkmac -s ::/0 -d ::/0 -i sit1 -j RETURN -A checkmac -s ::/0 -d ::/0 -j blocknlog COMMIT # Completed on Mon Feb 9 15:58:00 2004 signature.asc Description: Digital signature
Bug#321655: nvsetenv and nvsetvol should sync nvram after changes
Package: powerpc-utils
Version: 1.1.3-16
Tags: patch
Hi,
nvsetenv and nvsetvol can write to the nvram. The current generic_nvram
driver (2.6.12) has no special close function. The pmac_nvram driver can
only read or write whole nvram banks, so it allocates a ram buffer for
them.
This leads to the effect that the updated values can get lost, as
the changes are not written back to the physical nvram (happens e.g. on
an iBook G4)
The enclosed patch works around this problem from userspace by calling
the available ioctl for syncing the nvram explicitly.
Bye,
Joerg
--- nvsetenv.c.orig 2005-08-06 19:20:54.0 +0200
+++ nvsetenv.c 2005-08-06 19:31:54.0 +0200
@@ -14,6 +14,12 @@
#include
#include
#include
+#include
+#include
+#ifndef IOC_NVRAM_SYNC
+#warning IOC_NVRAM_SYNC undefined -- update your headers
+#define IOC_NVRAM_SYNC _IO('p', 0x43)
+#endif
#define NVSTART0x1800 // Start of the NVRam OF partition
#define NVSIZE 0x800 // Size of of the NVRam
@@ -325,6 +331,7 @@
else
nvOld(ac, av, i, nvfd);
+(void) ioctl(nvfd, IOC_NVRAM_SYNC);
(void) close(nvfd);
exit(EXIT_SUCCESS);
}
--- nvsetvol.c.orig 2005-08-06 19:21:09.0 +0200
+++ nvsetvol.c 2005-08-06 19:33:23.0 +0200
@@ -15,6 +15,12 @@
#include
#include
#include
+#include
+#include
+#ifndef IOC_NVRAM_SYNC
+#warning IOC_NVRAM_SYNC undefined -- update your headers
+#define IOC_NVRAM_SYNC _IO('p', 0x43)
+#endif
typedef struct {
unsigned char sig;
@@ -99,6 +105,7 @@
die("error writing /dev/nvram");
printf("new volume is %d\n", buf[VOLADDR]);
}
+ ioctl(fd, IOC_NVRAM_SYNC);
close(fd);
return 0;
}
--- nvvideo.c.orig 2005-08-06 19:21:46.0 +0200
+++ nvvideo.c 2005-08-06 19:38:25.0 +0200
@@ -3,6 +3,12 @@
#include
#include
#include
+#include
+#include
+#ifndef IOC_NVRAM_SYNC
+#warning IOC_NVRAM_SYNC undefined -- update your headers
+#define IOC_NVRAM_SYNC _IO('p', 0x43)
+#endif
#define NVSTART0x140f
#define NVSIZE 0x2
@@ -96,6 +102,7 @@
break;
}
+(void) ioctl(nvfd, IOC_NVRAM_SYNC);
(void) close(nvfd);
exit(EXIT_SUCCESS);
}
signature.asc
Description: Digital signature
Bug#402121: tracepath6 gives IPV6_HOPLIMIT: Protocol not available
Package: iputils-tracepath
Version: 3:20020927-3.1
Severity: important
Tags: patch, upstream, ipv6
Hi,
tracepath6 always gives IPV6_HOPLIMIT: Protocol not available (2.6.19
kernel). Tracing this down gives a similiar patch as the one applied to
ping6.
Please consider for inclusion.
Bye,
Joerg
--- tracepath6.c.orig 2006-12-08 11:05:40.0 +0100
+++ tracepath6.c2006-12-08 11:06:33.0 +0100
@@ -354,7 +354,7 @@
perror("IP_RECVERR");
exit(1);
}
- if (setsockopt(fd, SOL_IPV6, IPV6_HOPLIMIT, &on, sizeof(on))) {
+ if (setsockopt(fd, SOL_IPV6, IPV6_RECVHOPLIMIT, &on, sizeof(on))) {
perror("IPV6_HOPLIMIT");
exit(1);
}
signature.asc
Description: Digital signature
Bug#294325: Please include DvB-T file for Duesseldorf
Package: dvb-utils Version: 1.1.0-4 Tags: upstream Severity: wishlist Hi, enclosed is an input file for scan that meets the dvb signals receivable in Duesseldorf and the Ruhrgebiet (Germany). A channels.conf files does not make sense ATM, as they are going to change the frequency usage in April 2005. The frequencies themselves will stay the same, so you only have to rerun scan after that date. Maybe you like to include it in the examples section of the documentation as de-duesseldorf and forward it to the upstream authors. TIA, Joerg # DVB-T Ruhrgebiet/Düsseldorf # T freq bw fec_hi fec_lo mod transmission-mode guard-interval hierarchy T 53800 8MHz 2/3 NONE QAM16 8k 1/4 NONE T 58600 8MHz 2/3 NONE QAM16 8k 1/4 NONE T 74600 8MHz 2/3 NONE QAM16 8k 1/4 NONE T 81800 8MHz 2/3 NONE QAM16 8k 1/4 NONE T 83400 8MHz 2/3 NONE QAM16 8k 1/4 NONE signature.asc Description: Digital signature
Bug#294325: Please include DvB-T file for Duesseldorf
On Wed, Feb 09, 2005 at 11:51:23AM +0100, Thomas Schmidt wrote: > Thank you for this file, upstream allready has such a file, called > de-Ruhrgebiet, i will include all frequency-files from the > cvs-repository to the new package, which will be uploaded in a few > minutes. Great! > The file from upstream differs a little bit, but i hope it works too. My source is http://nrw.ueberallfernsehen.de/10.html The missing frequencies won't be searched, so you will miss some channels. Merging them won't hurt, I suppose. Bye, Joerg signature.asc Description: Digital signature
Bug#295179: gwhois with tpcd errs on postinstall
Package: gwhois Version: 20050115 Hi, I have setup gwhois as a daemon service with a line in /etc/inetd.conf like this: whois stream tcp nowait nobody /usr/sbin/tcpd /usr/bin/gwhois Note the tcpd. Now when I tried to upgrade to the version mentionend, I get this dialog: Trying to add the following entry: whois stream tcp nowait nobody /usr/bin/gwhois gwhois There is already an entry for whois in /etc/inetd.conf, but I don't recognise it. Here is what it looks like: whois stream tcp nowait nobody /usr/sbin/tcpd /usr/bin/gwhois Do you want to ignore this potential problem and continue, or would you rather not do so now ? Continue? (n/y) If I choose n, it leave the file as it is, but fails with an error, making e.g. apt-get bringing up this question over and over (and other annoyance) If I choose y, it overwrites the line and is happy, but I am not, as this interferes with the site security. IMHO there should be an option "leave the strange line as it is and do not exit with an error", or make the postinstall recognize tcpd. TIA, Joerg signature.asc Description: Digital signature
Bug#288942: acknowledged by developer (Re: Bug#288942: Dependency on xpdf-utils is too strong)
On Sun, Jan 09, 2005 at 09:48:03PM -0800, Debian Bug Tracking System wrote:
> Joerg Dorchain wrote:
> > I was printing pdf file directly before, with the help of
> > foomatic-filters. As this works as well, I suggest to replace the
> > dependency on xpdf-utils by xpdf-utils | foomatic-filters in order to
> > have pdf printing working.
>
> Hmm, it looks you're lucky :-)
>
> As far as I view, almost drivers in foomatic-filter couldn't print PDF
> directly but called pdftops internally.
>
> Here is my result.
>
> D [10/Jan/2005:14:02:39 +0900] start_process("/usr/lib/cups/filter/pdftops",
> 0xbffefc40, 0xbffeefb0, 9, 11, 8)
> I [10/Jan/2005:14:02:39 +0900] Started filter /usr/lib/cups/filter/pdftops
> (PID 11155) for job 188.
> D [10/Jan/2005:14:02:39 +0900] StartJob: filter =
> "/usr/lib/cups/filter/pstops"
> D [10/Jan/2005:14:02:39 +0900] StartJob: filterfds[1] = [ 9 12 ]
> D [10/Jan/2005:14:02:39 +0900] start_process("/usr/lib/cups/filter/pstops",
> 0xbffefc40, 0xbffeefb0, 10, 12, 8)
> I [10/Jan/2005:14:02:39 +0900] Started filter /usr/lib/cups/filter/pstops
> (PID 11156) for job 188.
> D [10/Jan/2005:14:02:39 +0900] StartJob: filter =
> "/usr/lib/cups/filter/foomatic-rip"
Hm, two questions remain to me:
- What was the filter chain for pdf files before this package version?
- Would it be an idea to use the gs-esp package to handle pdf? If so,
can you please consider it a wishlist? (I am kind of minimalist when
it comes the packages ;-)
TIA,
Joerg
signature.asc
Description: Digital signature
Bug#288942: acknowledged by developer (Re: Bug#288942: Dependency on xpdf-utils is too strong)
On Wed, Jan 12, 2005 at 10:17:10AM +0900, Kenshi Muto wrote: > > CUPS has own pdftops and old cupsys package included this. > This pdftops is derrived from pdftops in Xpdf (almost same), and I had > been affected many security hole problem by it :-) > So, I think there isn't so big size expansion by depending xpdf-utils. Ok, this explains it. > > > - Would it be an idea to use the gs-esp package to handle pdf? If so, > > can you please consider it a wishlist? (I am kind of minimalist when > > it comes the packages ;-) > > As far as I know, gs-esp's PDF conversion is far from perfect, and it > is too hard to modify pdftops for using gs-esp. Well, it is derived from gnu-ghostscript which I (together with ghostview) use for pdf handling. Some features really don't work, but those are mainly interactive one and cannot be reprodruced on paper anyway. > Furthermore gs-esp needs many packages than xpdf-utils :-) Yes, but it is needed by cupsys anyway. Bye, Joerg signature.asc Description: Digital signature
Bug#290259: Lengthy Install-report
Package: installation-reports
INSTALL REPORT
Debian-installer-version: Debian GNU/Linux testing _Sarge_ - Official Snapshot
powerpc Binary-1 (20041113)
uname -a: Sorry, got lost, but install kernel was kernel-image-2.6.8-powerpc
from above media
Date: Dec 30 2004
Method: How did you install? What did you boot off? If network
install, from where? Proxied?
