Bug#236243: adns: last problem hcommon.c.m4
On Fri, Dec 09, 2005 at 11:07:10AM +, Ian Jackson wrote: arrangements are supposed to make sure that the timestamps are correct and I will have to check this since obviously the shipped hcommon.c was out of date. well... this might be related to the fact that i had patched both in the debian package. :) So perhaps I should delete the file in the build rule for debian, only? I havent looked in detail what the m4 does, is it platform specific? wil lthe result be the same on all linux systems? Otherwise I will enforce the delete on clean. Gruss Bernd -- (OO) -- [EMAIL PROTECTED] -- ( .. )[EMAIL PROTECTED],linux.de,debian.org} http://www.eckes.org/ o--o 1024D/E383CD7E [EMAIL PROTECTED] v:+497211603874 f:+49721151516129 (OO) When cryptography is outlawed, bayl bhgynjf jvyy unir cevinpl! -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#236243: adns: last problem hcommon.c.m4
Hello Ian, me again :) there is a last problem, hcommon.c.m4 is missing ENOTSOCK, which will cause a false regression failure. However since it is shipping hcommon.c, it depends on the file timestamps on which system it will be used. I just uploaded an 1.1-3 which adds ENOTSOCK to .c.m4. So this means no matter if the generated or shipped file is used, the tests suceed. However, as already discussed in Bug #236243 it might be good to actually remove hcommon.c on clean, and/or not ship it? Gruss Bernd -- (OO) -- [EMAIL PROTECTED] -- ( .. )[EMAIL PROTECTED],linux.de,debian.org} http://www.eckes.org/ o--o 1024D/E383CD7E [EMAIL PROTECTED] v:+497211603874 f:+49721151516129 (OO) When cryptography is outlawed, bayl bhgynjf jvyy unir cevinpl! -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#340384: net-tools: arp Bus Error
On Fri, Dec 02, 2005 at 11:00:28PM -0800, Jurij Smakov wrote: In that case the cast from sockaddr to sockaddr_in is a problem. The sockaddr structure consists of a short (2 bytes) and an array of 14 chars, so it is allowed to be aligned (and is, in fact, aligned) on the 2-byte boundary, Yes, probaly sockad should be changed to contain an array of longs instead of chars, since it is supposed to be castable I really hate C :) Gruss Bernd -- (OO) -- [EMAIL PROTECTED] -- ( .. )[EMAIL PROTECTED],linux.de,debian.org} http://www.eckes.org/ o--o 1024D/E383CD7E [EMAIL PROTECTED] v:+497211603874 f:+49721151516129 (OO) When cryptography is outlawed, bayl bhgynjf jvyy unir cevinpl! -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#340384: net-tools: arp Bus Error
tag 340384 +patch
thanks
Hello Juri,
thanks for all your help (and please say thanks to Steve). I didnt want to
rewrite all of the code, so I just changed the alocation of the structures
in the following way:
Index: arp.c
===
RCS file: /cvsroot/net-tools/net-tools/arp.c,v
retrieving revision 1.24
diff -u -r1.24 arp.c
--- arp.c 16 May 2005 04:30:17 - 1.24
+++ arp.c 4 Dec 2005 02:49:42 -
@@ -100,7 +100,8 @@
{
char host[128];
struct arpreq req;
-struct sockaddr sa;
+struct sockaddr_storage ss;
+struct sockaddr *sa;
int flags = 0;
int deleted = 0;
@@ -112,12 +113,13 @@
return (-1);
}
safe_strncpy(host, *args, (sizeof host));
-if (ap-input(0, host, sa) 0) {
+sa = (struct sockaddr *)ss;
+if (ap-input(0, host, sa) 0) {
ap-herror(host);
return (-1);
}
/* If a host has more than one address, use the correct one! */
-memcpy((char *) req.arp_pa, (char *) sa, sizeof(struct sockaddr));
+memcpy((char *) req.arp_pa, (char *) sa, sizeof(struct sockaddr));
