Bug#444970: squirrelmail: Can't send attachment either
Hi, I can't send attachment either, and running PHP4. Downgrading back to 1.4.10a-2 fixes the problem so something has happened on the way to 1.4.11-1. - Toni -- Toni Timonen toni dot timonen at iki dot fi NP Solutions Ltd Helsinki University of Technology Department of Engineering Physics and Mathematics signature.asc Description: Digital signature
Bug#409710: ipython: Doesn't handle utf8 strings correctly
Package: ipython Version: 0.7.3-2 Severity: normal Tags: patch Behaviour in python interpreter: Python 2.4.4 (#2, Jan 13 2007, 18:15:24) Type help, copyright, credits or license for more information. u'Böö' u'B\xf6\xf6' Behaviour in ipython: Python 2.4.4 (#2, Jan 13 2007, 18:15:24) IPython 0.7.3 -- An enhanced Interactive Python. In [1]: u'Böö' Out[1]: u'B\xc3\xb6\xc3\xb6' In [2]: print u'Böö' Böö Expected behaviour: Same as in regular python interpreter. Behaviour after the patch: Python 2.4.4 (#2, Jan 13 2007, 18:15:24) IPython 0.7.3 -- An enhanced Interactive Python. In [1]: u'Böö' Out[1]: u'B\xf6\xf6' In [2]: print u'Böö' Böö -- System Information: Debian Release: 4.0 APT prefers unstable APT policy: (500, 'unstable') Architecture: amd64 (x86_64) Shell: /bin/sh linked to /bin/bash Kernel: Linux 2.6.18-3-amd64 Locale: LANG=fi_FI.UTF-8, LC_CTYPE=fi_FI.UTF-8 (charmap=UTF-8) Versions of packages ipython depends on: ii python2.4.4-2An interactive high-level object-o ii python-central0.5.12 register and build utility for Pyt Versions of packages ipython recommends: ii python-numeric24.2-7 Numerical (matrix-oriented) Mathem pn python-numeric-extnone (no description available) pn python-pexpectnone (no description available) -- no debconf information -- Toni Timonen toni dot timonen at iki dot fi NP Solutions Ltd Helsinki University of Technology Department of Engineering Physics and Mathematicsdiff -rub ipython-0.7.3.orig/IPython/iplib.py ipython-0.7.3/IPython/iplib.py --- ipython-0.7.3.orig/IPython/iplib.py 2006-12-11 18:52:16.0 +0200 +++ ipython-0.7.3/IPython/iplib.py 2007-02-04 23:37:03.0 +0200 @@ -1965,7 +1965,7 @@ try: -line = raw_input_original(prompt) +line = raw_input_original(prompt).decode(sys.stdin.encoding) except ValueError: warn(\n\nYou or a %run:ed script called sys.stdin.close() or sys.stdout.close()!\nExiting IPython!) self.exit_now = True signature.asc Description: Digital signature
Bug#379764: sysstat: iostat fails to show stats of one partition
Package: sysstat Version: 6.1.3-1 Severity: normal Tags: patch Hi, getting iostat of one partition fails. eg. [EMAIL PROTECTED]:~/work/m/tmp/sysstat-6.1.3$ ./iostat -p /dev/hda1 Linux 2.6.15-1-686 (sd033) 25/07/06 avg-cpu: %user %nice %system %iowait %steal %idle 10.860.64 35.550.720.00 52.23 Device:tps Blk_read/s Blk_wrtn/s Blk_read Blk_wrtn [EMAIL PROTECTED]:~/work/m/tmp/sysstat-6.1.3$ (but running 'iostat -p /dev/hda' gives full output). A patch to fix this issue is attached. Behaviour after the patch: [EMAIL PROTECTED]:~/work/m/tmp/sysstat-6.1.3-patched$ ./iostat -p /dev/hda1 Linux 2.6.15-1-686 (sd033) 07/25/06 avg-cpu: %user %nice %system %iowait %steal %idle 10.860.64 35.550.720.00 52.23 Device:tps Blk_read/s Blk_wrtn/s Blk_read Blk_wrtn hda1 0.00 0.00 0.00 7993 9460 [EMAIL PROTECTED]:~/work/m/tmp/sysstat-6.1.3-patched$ -- Toni Timonen 040-5111863,GPG 0x7984A4FD --- sysstat-6.1.3/iostat.c 2006-05-24 16:24:07.0 +0300 +++ sysstat-6.1.3-patched/iostat.c 2006-07-25 15:14:22.0 +0300 @@ -577,6 +577,17 @@ if (ok st_dev_list_i-disp_part) /* Also read stats for its partitions */ read_sysfs_dlist_part_stat(curr, st_dev_list_i-dev_name); + else { + char *parent=strdup(st_dev_list_i-dev_name); + char *p=strpbrk(parent,DIGITS); + if(p) { + *p=0; + sprintf(filename,%s/%s/%s/%s, + SYSFS_BLOCK,parent,st_dev_list_i-dev_name,S_STAT); + read_sysfs_file_stat(curr,filename,st_dev_list_i-dev_name,DT_PARTITION); + } + free(parent); + } } /* Free structures corresponding to unregistered devices */