Installation and Boot Media on builtin DVD drive, to external Firewire HD
Machine: iBook
Processor: G4
Memory: 128MB
Root Device: Firewire, /dev/sda4
Output of lspci and lspci -n:
lspci:
:00:0b.0 Host bridge: Apple Computer Inc. UniNorth 2 AGP
:00:10.0 VGA compatible controller: ATI Technologies Inc RV250 5c63 [Radeon
Mobility 9200 M9+] (rev 01)
0001:10:0b.0 Host bridge: Apple Computer Inc. UniNorth 2 PCI
0001:10:12.0 Network controller: Broadcom Corporation BCM4306 802.11b/g
Wireless LAN Controller (rev 03)
0001:10:17.0 ff00: Apple Computer Inc. KeyLargo/Intrepid Mac I/O
0001:10:18.0 USB Controller: Apple Computer Inc. KeyLargo/Intrepid USB
0001:10:19.0 USB Controller: Apple Computer Inc. KeyLargo/Intrepid USB
0001:10:1a.0 USB Controller: Apple Computer Inc. KeyLargo/Intrepid USB
0001:10:1b.0 USB Controller: NEC Corporation USB (rev 43)
0001:10:1b.1 USB Controller: NEC Corporation USB (rev 43)
0001:10:1b.2 USB Controller: NEC Corporation USB 2.0 (rev 04)
0002:20:0b.0 Host bridge: Apple Computer Inc. UniNorth 2 Internal PCI
0002:20:0d.0 ff00: Apple Computer Inc. UniNorth/Intrepid ATA/100
0002:20:0e.0 FireWire (IEEE 1394): Apple Computer Inc. UniNorth 2 FireWire (rev
81)
0002:20:0f.0 : Apple Computer Inc. UniNorth 2 GMAC (Sun GEM) (rev ff)
lspci -n:
:00:0b.0 0600: 106b:0034
:00:10.0 0300: 1002:5c63 (rev 01)
0001:10:0b.0 0600: 106b:0035
0001:10:12.0 0280: 14e4:4320 (rev 03)
0001:10:17.0 ff00: 106b:003e
0001:10:18.0 0c03: 106b:003f
0001:10:19.0 0c03: 106b:003f
0001:10:1a.0 0c03: 106b:003f
0001:10:1b.0 0c03: 1033:0035 (rev 43)
0001:10:1b.1 0c03: 1033:0035 (rev 43)
0001:10:1b.2 0c03: 1033:00e0 (rev 04)
0002:20:0b.0 0600: 106b:0036
0002:20:0d.0 ff00: 106b:003b
0002:20:0e.0 0c00: 106b:0031 (rev 81)
0002:20:0f.0 : 106b:0032 (rev ff)
Base System Installation Checklist:
[O] = OK, [E] = Error (please elaborate below), [ ] = didn't try it
Initial boot worked:[O]
Configure network HW: [O]
Config network: [ ]
Detect CD: [O]
Load installer modules: [O]
Detect hard drives: [O]
Partition hard drives: [O]
Create file systems:[O]
Mount partitions: [O]
Install base system:[O]
Install boot loader:[E]
Reboot: [E]
Comments/Problems:
The boot loader install failed due to the fact that ofpath can't tranlate
firewire devices into openfirmware paths. See bug #288008
Workaround: download and install a precompiled yaboot via macosx on the macos
partition and use it to boot into the newly installed system.
Reboot failed due to missing firewire-related modules in the initrd image
created by the installer and timing problems. See bug #288010. Workaround is
also described in the bug.
Additionally the current firewire code (testing with both the supplied 2.6.8
and a selfcompiled pristine 2.6.10) in the kernel does not reliabliy detect the
firewire disk (about 1 out of 5 times it fails). A possible workaround would be
a special script on the initrd (which provides a framwork for this) which loads
the sbp2 modules, waits some seconds, checks /rpoc/partitions wether the root
devices shows up, and if not unloads the spb2 module and starts over again. The
delay is necessary as the firewire code needs some time to recognize the
devices on the bus.
The is no keymap available that fits my iBook (G4 with a german keymap). None
of the supplied ones allowed me to a) type all keys and get the same character
on the screen as noted on the keyboard and b) get the essential non-german
characters (@|\{[]}) in a simple way (e.g. AltGr-q for @, or AltGr-< for |, as
on a PC keyboard). Besides, Alt-Fx to switch Consoles was not possible during
install, which hindered debugging.
The iBook keyboard is a bit special, as it has only the AlNum Block, which an
addtional layer reachable via the Fn-Key, overlapping with the Numlock-Layer,
reachable via F6, and Mac-Special buttons, converted to special keycodes by the
mac input handler in the kernel. Anyway, please find attache the keymap I
currently use. (In case you wonder: Yes, the power button has the same keycode
as the Alt key. Yes, it works that way, and yes, it is strange)
Additionally, it would be an idea to very strongly recommend the pbbuttond
package (currently optional, no references to it the Installation Manual), and
the powernowd package (at least as long as the on_demand cpufreq governor does
not work), as without them the system behaves strange. (E.g. the Mac special
keys produce very strange characters on the screen). Alt last, have those
package on the first cd.
In expert mode, dependencies between the
Bug#297290: Please include linux-input-layer remote config
Package: lirc Version: 0.7.1pre1-2 Severity: wishlist Hi, for using remote control via receivers that come with several tv-cards, please include the attached config file in the remotes directory and in the package setup dialogue. It is taken from http://linux.bytesex.org/v4l2/linux-input-layer-lircd.conf and works universally with any remote control via the input layer. The last entry at http://linux.bytesex.org/v4l2/faq.html gives a short resume on the setup. (for me, it works this way with a cinergy t2 usb device) TIA, Joerg begin remote name linux-input-layer bits 32 begin codes ESC 0x10001 10x10002 20x10003 30x10004 40x10005 50x10006 60x10007 70x10008 80x10009 90x1000a 00x1000b MINUS0x1000c EQUAL0x1000d BACKSPACE0x1000e TAB 0x1000f Q0x10010 W0x10011 E0x10012 R0x10013 T0x10014 Y0x10015 U0x10016 I0x10017 O0x10018 P0x10019 LEFTBRACE0x1001a RIGHTBRACE 0x1001b ENTER0x1001c LEFTCTRL 0x1001d A0x1001e S0x1001f D0x10020 F0x10021 G0x10022 H0x10023 J0x10024 K0x10025 L0x10026 SEMICOLON0x10027 APOSTROPHE 0x10028 GRAVE0x10029 LEFTSHIFT0x1002a BACKSLASH0x1002b Z0x1002c X0x1002d C0x1002e V0x1002f B0x10030 N0x10031 M0x10032 COMMA0x10033 DOT 0x10034 SLASH0x10035 RIGHTSHIFT 0x10036 KPASTERISK 0x10037 LEFTALT 0x10038 SPACE0x10039 CAPSLOCK 0x1003a F1 0x1003b F2 0x1003c F3 0x1003d F4 0x1003e F5 0x1003f F6 0x10040 F7 0x10041 F8 0x10042 F9 0x10043 F10 0x10044 NUMLOCK 0x10045 SCROLLLOCK 0x10046 KP7 0x10047 KP8 0x10048 KP9 0x10049 KPMINUS 0x1004a KP4 0x1004b KP5 0x1004c KP6 0x1004d KPPLUS 0x1004e KP1 0x1004f KP2 0x10050 KP3 0x10051 KP0 0x10052 KPDOT0x10053 103RD0x10054 F13 0x10055 102ND0x10056 F11 0x10057 F12 0x10058 F14 0x10059 F15 0x1005a F16 0x1005b F17 0x1005c F18 0x1005d F19 0x1005e F20 0x1005f KPENTER 0x10060 RIGHTCTRL0x10061
Bug#409170: ntp does not react to multicast or broadcast ntp packets
On Tue, Feb 27, 2007 at 07:53:34PM +0100, Kurt Roeckx wrote: > > multicastclient > > > > as last lines. > > > > netstat -g does not show a multicast group membership, and an lsof does > > not show an open multicast socket of ntpd. > > > > Replacing the 224.0.1.1 on the server with the network broadcast address > > and replacing multicastclient with broadcastclient on the client does > > not change the behaviour. > > Can you take a look at > https://ntp.isc.org/bugs/show_bug.cgi?id=495 and see if that's about > your problem? From the description, it seems to be. Seems to be related. From the last report, some things changed. - I am now using br0 as main interface (with eth0 as a bridge member and a tap interface from qemu as the other bridge port) - lsof and netstat -g now report the proper multicast membership An strace shows the ntp multicasts are received on the proper socket. However they seem to be ignored for obscure reasons. Is there a way to get more debugging out of ntpd directly, e.g. a version with -d enabled? Bye, Joerg signature.asc Description: Digital signature
Bug#409170: ntp does not react to multicast or broadcast ntp packets
Addendum: With the nopeer option removed from the default restrict statement, multicast works as expected. The open question for me now is, why is there a difference between using eth0 directly and br0? Bye, Joerg signature.asc Description: Digital signature
Bug#409170: ntp does not react to multicast or broadcast ntp packets
On Sat, Mar 03, 2007 at 01:31:49PM +0100, Kurt Roeckx wrote: > On Wed, Feb 28, 2007 at 12:43:55PM +0100, Joerg Dorchain wrote: > > Addendum: > > > > With the nopeer option removed from the default restrict statement, > > multicast works as expected. > > So this was just a configuration error? Not solely. The multicast socket was not open with eth0. > > > The open question for me now is, why is there a difference between using > > eth0 directly and br0? > > I have no idea, but I assume this is a kernel problem? Hm, not necessarily. Something behaves different with eth0 and br0. I'll try to dig into this more deeply. But as it works-for-me-now(tm), my boss will most likely adjust my priority list. Bye, Joerg signature.asc Description: Digital signature
Bug#413236: dnstop should support ppp interfaces
Package: dnstop
Version: 20060517-1
Severity: wishlist
Tags: patch, upstream
Hello,
with the attached patch, dnstop works with linux ppp interfaces. Please
consider it or inclusion and upstream forwarding.