if (hw_set)
req.arp_ha.sa_family = hw-type;
@@ -177,11 +179,11 @@
usage();
if (strcmp(*args, 255.255.255.255) != 0) {
strcpy(host, *args);
- if (ap-input(0, host, sa) 0) {
+ if (ap-input(0, host, sa) 0) {
ap-herror(host);
return (-1);
}
- memcpy((char *) req.arp_netmask, (char *) sa,
+ memcpy((char *) req.arp_netmask, (char *) sa,
sizeof(struct sockaddr));
req.arp_flags |= ATF_NETMASK;
}
@@ -266,7 +268,8 @@
{
char host[128];
struct arpreq req;
-struct sockaddr sa;
+struct sockaddr_storage ss;
+struct sockaddr *sa;
int flags;
memset((char *) req, 0, sizeof(req));
@@ -277,12 +280,13 @@
return (-1);
}
safe_strncpy(host, *args++, (sizeof host));
-if (ap-input(0, host, sa) 0) {
+sa = (struct sockaddr *)ss;
+if (ap-input(0, host, sa) 0) {
ap-herror(host);
return (-1);
}
/* If a host has more than one address, use the correct one! */
-memcpy((char *) req.arp_pa, (char *) sa, sizeof(struct sockaddr));
+memcpy((char *) req.arp_pa, (char *) sa, sizeof(struct sockaddr));
/* Fetch the hardware address. */
if (*args == NULL) {
@@ -352,11 +356,11 @@
usage();
if (strcmp(*args, 255.255.255.255) != 0) {
strcpy(host, *args);
- if (ap-input(0, host, sa) 0) {
+ if (ap-input(0, host, sa) 0) {
ap-herror(host);
return (-1);
}
- memcpy((char *) req.arp_netmask, (char *) sa,
+ memcpy((char *) req.arp_netmask, (char *) sa,
sizeof(struct sockaddr));
flags |= ATF_NETMASK;
}
@@ -525,7 +529,8 @@
static int arp_show(char *name)
{
char host[100];
-struct sockaddr sa;
+struct sockaddr_storage ss;
+struct sockaddr *sa;
char ip[100];
char hwa[100];
char mask[100];
@@ -538,14 +543,15 @@
host[0] = '\0';
+sa = (struct sockaddr *)ss;
if (name != NULL) {
/* Resolve the host name. */
safe_strncpy(host, name, (sizeof host));
- if (ap-input(0, host, sa) 0) {
+ if (ap-input(0, host, sa) 0) {
ap-herror(host);
return (-1);
}
- safe_strncpy(host, ap-sprint(sa, 1), sizeof(host));
+ safe_strncpy(host, ap-sprint(sa, 1), sizeof(host));
}
/* Open the PROCps kernel table. */
if ((fp = fopen(_PATH_PROCNET_ARP, r)) == NULL) {
@@ -581,10 +587,10 @@
if (opt_n)
hostname = ?;
else {
- if (ap-input(0, ip, sa) 0)
+ if (ap-input(0, ip, sa) 0)
hostname = ip;
else
- hostname = ap-sprint(sa, opt_n | 0x8000);
+ hostname = ap-sprint(sa, opt_n | 0x8000);
if (strcmp(hostname, ip) == 0)
hostname = ?;
}
this might not be the cleanest way, but i think it is ok this way. I hope
arp is the only afected tool: I dont have yet a sparc64 account, but
hopefully soon, and then i will test it before upload.
Gruss
Bernd
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#340384: backtrace running nonstripped
On Fri, Dec 02, 2005 at 09:26:46AM -0800, Blars Blarson wrote:
Sorry it has taken so long for me to get you this info.
thank you very much Blars for the help, this is much apreciated.
the bus error on sparc64:
#0 0x701029cc in inet_aton () from /lib/libc.so.6
#1 0x00013f84 in INET_resolve ()
#2 0x000148f4 in INET_input ()
#3 0x00011608 in arp_show ()
#4 0x00012554 in main ()
static int arp_show(char *name)
{
struct sockaddr sa;
if (ap-input(0, ip, sa) 0)
static int INET_input(int type, char *bufp, struct sockaddr *sap)
{
return (INET_resolve(bufp, (struct sockaddr_in *) sap, 1));
static int INET_resolve(char *name, struct sockaddr_in *sin, int hostfirst)
{
if (inet_aton(name, sin-sin_addr)) {
Hmmm... it looks a bit strange. Maybe it is a header or compiler problem
since you pointed out that an older version works until you recompile it.