Bye,
Joerg
--- dnstop.c.orig 2007-03-03 16:19:41.0 +0100
+++ dnstop.c2007-03-03 16:29:17.0 +0100
@@ -49,6 +49,10 @@
#define ETHERTYPE_8021Q 0x8100
#endif
+#ifndef SLL_HDR_LEN
+#define SLL_HDR_LEN 16
+#endif
+
#if USE_PPP
#include
#define PPP_ADDRESS_VAL 0xff /* The address byte value */
@@ -605,6 +609,14 @@
#endif
+#ifdef DLT_LINUX_SLL
+int
+handle_sll(const u_char * pkt, int len)
+{
+return handle_ip((struct ip *) (pkt + SLL_HDR_LEN), len - SLL_HDR_LEN);
+}
+#endif
+
int
handle_ether(const u_char * pkt, int len)
{
@@ -1268,6 +1280,11 @@
handle_datalink = handle_raw;
break;
#endif
+#ifdef DLT_LINUX_SLL
+case DLT_LINUX_SLL:
+ handle_datalink = handle_sll;
+ break;
+#endif
case DLT_NULL:
handle_datalink = handle_null;
break;
signature.asc
Description: Digital signature
Bug#413236: Acknowledgement (dnstop should support ppp interfaces)
Addendum: with this patch, the "any" device works as well. Bye, Joerg signature.asc Description: Digital signature
Bug#400704: Greyed Recently Closed Tabs is reproduceable
Hello, Wojciech comment works in both ways: Setting a homepage to any URL makes the tab work. Setting it back to the empty string greyes it out. Bye, Joerg signature.asc Description: Digital signature
Bug#412092: [Patch] bug in gdth.c crashing machine
Package: linux-image-2.6.18-3-686
Version: 2.6.18-7
Tags: upstream,patch
Severity: important
Hello,
there is a bug in the gdth driver causing crashes and potential data
losses when using tape. The bug is also described and confirmed on the lklm
on 13 Oct 2006
Message-ID: <[EMAIL PROTECTED]>
The included patch is derived directly from that mail.
Please consider including it until it has made it upstream, as it
endangers machines doing tape backups.
Bye,
Joerg
--- gdth.c.orig 2007-02-23 12:48:56.0 +0100
+++ gdth.c 2007-02-23 12:51:44.0 +0100
@@ -3092,6 +3092,7 @@
cmdp->u.raw64.direction =
gdth_direction_tab[scp->cmnd[0]]==DOU ?
GDTH_DATA_OUT:GDTH_DATA_IN;
memcpy(cmdp->u.raw64.cmd,scp->cmnd,16);
+cmdp->u.raw64.sg_ranz= 0;
} else {
cmdp->u.raw.reserved = 0;
cmdp->u.raw.mdisc_time = 0;
@@ -3108,6 +3109,7 @@
cmdp->u.raw.direction =
gdth_direction_tab[scp->cmnd[0]]==DOU ?
GDTH_DATA_OUT:GDTH_DATA_IN;
memcpy(cmdp->u.raw.cmd,scp->cmnd,12);
+cmdp->u.raw.sg_ranz= 0;
}
if (scp->use_sg) {
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#408949: After upgrade torrus fails with DB_VERSION_MISMATCH
Package: torrus-common Version: 1.0.4-1 Severity: important Hi, after upgrading the torrus-common packe after a while (about last August) to the named version, torrus does not work anymore. in /var/log/torrus/collector.main.log the following lines show up: [29-Jan-2007 15:23:56 ] Torrus version 1.0.4 [29-Jan-2007 15:23:56 ] /usr/share/torrus/bin/collector started for tree main [29-Jan-2007 15:23:56*] Cannot create BerkeleyDB Environment: DB_VERSION_MISMATCH: Database environment version mismatch [29-Jan-2007 15:23:56*] Cannot create BerkeleyDB Environment: DB_VERSION_MISMATCH: Database environment version mismatch [29-Jan-2007 15:23:56*] Scheduler initialization error. Sleeping 1800 seconds Additionally I find file in /var/log/torrus/ like dbenv_errlog_3528 containing Program version 4.4 doesn't match environment version 27.2014 There seems to be a conversion tools for existing data missing, or at least a hint somewhere in /usr/share/doc/torrus-common/ Kind regards, Joerg -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#409170: ntp does not react to multicast or broadcast ntp packets
Package: ntp Version: 1:4.2.2.p4+dfsg-1 This package is running on a time server configured to send out multicast time informformation via broadcast 224.0.1.1 ttl 1 They arrive at a client (tcpdump shows it, strange enough with a ttl of 32), but it does not react to it. The client uses the supplied ntp.conf, with server statements commented out and disable auth multicastclient as last lines. netstat -g does not show a multicast group membership, and an lsof does not show an open multicast socket of ntpd. Replacing the 224.0.1.1 on the server with the network broadcast address and replacing multicastclient with broadcastclient on the client does not change the behaviour. Kind regards, Joerg signature.asc Description: Digital signature
Bug#408949: Correction
Jurij Smakov wrote: Hi, The recovery command should actually read db4.4-recover -h /var/lib/torrus/db It worked. Many thanks, Joerg -- - infeurope S.A. 62, rue Charles Martel L-2134 Luxembourg Luxembourg Tel: (+352) 25.22.33.1 Fax: (+352) 25.22.33.222 Web: www.infeurope.lu - Any views expressed are purely those of the writer and may not in any circumstances be regarded as stating an official position of infeurope S.A.! -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#394097: auth_pam does not work
Hello, thanks to all the previous posters. Unfortunately it does not work for me. My config snippet: DAV svn SVNPath /var/lib/svn ForceType text/plain SVNAutoversioning on Satisfy All AuthPAM_Enabled On AuthPAM_FallThrough off AuthType Basic AuthBasicAuthoritative off AuthName "Subversion Repository" Require group subversion My modules: alias.load authz_user.load dir.conf php4.load auth_basic.load autoindex.load dir.load proxy.conf auth_pam.load dav.load env.load proxy.load authn_default.load dav_fs.conf info.load rewrite.load authn_file.load dav_fs.load mime.load setenvif.load authz_default.load dav_svn.conf negotiation.load status.load authz_host.load dav_svn.load php4.conf Could one of the people who say it works please also post their list of included modules, esp. the auth* ones? Kind regards, Joerg signature.asc Description: Digital signature
Bug#409801: RFP: please include mod_authn_pam packages
Package: wnpp Severity: wishlist Hello, with reference to bug 394097 it would seem a good idea to have a mod_authn_pam package. This fits in the apache 2.2 athn framework. From the webpage: http://sourceforge.net/projects/mod-auth/ The mod_auth project provides authn and authz modules for the apache httpd 2.1 authentication framework. Project Admins: firechipmunk, honx Operating System: OS Independent (Written in an interpreted language) License: Apache Software License Category: Database, HTTP Servers, Site Management, Security Many thanks, Joerg signature.asc Description: Digital signature
Bug#419838: clamav-milter fails at system startup when using clamd
Package: clamav-milter Version: 0.90.2-1 Hi, I configured clamav-milter in /etc/default/clamav-milter with OPTIONS="--max-children=5 -ol -e -k 60 -q" Note especially the -e, which asks clamav-milter to use a running clamd for scanning. The /etc/init.d scripts are setup correctly so that the clamd script is run before the clamav-milter script. Unfortunately there is a race condition left. On my system though the clamd script returns before clamd has become fully operational, i.e. when its socket has not been created yet. So when clamav-milter starts up, it does not find the clamd-socket and refuses to start. I see several possible solution for this: - make the clamav-milter start script wait until the socket is created. Might be problematic when clamav-milter is configured not to use clamd - make the clamd start script wait until the socket has been created. - change clamav-milter to behave the same when when the socket does not exist on startup and when clamd is stopped after startup. This requires some change in the upstream code. Some screaming lines in a logfile would be usefull, too. This would be my personal favourite. Bye, Joerg signature.asc Description: Digital signature
Bug#420009: Apache2.conf does not work
Package: nagios2 Version: 2.6-3 Hello, I just installed a fresh naios2 from testing and found /etc/nagios2/apache2.conf not owkring. Apach2.2 even complains about a potential misconfiguration. The stylesheet aliases part # Where the stylesheets (config files) reside Alias /nagios2/stylesheets /etc/nagios2/stylesheets # nagios 1.x: #Alias /nagios/stylesheets /etc/nagios2/stylesheets needs to be in front of the html page alias: # Where the HTML pages live Alias /nagios2 /usr/share/nagios2/htdocs # nagios 1.x: #Alias /nagios /usr/share/nagios2/htdocs because otherwise it is never evaulated as the earlier and shorter alias matches first. IMHO this was introduced with the merge of html and and cgis into a single DirectoryMatch clause. Bye, Joerg signature.asc Description: Digital signature
Bug#420011: Stylesheets are missing
Package: nagios2 Version: 2.6-3 Hello, I just installed a fresh naios2 from testing and found /etc/nagios2/stylesheets empty. This gives a strange view via the web interface. For nagios, these files are in nagios-common. Maybe they belong into nagios2-common. Feel free to reassign if appropriate. Bye, Joerg -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#390191: mountall.sh gives error for /sys
Package: initscripts Version: 2.86.ds1-20 Severity: minor Hi, my /etc/fstab looks like: /dev/hda1 / ext3errors=remount-ro 0 1 /dev/hda2 noneswapsw 0 0 proc/proc procdefaults0 0 nfsd/proc/fs/nfsd nfsddefaults0 0 sysfs /syssysfs defaults0 0 tmpfs /dev/shmtmpfs defaults0 0 devpts /dev/ptsdevpts defaults0 0 usbfs /proc/bus/usb usbfs defaults,user,noauto0 0 I get an error message while booting: Fri Sep 29 18:58:26 2006: Will now mount local filesystems:mount: sysfs already mounted or /sys busy Fri Sep 29 18:58:26 2006: mount: according to mtab, /sys is already mounted on /sys Fri Sep 29 18:58:26 2006: ^[[31mfailed!^[[39;49m I noticed that /sys (and others) are mounted by mountkernfs.sh. IMHO there should be a handling for all filesystems handled by mountkernfs.sh as there is for /proc in mountall.sh to avoid this error message. Bye, Joerg signature.asc Description: Digital signature
Bug#733032: modprobe -d does the wrong thing
Package: kmod Version: 9-3 Hello, when I use the -d option to modprobe in order to load modules from a non-default directory, it is searched only for the first module. For the modules.dep.bin and all modules required by the given modules, it *prepends* the -d path to the default paths, resulting e.g. in /out/of/tree/modules//lib/modules/`uname -r`/... I would rather expect it to *replace* the default path with it, e.g. when I give # modprobe -d /out/of/tree/modules/ mymodule it should search /out/of/tree/modules/ for modules.dep.bin, and its subdirectories for the dependand modules, i.e. prefix the content of modules.dep with /out/of/tree/modules/ (just as it would use /lib/modules/`uname -r`/ without any options) Bye, Joerg signature.asc Description: Digital signature
Bug#651691: closed by "Rodolfo García Peñas (kix)" (Re: uswsusp needs specialkernel parameter to resume)
On Fri, Dec 27, 2013 at 12:06:05PM +, Debian Bug Tracking System wrote:
> I am closing this bug because I have no reply with in the last email.
Sorry, I did not notice it.
>
> uswsusp didsn't change their behavior, but other packages (like
> initramfs-tools) had changes. Please, test with other versions of
> uswsusp and initramfs-tools and reopen this bug if you think that the
> bug was not solved.
>
> I tested some versions here, and I cannot reproduce the problem. If
> you reopen the bug, please, include the info about the latest
> initramfs-tools and uswsusp package, and the configuration for
> /etc/uswsusp.conf and /etc/initramfs-tools/conf.d/resume files.
I checked the initramfs generated by a version of debian testing
of today (including uswsusp 1.0+20120915-4)
The file in question is scripts/local-premount/uswsusp, which
belongs to the uswsusp package.
Here, the line
[ -n "${resume}" ] || exit 0;
implies the resume variable (set in init from parsing the command
line) must be set in order to resume.
Former versions checked for the noresume variable and did not
resume when it was set (also set in init)
Bye,
Joerg
signature.asc
Description: Digital signature
Bug#651691: closed by "Rodolfo García Peñas (kix)"
On Fri, Dec 27, 2013 at 09:27:18PM +0100, Rodolfo García Peñas wrote: > > I don't agree :-) > > Yes, the file /usr/share/initramfs-tools/scripts/local-premount/uswsusp > belongs to uswsusp. The script checks if the $resume variable is set. If is > set, then the script calls /sbin/resume to load the save image. If is not > set, is because the system was not hibernated (normal boot). > > But it doesn't mean that the resume variable is set as kernel parameter. To my understanding /proc/cmdline is used to parse the variables in the init script parameters, but let's not be picky about it. [...] > If the variable resume was set, with the noresume option as kernel parameter, > is unset. I think I understand the resulting semantics of the scripts. My point is more but I preferred the old behaviour, and changing user visible behaviour without a mentioning it somewhere prominently is what I would call a regression. Esp. since if you boot without resuming once, you have no chance to resume to the old state again, since the swap device is overwritten, which can be quite unpleasant. Bye, Joerg signature.asc Description: Digital signature
Bug#637332: uswsusp: resume problem
On Fri, Dec 27, 2013 at 10:56:06PM +0100, Rodolfo García Peñas wrote: > > I am trying to solve this bug, but I need more info. Thanks, I'll do what I can. > > Please, could you try the latest initramfs-tools package (version 0.115)? The > initramfs-tools package had a problem with blkid, and this problem affected > uswsusp. This problem is solved now, but I need that you test this new > version. I am using initramfs-tools 0.115 to generate the initramfs I examined. > OTOH, if you are using lvm2, this package has a problem because doesn't > initialize the swap device if is in a LVM device. You can see the bug number > #544641 I use the workaround suggested in this bug report: I give resume=/dev/mapper/cryptvg-swap as kernel commandline parameter To my reading of the initscripts, this variant should work. (See end of scripts/local-top/lvm2) Practical test with debugging and breakpoints confirms it. IMHO the meaning of the resume variable varies a bit across all involved initscripts. I have not yet looked up if there is a debian policy about it, as said, I had a bad surprise with the change in behaviour. Bye, Joerg signature.asc Description: Digital signature
Bug#651691: closed by "Rodolfo García Peñas (kix)"
On Sat, Dec 28, 2013 at 01:02:38AM +0100, Rodolfo García Peñas wrote: > > Hi Joerg, > > I don't understand when you say "old behaviour". I was working in uswsusp but > I didn't change the behaviour. I added some things, like UUID support,... but > with the same behaviour. The old behaviour was that the suspended machine resumed without explicitly needing to give the swap partition as an argument to the resume option. > > I was checking previous uswsusp versions, and I didn't see big differences > with the old code. I am not sure if the problem is with uswsusp or with other > Debian packages (initramfs-tools?). Could you help me to discover when the > behaviour changed? Hm, if you checked uswsusp (and I agree that the current code does not look like it changed functionality), then others broke it. > > Old uswsusp package versions are here: > http://snapshot.debian.org/package/uswsusp/ > > You can download the old binary packages. I started to work in uswsusp in > version 1.0-1, so testing versions 0.8-1.2, 1.0-1 could help. I will try. However, while reading more initramfs scripts, I tend to wonder if not changes initramfs-tools (maybe in combination with the lvm2 initramfs script) led to the unexpected effects. Bye, Joerg signature.asc Description: Digital signature
Bug#732920: procps: sysctl --system fails to load /etc/sysctl.conf
Hello, would you mind uploading an interim packet with the fix? The broken version is now in testing. And a router not routing packets anymore due to sysctl not enabling forwarding is a serious functionality bug. Bye, Joerg signature.asc Description: Digital signature
Bug#711486: import_request_variables replacement
tags 711486 + patch
thanks
Hello,
replacing import_request_variables with extract is relatively
easy, e.g.
-import_request_variables('GPC', 'i_');
+extract($_GET, EXTR_PREFIX_ALL, 'i');
+extract($_POST, EXTR_PREFIX_ALL, 'i');
+extract($_COOKIE, EXTR_PREFIX_ALL, 'i');
as already written by the original poster.
Please note that this feature was removed from php for security
reasons (turning potentially unverified user input into code).
Also, please find a way to get the working parts a freeradius into
testing again. I would appreciate.