Anybody has an idea?
Gruss
Bernd
--
(OO) -- [EMAIL PROTECTED] --
( .. )[EMAIL PROTECTED],linux.de,debian.org} http://www.eckes.org/
o--o 1024D/E383CD7E [EMAIL PROTECTED] v:+497211603874 f:+49721151516129
(OO) When cryptography is outlawed, bayl bhgynjf jvyy unir cevinpl!
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#340384: sparc64 system?
Hallo, according to db.debian.org we have two sparc machines (aric and vore) for Developers. However vore seems to be unable, and on auric I can't log in. (it wont ask me for the pubkey i have in the databse and it wont accept the password) I need a sparc64 test system to debug a bus error (#340384). Cananybody tell me where I get an account, or do the debug symbols compile and backtrace for me? And also update the availability information of vore in the db. Gruss Bernd -- (OO) -- [EMAIL PROTECTED] -- ( .. )[EMAIL PROTECTED],linux.de,debian.org} http://www.eckes.org/ o--o 1024D/E383CD7E [EMAIL PROTECTED] v:+497211603874 f:+49721151516129 (OO) When cryptography is outlawed, bayl bhgynjf jvyy unir cevinpl! -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#340384: net-tools: arp Bus Error
do you know if this was introduced in -16 (or earlier?) i.e. what was the last version which worked? #0 0x70104220 in inet_aton () from /lib/libc.so.6 #1 0x00014028 in ?? () #2 0x00014028 in ?? () if you have some time at hand you could try to debug this with a debug build, however it might also be obvious from source where the unaligned access happens, will check that at the weekend. Thanks for your report. Bernd -- (OO) -- [EMAIL PROTECTED] -- ( .. )[EMAIL PROTECTED],linux.de,debian.org} http://www.eckes.org/ o--o 1024D/E383CD7E [EMAIL PROTECTED] v:+497211603874 f:+49721151516129 (OO) When cryptography is outlawed, bayl bhgynjf jvyy unir cevinpl! -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#328602: very old package, should this be removed?
reassign 328602 wnpp retitle 328602 RFA: redir needs new maintainer severity 328602 normal thanks On Fri, Sep 16, 2005 at 12:32:32PM +0200, Marc 'HE' Brockschmidt wrote: After 7 days without answer from you (the maintainer) we will reassign this bug to either WNPP (in case we propose to orphan it) or ftp.debian.org (in case we propose to remove it). I have uploaded a new version 2.1-3. This package is still usefull and works. There is a new upstream version which should be packaged, I am looking for a new Maintainer. Gruss Bernd -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#322541: patch to fix'
On Sat, Aug 13, 2005 at 01:06:45AM -0400, Ari Pollak wrote: I'll be doing a 7-day-delayed NMU of this, with only the attached patch as the changes. thanks. Please observe version numbering. It is the second NMU Bernd -- (OO) -- [EMAIL PROTECTED] -- ( .. )[EMAIL PROTECTED],linux.de,debian.org} http://www.eckes.org/ o--o 1024D/E383CD7E [EMAIL PROTECTED] v:+497211603874 f:+49721151516129 (OO) When cryptography is outlawed, bayl bhgynjf jvyy unir cevinpl! -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#273338: lost some mails
Re NMU of symlinks: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=273338 On Sat, May 14, 2005 at 05:49:24AM -0500, Bill Allombert wrote: if anybody send me mails in the last two month to ecki at debian.org the chances are high I havent received it due to some missconfig on my site. Please resend. Hello Bernd, That might explain we got no answer from you about the package 'symlinks'... In any event, I have NMUed this package. Please find more information in the bug log for #273338. thanks for the pointer, yes indeed, i was totally unaware of that :) I will not upload a maintainer version before the release (unless somebody asks me to) Thans Mark for your Help, also. Greetings Bernd -- (OO) -- [EMAIL PROTECTED] -- ( .. ) [EMAIL PROTECTED],linux.de,debian.org} http://www.eckes.org/ o--o 1024D/E383CD7E [EMAIL PROTECTED] v:+497211603874 f:+49721151516129 (OO) When cryptography is outlawed, bayl bhgynjf jvyy unir cevinpl! -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