Bye,
Joerg
signature.asc
Description: Digital signature
Bug#637346: Please create /etc/video_format for default video format
retitle 637346 Please create /etc/video_format for default video format thanks This version of dvdauthor needs to have explicitly set a default video format in order to be useful. Options include environment variable, user config files or system config file. The file /etc/video_format is the system wide config file and should contain one line: "PAL" or "NTSC", as # cat /etc/video_format PAL The value could be set via debconf (just as e.g. default paper format in /etc/papersize) Thank you for considering it. Bye, Joerg -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#726755: Please include High-performane-patch
Package: openssh Version: 1:6.2p2-6 Severity: wishlist Tags: patch Hello, I would appreciate if you consider include the high performance patches for openssh from http://www.psc.edu/index.php/hpn-ssh It allow significantly better throughput for high-bandwidth networks. A detailed discussion of the background and rationales is available on that site, to.. Bye, Joerg signature.asc Description: Digital signature
Bug#746500: Please add nfs support
Package: ipxe-qemu Version: 1.0.0+git-2013.c3d1e78-2 Severity: wishlist Hello, please consider adding nfs support, at least for the qemu-virtio rom, by adding #define DOWNLOAD_PROTO_NFS to config/local/general.h as described at http://ipxe.org/gsoc/nfs We use it here, space is not an issues when it is not burned into a real rom, and it is an available upstream feature. Thanks, Joerg signature.asc Description: Digital signature
Bug#747197: Please move capituils to this package, too
Package: libcapi20-3 Version: 1:3.27-1 Severity: wishlist Dear maintainer, Not sure what package to file the bug against, please excuse if it is the wrong one. I would appreciate the capiutils package (basically, the capiinit and the capiinfo binary) would move from source isdnutils to the same source as libcapi20. Currently, capiutils version-depends on a specific version of libcapi, besides, they are very closely related to the library, so having them generated by the same package would help to keep them consistent. Thank you for considering it. Bye, Joerg signature.asc Description: Digital signature
Bug#553426: #553426 - choose-session option missing with no upgrade path
On Wed, May 14, 2014 at 05:41:28PM +0100, althaser wrote: > > Could you please still reproduce this issue with newer gnome-session > version like 3.4.2.1-4 or 3.8.4-4 ? Actually, not only but also for this bug I moved off gnome completely. So I am sorry to say I do not have any machine with gnome to test it. However, if you have a gnome machine, it should be relatively easy to see if gnome-session supports a choose-session option. Bye, Joerg signature.asc Description: Digital signature
Bug#698040: Filenames with spaces need quotes
Package: grub-imageboot Version: 0.6 Hello, in order for grub to find image names with spaces, they need to be quoted. E.g. TrueCrypt Rescue Disk.iso is not found. When quotes around the file name do not have other side effects (e.g. on filenames containing quotes) they could be enabled by default. Bye, Joerg signature.asc Description: Digital signature
Bug#696926: dovecot update breaks with tcpwrap
Package: dovecot-core Version: 1:2.1.7-6 Severity: grave Hello, the new version 1:2.1.7-6 lacks support for tcpwrappers. Despite the mention in the NEWS file, it does not contain a /usr/lib/dovecot/tcpwrap and is obviously not compiled with tcpwrap support: dovecot --build-options Build options: ioloop=epoll notify=inotify ipv6 openssl io_block_size=8192 Mail storages: shared mdbox sdbox maildir mbox cydir imapc pop3c raw SQL driver plugins: mysql postgresql sqlite Passdb: checkpassword ldap pam passwd passwd-file shadow sql Userdb: checkpassword ldap(plugin) nss passwd prefetch passwd-file sql This breaks existing installations, as it even fails to install. Thanks for fixing it. Bye, Joerg signature.asc Description: Digital signature
Bug#696928: Please provide IPv6 MX for mail connectivity
Package: bugs.debian.org Severity: wishlist Hello, as I am owning an IPv6-only sytem, I have a vital interest that I am able to submit bug report also from that system. Today (29.12.2012) there is no MX with an record available. Thank you for setting up one. Bye, Joerg -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#696926: dovecot update breaks with tcpwrap
On Sat, Dec 29, 2012 at 02:29:01PM +0100, Julien Cristau wrote: > > This breaks existing installations, as it even fails to install. > > > This is on purpose. Jaldhar, I guess the mention of tcp wrapper support Can you please add a bit of explanation what would be bad about tcpwrapper support? It does not hurt, adds <10k to the package, and in case of dovecot needs to be explicitly enabled in the configuration. Besides, silently breaking existing installation on upgrade without mentioning is a bug, not a wishlist item. Bye, Joerg signature.asc Description: Digital signature
Bug#696926: dovecot update breaks with tcpwrap
On Sat, Dec 29, 2012 at 05:42:42PM +0100, Julien Cristau wrote: > > It does not hurt, adds <10k to the package, and in case of > > dovecot needs to be explicitly enabled in the configuration. > > > It's a new feature, and as such not appropriate during a freeze. I I have to apologize, I just noticed I compiled my own version with tcpwrappers enabled. As such, wishlist is appropriate. Sorry for the confusion. > don't know that it doesn't hurt, and don't want to go read the code to > find out. I am a bit surprised that tcpwrappers are not common knowledge around network daemons. Anyway, I would encourage reading, the code is really small. > > > You're running unstable, you can deal with breakage. It's mentioned JFTR, I am following testing, which dovecot entered recently. So I am looking forward to a native debian package. Happy new year anyway! Bye, Joerg -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#625020: Update
On Thu, Oct 18, 2012 at 12:36:18PM +1100, Mark Purcell wrote: > > ... > > 13:51 < tzafrir_laptop> stappers, th fix is to get a newr chan_capi from > > SVN 13:52 < tzafrir_laptop> It seems that thy don't intend to release > > anything. But thir branch 1.1.5 is old. > > Thanks Geert, > > If upstream don't intend to release then I would argue the viability of the > package in Debian is questionable. I would not mind a packages from ftp://ftp.chan-capi.org/chan-capi/chan_capi-HEAD.tgz , which works with the current asterisk in testing. Best regards, Joerg signature.asc Description: Digital signature
Bug#656320: Please make $RepeatedMsgReduction on default
Package: rsyslog Version: 5.8.6-1 Severity: wishlist Hello, for increased compatibility with sysklogd could you please consider adding a line $RepeatedMsgReduction on to the default rsyslogd.conf file shipped with debian? At least, it would be great to mentioned it here even when it is commented out, e.g. # $RepeatedMsgReduction on # "Last line repeated n times" feature The is the default behaviour of the traaditional syslog daemon, and I would appreciate to see it with rsyslogd, too. The documentation for this feature hides at http://rsyslog.com/doc/rsconf1_repeatedmsgreduction.html Bye, Joerg signature.asc Description: Digital signature
Bug#656320: Please make $RepeatedMsgReduction on default
On Wed, Jan 18, 2012 at 02:24:52PM +0100, Michael Biebl wrote: > > rsyslog has rate limiting on by default [1] (which works a bit Rate limiting is not my main concern, but thanks for mentioning it. > differently then message reduction, though). RepeatedMsgReduction was > turned off by default as it has some unpleasant side effects (like > regular wakeups etc). rate limiting has it's own share of problems so Regular wakeups are also not really a special case in my setup, though I take the point of having safe defaults is a good thing(tm) > > compatibility with the old sysklogd daemon alone is not enough reason in > this case to enable this feature imho. Well, for me, mentioning it commented out together with other features like $ModLoad imudp, $ModLoad imtcp,.. would be perfectly ok. IMHO the best place might be the after the line #$ModLoad immark # provides --MARK-- message capability adding #$RepeatedMsgReduction on # "Last line repeated n times" feature which for me is a compatibility feature, too. Bye, Joerg signature.asc Description: Digital signature
Bug#656320: Please make $RepeatedMsgReduction on default
On Wed, Jan 18, 2012 at 03:35:28PM +0100, Michael Biebl wrote: > On 18.01.2012 14:51, Joerg Dorchain wrote: > > Well, for me, mentioning it commented out together with other > > features like $ModLoad imudp, $ModLoad imtcp,.. would be perfectly > > ok. > > IMHO the best place might be the after the line > > #$ModLoad immark # provides --MARK-- message capability > > adding > > #$RepeatedMsgReduction on # "Last line repeated n times" feature > > I could be convinced to do that :-) Let's try some great magic words: Thank you! Bye, Joerg signature.asc Description: Digital signature
Bug#533031: Still can't start gmplayer
Hello, just to mention it, this bug is still present with mplayer-skin-blue 1.6-2 and mplayer-gui 2:1.0~rc4.dfsg1+svn34540-1 Bye, Joerg signature.asc Description: Digital signature
Bug#685840: squirrelmail gives Undefined variable: boxes and does not show any folders
Package: squirrelmail Version: 2:1.4.23~svn20120406-1 Hi, I recently started using dovecot as imap server (used be uw-imap). With squirrelmail on localhost, I cannot subscribe to any folders. When I click on the folders button, it just shows "No folders found". In the apache error log I find PHP Notice: Undefined variable: boxes in /usr/share/squirrelmail/functions/imap_mailbox.php on line 854 PHP Notice: Undefined variable: boxes in /usr/share/squirrelmail/functions/imap_mailbox.php on line 858 The only folder I see is INBOX. Bye, Joerg signature.asc Description: Digital signature
Bug#685850: Please compile dovecot --with-libwrap
Package: dovecot-core
Version: 1:2.1.7-2
Hi,
in order be be able to use access control via
/etc/hosts.{allow|deny}, please compile dovecot with the
--with-libwrap configure option.
Direct support for starting via inetd and tcpd has been dropped,
but direct support for libwrap has been added.
See also mailthread at
http://www.dovecot.org/list/dovecot/2012-July/067277.html
Bye,
Joerg
signature.asc
Description: Digital signature
Bug#685850: dovecot and libwrap
Hi,
i got an addendum. Libwrap use is actually documented in the wiki
at http://wiki2.dovecot.org/LoginProcess
(group = $default_login_user does not work)
I complied dovecot successfully by just adding --with-libwrap to
the configure options.
However, a config file like the one below, with some comments
would probably help
# 10-tcpwrapper.conf
#
# service name for hosts.{allow|deny} are those defined as
# inet_listener in master.conf
#
login_access_sockets = tcpwrap
service tcpwrap {
unix_listener login/tcpwrap {
#group = $default_login_user
mode = 0600
user = $default_login_user
}
}
This is in works-for-me state now ;-)
Bye,
Joerg
signature.asc
Description: Digital signature
Bug#683181: capiinit fails for fcpci
Package: capiutils Version: 1:3.25+dfsg1-1 Hello, following testing I just upgraded from the previous capiutils to version 1:3.25+dfsg1-1. With this, the package has an install error, as /etc/init.d/capiutils start gives an error. Following that, it shows that capiinit fails: # capiinit ERROR: controller 1: firmware file "-" not found ERROR: failed to load firmware for controller 1 driver fcpci name fcpci-ec80-16 The line in /etc/isdn/capi.conf is unchanged since ages and reads fcpci - - - - - - Meaning that everything is autoconfigured and no firmware is needed. The running kernel is linus's 3.2.0 Best regards, Joerg signature.asc Description: Digital signature
Bug#683181: Works with kernel 3.5
Hi, I upgraded to kernel version 3.5. With this, capiinit works as expected. Bye, Joerg signature.asc Description: Digital signature
Bug#702245: Thinkpad T61p cannot detect link, too
Hello, same here for a thinkpad T61. ethtool show output whenI do ifconfig eth0 down, but nothing when the interface is up and a cable plugged in. Bye, Joerg signature.asc Description: Digital signature
Bug#510476: linuxcallrouter: changing back from ITP to RFP
retitle 510476 ITP: linuxcallrouter -- an ISDN based PBX for Linux owner 510476 ! thanks -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#572516: Please make dsniff handle ppp interfaces
Package: dsniff
Version: 2.4b1+debian
Severity: wishlist
Tags: patch, upstream
Hello,
e.g. tcpkill does not handle ppp interfaces. The following small
patch adds support for it (and others). Please include.
Bye,
Joerg
--- pcaputil.c.orig 2010-03-04 17:15:47.0 +0100
+++ pcaputil.c 2010-03-04 17:17:21.0 +0100
@@ -52,6 +52,9 @@
case DLT_NULL:
offset = 4;
break;
+ case DLT_LINUX_SLL: /* e.g. ppp */
+ offset = 16;
+ break;
default:
warnx("unsupported datalink type");
break;
signature.asc
Description: Digital signature
Bug#630346: Please include minissdp support
Package: minidlna Version: 1.0.18-5 Severity: wishlist Tags: upstream, patch Hello, a patch for support of minissdp (also in the debian distribution) exist by the author of it at http://miniupnp.free.fr/files/minidlna_1.0.18_minissdp1.patch Please consider including it. Bye, Joerg signature.asc Description: Digital signature
Bug#637332: resume now needs explicit resume= option
Package: uswsusp
Version: 1.0+20110509-1.1
Hi,
with this version, the initramfs script has changed so it needs
an explicit resume= boot parameter. I have several issues
with that:
- At least, it should be documented, as all of a sudden resume
does not work anymore as it used to be without changes by the user.
- If you do not want to resume, the boot paramter noresume is enough.
- This duplicates information from /etc/uswsusp.conf. Bug #544641
contains a solution for this unnecessary duplication.
- Actually, just in case the root and the swap partition are on
different volume groups, there should be an additional script
in local-top that runs before lvm2 and sets the resume variable
to the configured partition. See below.
Bye,
Joerg
# As /usr/share/initramfs-tools/scripts/local-top/uswsusp
#!/bin/sh
PREREQ="" # Actually, this must run before lvm2 to make sense
prereqs()
{
echo "$PREREQ"
}
case $1 in
# get pre-requisites
prereqs)
prereqs
exit 0
;;
esac
# From /usr/share/initramfs-tools/hooks/uswsusp
# Works, as scripts are sourced
resume=$(sed -rn 's/^resume device[[:space:]]+[:=][[:space:]]+// p'
/etc/uswsusp.conf)
signature.asc
Description: Digital signature
Bug#637856: /run/lock should be owned by uucp
Package: initscripts Version: 2.88dsf-13.11 Hi, there are programs that stat() /var/lock to test permissions instead of actually trying to create lockfiles, the java rxtx library for example. Before making this symlink to the /run/lock tmpfs, ownership was preserved on harddisk. A simple chgrp uucp /run/lock || true after creation/mounting of /run/lock solves this and does not hurt programmes just trying to write the lockfile. A more general apporach is also welcome. Bye, Joerg signature.asc Description: Digital signature
Bug#637856: /run/lock should be owned by uucp
On Mon, Aug 15, 2011 at 10:15:29AM +0100, Roger Leigh wrote:
> On Mon, Aug 15, 2011 at 09:43:12AM +0200, Joerg Dorchain wrote:
> > there are programs that stat() /var/lock to test permissions
> > instead of actually trying to create lockfiles, the java rxtx
> > library for example. Before making this symlink to the /run/lock
> > tmpfs, ownership was preserved on harddisk.
>
> I think the check is broken. They should just create the lockfile
Well, there are issues with setuid()-binaries that need to do
their own access checks. A more common example would be the
sendmail binary.
> Regarding the ownership change (I assume here you changed it from
> root:root to root:uucp?), this has never been supported. And in
> any case, /run/lock and /var/lock are writable by anyone at
> present--the ownership should not matter:
Yes, for a certain point of view you can argue that the
application is broken and that this is just a bad excuse for a
workaround of closed source sofware.
On the other hand, following the first paradigma of operations
("It has to work") this is the easiest place for a change that
makes it work with almost no side effects.
IMHO it would be a nice little feature of debian to be able to
cope with commercial software of certain quality.
>
> Why uucp? As shown above, the user and group owning the directory
> should not matter--it's globally writable and has the sticky-bit set.
As uucp is traditionally (yes, that means it is more than a
decade old and nobody but but long and greyhaired grandfathers
remember how it came to that) the user who uses lockfiles.
Specifically, that is what I need for my application to work.
I don't mind to have this configurable somewhere, e.g. yet
another rcS-variable would still be better than editing the
init.d script directly.
>
> > A more general apporach is also welcome.
>
> We have been discussing moving to having a "lock" group as used by
> other distributions, but this hasn't happened yet. /run/lock
> would then be run:lock 02775 I think. Programs creating lockfiles
> would then need to be running/started as root or setgid lock.
No problem with that. IIRC this is even supported by the rxtx
library my questionable binary uses. This would also separate the
locking functionally from the rest of the uucp stuff. setgid lock
would not even be necessary, just having lock amongst the
additional groups of the calling user would be sufficient.
I would still propose mode 3775, though.
On the other hand, uucp style lockfiles are typically used for
accessing devices owned by the dialout group.
Bye,
Joerg
signature.asc
Description: Digital signature
Bug#637856: /run/lock should be owned by uucp
On Mon, Aug 15, 2011 at 02:59:04PM +0100, Roger Leigh wrote: > > > I think the check is broken. They should just create the lockfile > > > > Well, there are issues with setuid()-binaries that need to do > > their own access checks. A more common example would be the > > sendmail binary. > > I'm not sure I follow. A stat(2) on a symlink is equivalent to a > stat(2) on the pointed-to object, unless you use lstat(2). Since > the directory is world-writable, any access checks will always succeed. > The introduction of the /var/lock->/var/run symlink will not change > the behaviour of stat(2). I am under the impression that these applications check for group permissions, and consider a world-writeable directory unsafe. I have only looked at strace output, not disassembled code. > > I don't think this is an issue restricted to Debian--it is presumably > already broken on most other distributions as well? This isn't a > recent change, or a change made in isolation. AFAICT it was made > for FHS compliance reasons--nowadays /var/lock is used for a lot > more than just uucp, so expecting it to be owned by the uucp group > is, I think, unrealistic. root:root or root:lock are the way things > have gone. I use this application roughly every other week, on a machine following testing, so I am quite sure the change happend in the last few weeks. As said, root:lock ownership with group-write permissions is probably ok. > > > We don't currently permit changing of the ownership; the permissions > are entirely configurable in /etc/fstab. You could use the tmpfs > uid= and gid=options to set the uucp group there. Thanks for that hint. > We don't explicitly > cater for customisation here because the admin should never need > to change the ownership of these FHS-specified locations. Well, if all developers would follow the FHS, it would be great. > > > No problem with that. IIRC this is even supported by the rxtx > > library my questionable binary uses. This would also separate the > > locking functionally from the rest of the uucp stuff. setgid lock > > would not even be necessary, just having lock amongst the > > additional groups of the calling user would be sufficient. > > I would still propose mode 3775, though. > > > > On the other hand, uucp style lockfiles are typically used for > > accessing devices owned by the dialout group. > > The liblockdev library exists to lock ttyS devices portably; any > program creating uucp-stype LCK..* files under /var/lock should be > using it. I do not see a realistic chance to have this application changed. > > Looking to the future, Linux permits the use of flock(2) on devices, > so the reality is that "lockfiles" are entirely redundant when you > can lock the device directly using an proper kernel-provided > advisory lock. This is much more robust. When I have time, I'll be > looking at making liblockdev use flock directly. I completely agree. However, this is an application from the past, and not open source. Nevertheless I have not much choice but using it. > > From what I've just been reading on RXTX file locks, it's rather > broken. > (including http://rxtx.qbang.org/wiki/index.php/Trouble_shooting) > I think that, first and foremost, RXTX needs fixing. It can use > liblockdev, which I see has already been suggested looking at list > archives with google. It's basically using broken assumptions and > broken checks. Even if librxtx is ever fixed, that fixed version needs to make it into the closed source apps of a commercial vendor. It will take years. > I don't think that breakage in a Java library can > really warrant changes to the default ownership of an FHS-standardised > directory. As said, root:lock would also be ok. Even if you are not moving away from FHS for policy reasons, just do not make it unnecessarly hard to run with slight derivations from it. Besides, I am sure it is not the only pre-FHS application out there. Bye, Joerg signature.asc Description: Digital signature
Bug#645666: libcapi20 breaks avm kernel module
Package: libcapi20-3 Version: 1:3.9.20060704+dfsg.2-11 Hi, following testing, the recent upgrade to 1:3.9.20060704+dfsg.2-11 breaks the avm kernel module for fcpci with kernel 3.0 (FYI, the module is self compiled using patches from archlinux) - one issue is that the package is looking for capifs, which does not exist anymore (cfr https://bugs.launchpad.net/ubuntu/+source/isdnutils/+bug/873892) - even with capifs mounting attempt removed, capiinit works, but capiinfo does not show the card. strace shows no attempt to open /dev/capi20 This bug affects also asterisk chan_capi and any other application using libcapi. My current workaround is to use version 1:3.9.20060704+dfsg.2-4.1 from stable. Bye, Joerg signature.asc Description: Digital signature
Bug#641257: apt-cacher-ng should support prefer IPv4/IPv6 setting
Package: apt-cacher-ng Version: 0.6.6-1 Severity: wishlist Hello, I'd like to have an option for apt-cacher-ng to be able to choose the preference for the protocol of outgoing connections to be IPv4 or IPv6, with the default, i.e. not set explicitly, being the system default, i.e. the current behaviour. Option could look like: PreferredProtocol: IPv4|Ipv6|default Uses include an IPv4/IPv6 application layer gateway or situations where both protocols are available, but one is preferred for some reason (bandwidth differences, billing, policies etc.) Bye, Joerg signature.asc Description: Digital signature
Bug#639872: /etc/init.d/sysklogd reload-or-restart does not work
Package: sysklogd Version: 1.5-6.1 Hello, The /etc/init.d/sysklogd reload-or-restart clause used in /etc/cron.weekly/sysklogd does not work. Precisely, I have the folowing effect: At system startup, all is fine: # lsof /var/log/mail/* COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME syslogd 30307 syslog6w REG 254,5 4734 423 /var/log/mail/mail.log syslogd 30307 syslog9w REG 254,5 4734 75 /var/log/mail/mail.info syslogd 30307 syslog 10w REG 254,5 75 783 /var/log/mail/mail.warn syslogd 30307 syslog 11w REG 254,5 75 765 /var/log/mail/mail.err Then: # /etc/init.d/sysklogd reload-or-restart Reloading system log daemon # lsof /var/log/mail/* COMMANDPID USER FD TYPE DEVICE SIZE/OFF NODE NAME Only # /etc/init.d/sysklogd restart Restarting system log daemon # lsof /var/log/mail/* COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME syslogd 30470 syslog6w REG 254,5 4734 423 /var/log/mail/mail.log syslogd 30470 syslog9w REG 254,5 4734 75 /var/log/mail/mail.info syslogd 30470 syslog 10w REG 254,5 75 783 /var/log/mail/mail.warn syslogd 30470 syslog 11w REG 254,5 75 765 /var/log/mail/mail.err makes it work again. First suspicion is that this effect is related to the newly run-as-user feature. Bye, Joerg signature.asc Description: Digital signature
Bug#641257: apt-cacher-ng should support prefer IPv4/IPv6 setting
On Thu, Sep 15, 2011 at 03:51:31PM +0200, Eduard Bloch wrote: > Ok. This is how it will look like: > > # Specifies the IP protocol families to use for remote connections. Order does > # matter, first specified are considered first. Possible combinations: > # v6 v4 > # v4 v6 > # v6 > # v4 > # (empty or not set: use system default) > # > # ConnectProto: v6 v4 Great! Thank you for the quick response. Bye, Joerg signature.asc Description: Digital signature
Bug#642228: gwhois does not lookup .co correctly
Package: gwhois Version: 20100728 Hello, # gwhois g.co Process query: 'g.co' Querying www.nic.co with cgi. 404 Not FoundNot Found The requested URL /pls/dominio/Pwhois.DatosDominio was not found on this server.Apache/2.2.3 (CentOS) Server at www.nic.co Port 443 whois.nic.co. works, using the regular whois protocol. Bye, Joerg signature.asc Description: Digital signature
Bug#476356: After upgrade torrus fails with DB_VERSION_MISMATCH
On Sun, Sep 25, 2011 at 01:03:02PM +0200, Bernhard Schmidt wrote: > Debian. I've tried the BDB Upgrade script Marc Haber wrote with your > assistance, but it does not work (at least not for the upgrade > squeeze->wheezy (4.8 -> 5.1). [...] > > Since I do not have much knowledge about BDB, I'm at loss here. Do > you have any idea? This script is a workaround for the lack of a version independant dbX.X-recover. You can try to run db5.1_recover in the /var/lib/torrus/db/ directory. You may wish to read the man page, though. Please stop anything acessing the db before, which inlcudes torrus itself, but also apache for its embedded perl module. Best regards, Joerg signature.asc Description: Digital signature
Bug#715486: Please seperate mysql dependancy for snmptrapd
Package: snmpd Version: 5.7.2~dfsg-8 Severity: wishlist Hello, I am running debian of a small box (raspberry pi). Since the recent upgrade, the snmpd package depends on a mysql client library. The mysql client library is not needed for the snmpd itself, just for snmptrapd. Addionally, snmptrapd is used far less often than snmpd, so splitting this into a separate package might make sense under this aspect, too. So please consider shipping a package with snmpd (without the mysql dependency) and snmptrapd (which for me would be ok to depend on snmpd package) Bye, Joerg signature.asc Description: Digital signature
Bug#655086: this seems to be a problem in asterisk
On Sun, Feb 26, 2012 at 12:15:32PM +0800, Rolf Leggewie wrote: > > I consulted with my co-maintainer and he concluded this was a problem in > asterisk, possibly linking to a module instead of libcapi20 itself. I'm > not sure that's indeed the case, but he's the libcapi guy between us. Well, the actual module chain is asterisk dlopens chan_capi.so links libcapi20.so dlopens lib_capi_mod_std.so If you are going to investigate in that direction, please notice the chan_capi package, too. The effect of this bug report comes from the fact that lib_capi_mod_std.so references a symbol defined in libcapi20.so. The proposed patch just prolongs this chain at the end to asterisk dlopens chan_capi.so links libcapi20.so dlopens lib_capi_mod_std.so links libcapi20.so E.g. the package built with the proposed patch shows # ldd /usr/lib/capi/lib_capi_mod_std.so linux-gate.so.1 => (0xb7721000) libcapi20.so.3 => /usr/lib/libcapi20.so.3 (0xb76f9000) libc.so.6 => /lib/i386-linux-gnu/i686/cmov/libc.so.6 (0xb759c000) libdl.so.2 => /lib/i386-linux-gnu/i686/cmov/libdl.so.2 (0xb7597000) /lib/ld-linux.so.2 (0xb7722000) In the resulting package there are no .la files contained. Thanks for digging into this complicated shared library problem. Best regards, Joerg signature.asc Description: Digital signature
Bug#645666: libcapi20 breaks avm kernel module
On Sun, Nov 27, 2011 at 01:54:38PM +0800, Rolf Leggewie wrote: > On 18.10.2011 01:54, Joerg Dorchain wrote: > > - even with capifs mounting attempt removed, capiinit works, but > > capiinfo does not show the card. strace shows no attempt to > > open /dev/capi20 > > Jörg, > > thank you for your report. About the problem with capiinfo, I am able > to reproduce and confirm that. Can you please verify if "cd > /usr/lib/capi/ ; for file in lib_capi_mod_fritzbox.so > lib_capi_mod_rcapi.so lib_capi_mod_std.so; do ln -sv $file.0.0.0 > $file;done;cd -" run as root fixes the problem? This does indeed fix the capiinfo problem. However, now chan_capi of asterisk has a problem: /usr/sbin/asterisk: symbol lookup error: /usr/lib/capi/lib_capi_mod_std.so: undefined symbol: processMessage There are several bug reports and fixes around this, the first I found is https://bugzilla.novell.com/show_bug.cgi?id=475808 (llok for the patch) I'd be happy to test a fixed package ;) Bye, Joerg signature.asc Description: Digital signature
Bug#645666: Patch link - any news
retitle 645666 libcapi has symbol lookup error "undefined symbol: processMessage" tags 645666 +patch thanks Hello, just wondering if there are any news about this problem. I also changed the title in the hope to make it clearer. The patch pasted below (watch for whitespace errors) is at https://bugzillafiles.novell.org/attachment.cgi?id=460626 and helps the symbol lookup error: /usr/lib/capi/lib_capi_mod_std.so: undefined symbol: processMessage message. Please include it in order to make the package usable again. Bye, Joerg Index: isdn4k-utils/capi20/Makefile.am === --- isdn4k-utils.orig/capi20/Makefile.am +++ isdn4k-utils/capi20/Makefile.am @@ -19,14 +19,17 @@ lib_capi_mod_std_la_SOURCES = capi_mod_std.c lib_capi_mod_std_la_CFLAGS = -fno-strict-aliasing lib_capi_mod_std_la_LDFLAGS = -shared +lib_capi_mod_std_la_LIBADD = libcapi20.la lib_capi_mod_fritzbox_la_SOURCES = capi_mod_fritzbox.c lib_capi_mod_fritzbox_la_CFLAGS = -fno-strict-aliasing lib_capi_mod_fritzbox_la_LDFLAGS = -shared +lib_capi_mod_fritzbox_la_LIBADD = libcapi20.la lib_capi_mod_rcapi_la_SOURCES = capi_mod_rcapi.c lib_capi_mod_rcapi_la_CFLAGS = -fno-strict-aliasing lib_capi_mod_rcapi_la_LDFLAGS = -shared +lib_capi_mod_rcapi_la_LIBADD = libcapi20.la libcapi20dyn_a_SOURCES = capidyn.c libcapi20dyn_a_CFLAGS = -fPIC signature.asc Description: Digital signature
Bug#655086: libcapi has symbol lookup error "undefined symbol: processMessage"
Package: libcapi20-3 Version: 1:3.9.20060704+dfsg.2-12 Tags: patch Hello, as requested by you in bug #645666, here is a new bug report for the problem. Using chan_capi with asterisk shows the problem: /usr/sbin/asterisk: symbol lookup error: /usr/lib/capi/lib_capi_mod_std.so: undefined symbol: processMessage To my understanding, there is a circular symbol dependancy between libcapi20.so and lib_capi_mod_std.so (and the other libs in /usr/lib/capi/, but I am only using local capi) One way to solve this is to give libcapi twice to the linker at appropriate positions (cfr. -lgcc appearing several times during gcc build) A proposal that works for me, i.e. the package builds and works properly, is mentioned at https://bugzillafiles.novell.org/attachment.cgi?id=460626 (in clear text below) As you mentioned in the previous bug report, you don't like touching the .la files. You do not need to ship them in any packages, just use them at build time, as the lib_capi_mod* files are sort of plugins to libcapi20. The goal is to have the linker for lib_capi_mod_std.so resolving symbols via libcapi20.so. By using the .la files you do not have to give the full path for it, which might be tricky as it is different at built time and at package using time, so the .la files provide a convinience here instead of using libcapi20.so and -R. For alternative ways the description of the -rpath and -rpath-link in the ld documetation might be helpful. However, to my understanding the proposed usage of the .la files is ok with http://www.netfort.gr.jp/~dancer/column/libpkg-guide/libpkg-guide.html Nevertheless, any alternative solution is very welcome. I would appreciate if for pragmatic reasons this fix could be included to make it work for the time being, and, in case it conflicts with the long run policy, have another bug (cloned, e.g. ;-) to implement the final solution. As for the other bug reports, I am willing and able to test any packages if you wish. Bye, Joerg Index: isdn4k-utils/capi20/Makefile.am === --- isdn4k-utils.orig/capi20/Makefile.am +++ isdn4k-utils/capi20/Makefile.am @@ -19,14 +19,17 @@ lib_capi_mod_std_la_SOURCES = capi_mod_std.c lib_capi_mod_std_la_CFLAGS = -fno-strict-aliasing lib_capi_mod_std_la_LDFLAGS = -shared +lib_capi_mod_std_la_LIBADD = libcapi20.la lib_capi_mod_fritzbox_la_SOURCES = capi_mod_fritzbox.c lib_capi_mod_fritzbox_la_CFLAGS = -fno-strict-aliasing lib_capi_mod_fritzbox_la_LDFLAGS = -shared +lib_capi_mod_fritzbox_la_LIBADD = libcapi20.la lib_capi_mod_rcapi_la_SOURCES = capi_mod_rcapi.c lib_capi_mod_rcapi_la_CFLAGS = -fno-strict-aliasing lib_capi_mod_rcapi_la_LDFLAGS = -shared +lib_capi_mod_rcapi_la_LIBADD = libcapi20.la libcapi20dyn_a_SOURCES = capidyn.c libcapi20dyn_a_CFLAGS = -fPIC signature.asc Description: Digital signature
Bug#645666: Patch link - any news
On Sun, Jan 08, 2012 at 11:27:21AM +0800, Rolf Leggewie wrote: > On 06.01.2012 22:02, Joerg Dorchain wrote: > > just wondering if there are any news about this problem. I also > > changed the title in the hope to make it clearer. > > Well, in all honesty you've now added a third, separate issue to the two > you already had in your original report. "one issue, one ticket", > please. Anything else makes it very hard to deal with the problems and > track progress. With all understanding of issue tracking and bug handling (I am earning part of my living with that), from my point of view I have one problem, that my relativly critical phone application ceases to work after an upgrade, that is one ticket. If this problem (previous version works, new one does not), which I would classify as a regression, needs several fixes, well, then I expect them be handled, one after another in the worst case if they come completely unexpected. I am sorry that most of this mess concerns the upstream part of the package, but heck, I am using the debian package and not upstream directly. As we are working in a volunteer organsation, which I would interpret in a way that there is no KPI or anything linked with it with bug handling that endorses many small bugs opened and closed even faster, I would could consider it good and friendly service when you split bug reports yourself as you see them fit. I mean bug cloning and merging is available, and who knows better than the maintainer how to organise "his" bug reports. Just closing them is what I consider extremly impolite. You might want to have a look at some ITIL documentation, some principles are not totally wrong and applicable also to debian bug handling processes. On the positive side, speaking of a volunteer and not a professional organisation, I think I should be happy that there was a fast reponse. > > Please open a new ticket about about the undefined symbol issue. After > a brief look at the proposed patch I'm not sure it's appropriate. There > are suggestions to stop shipping the .la files entirely. FWIW, I have > not looked into this newly uncovered problem at all, yet. I will open a new ticket for this in the hope it will be dealt with. I tried to be as constructive as I can by providing a patch which works for me. From what I see from searches on the web chan_capi is not the only application that breaks, it is just anything that actually tries to open an isdn channel. I will further go into details in the next bug report. Bye, Joerg signature.asc Description: Digital signature
Bug#651691: uswsusp needs special kernel parameter to resume
Package: uswsusp Version: 1.0+20110509-2 Hi, with the current version of initramfs-tools, uswsuspend needs a paramter to resume, namely the current swap partition, e.g. resume=/dev/swap. This this different from previous behaviour, where it needed a special parameter (noresume) in order to not to resume. This change in behaviour is not documented. (e.g. in the README. debian) Besides, the same information is also available in /etc/uswsusp.conf (resume device), and both need to match in order for resume to work. For me, the general assumtion that you would like to resume when the computer has been shutdown via suspend is much more likely than not to resume, so I'd like to see the previous behaviour (booting with resume and no special parameter) again. Even if a parameter is needed, having to update the same information at two location manually without documentation is error prone and ugly. Reuse the information in /etc/uswsusp.conf (or from debconf) for generating usefull defaults for bootloader configuration would be more adeuqate. The minimum would be to documented it, first step being this report. Last I'd like to repeat that resume on boot after suspend without special measures would follow the principle of least surprise and would be most preferred for me. Bye, Joerg -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#857906: iaxmodem 1.2.0~dfsg-2+b2 does not work anymore
Package: iaxmodem Version: 1.2.0~dfsg-2+b2 Hello, following testing, the new version 1.2.0~dfsg-2+b2 does not work anymore: - It does not register itself to the aix server (here: asterisk 1:13.14.0~dfsg-1) anymore - When you configure the relation statically, faxes are not received. No page arrives, the modem decodes only garbage. - The stop initscript action does not work. For the time being, downgrading to 1.2.0~dfsg-2 is a workaround. Bye, Joerg signature.asc Description: PGP signature
Bug#855281: userspace ct_helper_*.so modules are missing
Package: nfct Version: 1:1.4.4+snapshot20161117-4 Hello, on the search for ssdp connection track helper I learned that this is available in upstream since version 1.4.3. However I do not find any of the userspace helpers in any of the packages derived from conntrack-tools. An strace of nfct shows that it search for them in /usr/lib/conntrack-tools/, so I presume that the nfct package would be the most approriate place to include them. Please include them in order to avoid messages like # nfct add helper ftp inet tcp nfct v1.4.4: that helper is not supported: No such file or directory Bye, Joerg signature.asc Description: PGP signature
Bug#784682: Make wishlist
reopen 784682 retitle 784682 "Please backport new version to jessie" severity wishlist thanks Hello Zed, reopening this bug as a hint for people like me who ran into problems with the version of snoopy currently in stable. The dangling file descriptor that is fixed in the mentioned upstream bug report also affects co-existence with other packages, for me notably puppet. The package recompiles without change on jessie, so backport is rather trivial, just needs an upload to the backport queue. A recent version of snoopy might also fix other bugs. TIA for considering. Bye, Joerg
Bug#784682: Make wishlist
On Mon, Jun 06, 2016 at 05:02:28PM -0400, Zed Pobre wrote: > On Mon, Jun 06, 2016 at 07:41:40PM +0200, Joerg Dorchain wrote: > > Hello Zed, > > > > reopening this bug as a hint for people like me who ran into problems > > with the version of snoopy currently in stable. > > I'm afraid I am seriously considering abandoning this package at the > moment. :/ It turned out to be trivial to bypass because any user can > recompile it and preload a version that doesn't log. We moved to > auditd as a replacement. > > If you are interested in maintaining this package going forwards, I > will happily turn it over to you. In principle yes. However, technically, I am just a mere debian user, no maintainer nor uploader, so I would need help anyway. I have tried to become a maintainer, but that went unlucky because first the mentor lacking time and then me being busy. Replying from my private address, as for sure my company will not pay extra for this ;-) Bye, Joerg signature.asc Description: PGP signature
Bug#827721: apt-cacher-ng stalls on upgrade
Package: apt-cacher-ng Version: 0.9.3.1-1 Hello, on upgrading via apt-get upgrade, apt-cacher-ng tries to download something with wget. It does this with proxy settings read from apt config. Here, Acquire::http::Proxy setting points to the apt-cacher-ng itself running on the local machine. This stalls, as at this time of the upgrade cycle, the old apt-cacher-ng version has been stopped by apt-get, but the new version has not yet been started. The script tries several methods, each having a timeout at least 300 seconds. Please think of a "smarter" way of upgrading. Options include a trigger that is run at a very late stage of the upgrade, or a debconf question which proxy to use during the upgrade, or maybe something else. Bye, Joerg signature.asc Description: PGP signature
Bug#812079: apt-cacher-ng delivers outdated release files
On Wed, Jan 20, 2016 at 08:05:46PM +0100, Eduard Bloch wrote: > > > > I am following debian testing, and do updates almost every other > > day. I use apt-cacher-ng via the Acquire::http::Proxy setting in > > apt. > > > > Apt-cacher-ng seems not to expire the metadata, so doing apt-get > > update does not show any new packages, not even after a week. > Finally, I was able to track this behaviour down to a specific mirror I had setup in /etc/apt-cacher-ng/backends_debian When I use http://ftp.freenet.de/debian/ as a first line in it. strange things (i.e. keeping outdated data in the cache) happen. Other mirrors, e.g. http://ftp.uni-kl.de/debian/ work fine. I was not able to dig into difference between those two mirror servers. Bye, Joerg signature.asc Description: PGP signature
Bug#736014: iaxmodem is marked for autoremoval from testing
On Mon, Jun 30, 2014 at 09:11:09AM +0200, Joerg Dorchain wrote: > > If nobody else volunteers, I will try to care care of that. Fix > is trivial, obviously. I found that the fix for this bug is in the source (http://anonscm.debian.org/viewvc/pkg-voip/iaxmodem/trunk/debian/) for quite some time, so nothing to do but upload a new package, which I would really appreciate. Thanks for taking care of it, Joerg signature.asc Description: Digital signature