Bug#374524: pidentd: ERROR : UNKNOWN-ERROR
On Mon, Jun 19, 2006 at 03:50:55PM -0400, Mark Whitis wrote: postgresql was failing due to identd failures telnet localhost 113 113,59699 113 , 59699 : ERROR : UNKNOWN-ERROR Jun 19 15:27:13 localhost identd[4389]: started Jun 19 15:27:19 localhost identd[4389]: netlink_lookup: write failed: Connection refused http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=305926 reports of this bugs demise are greatly exagerated. fgrep TCPDIAG /usr/src/linux-source-2.6.16/.config shows nothing I think it is called CONFIG_INET_TCP_DIAG in 2.6.16 and should be there already. Loading module tcp_diag should do the trick. -- Toni Timonen toni dot timonen at iki dot fi NP Solutions Ltd Helsinki University of Technology Department of Engineering Physics and Mathematics signature.asc Description: Digital signature
Bug#251885: ITP: cgal -- C++ library for computational geometry
On Tue, Feb 07, 2006 at 08:04:40PM +0100, Joachim Reichel wrote: How does upstream release version 3.1 of CGAL, actually fixes these issues? The license of some files was changed such that all files in the Kernel and Support Library are licensed under LGPL, the files in the Basic Library are licensed under QPL. In particlar, all *.C files that end up in libCGAL.so are licensed under LGPL. ok. I got a different impression at first about the libCGAL.so. After doing QPL-cleanup: $ rm `grep -lr LICENSE.QPL .` Build do fails. There seems to be some debug headers that the core uses. Fortunately it is rather easy to fix (either fixing the license of the debug headers or the use of the debug header with an attached patch). -- Toni Timonen toni dot timonen at iki dot fi NP Solutions Ltd Helsinki University of Technology Department of Engineering Physics and Mathematicsdiff -rub CGAL-3.1-orig/include/CGAL/Nef_2/Polynomial.h CGAL-3.1/include/CGAL/Nef_2/Polynomial.h --- CGAL-3.1-orig/include/CGAL/Nef_2/Polynomial.h 2006-02-07 00:21:27.0 +0200 +++ CGAL-3.1/include/CGAL/Nef_2/Polynomial.h2006-02-07 23:29:44.0 +0200 @@ -35,7 +35,6 @@ #include cstddef #undef _DEBUG #define _DEBUG 3 -#include CGAL/Nef_2/debug.h #include vector diff -rub CGAL-3.1-orig/src/Polynomial.C CGAL-3.1/src/Polynomial.C --- CGAL-3.1-orig/src/Polynomial.C 2006-02-07 00:21:26.0 +0200 +++ CGAL-3.1/src/Polynomial.C 2006-02-07 07:27:24.0 +0200 @@ -24,6 +24,9 @@ #include CGAL/Nef_2/Polynomial.h +#define CGAL_NEF_TRACEN(foo) +#define CGAL_NEF_TRACEV(foo) + namespace CGAL{ signature.asc Description: Digital signature
Bug#251885: ITP: cgal -- C++ library for computational geometry
Hi, How does upstream release version 3.1 of CGAL, actually fixes these issues? I don't see any changes in the license since lgpl/qpl dual licensing. -- Toni Timonen NP-Ratkaisut Oy Teknillinen Korkeakoulu/Teknillinen Fysiikka 040-5111863,GPG 0x7984A4FD, [EMAIL PROTECTED] ICQ #45732842 irc://irc.npr.fi/#perunamaa signature.asc Description: Digital signature
Bug#38857: A patch for handling readonly keyrings in gnupg
Package: gnupg Version: 1.4.1-1.1 Followup-For: Bug #38857 tags 38857 + patch thanks Hi, I wrote a patch for gnupg to handle readonly keyrings correctly. What it does now, is that when you try to modify a key in a keyring that is not writable it makes a copy of that keyring to a writable one and modifies it. Patch attached. -- Toni Timonen toni dot timonen at iki dot fi NP Solutions Ltd Helsinki University of Technology Department of Engineering Physics and Mathematics diff -Nrub orig/gnupg-1.4.1/debian/patches/00list new/gnupg-1.4.1/debian/patches/00list --- orig/gnupg-1.4.1/debian/patches/00list 2005-07-18 19:17:32.0 +0300 +++ new/gnupg-1.4.1/debian/patches/00list 2005-07-18 19:53:17.0 +0300 @@ -1,2 +1,3 @@ 15_free_caps 16_min_privileges +20_readonly_keyrings diff -Nrub orig/gnupg-1.4.1/debian/patches/20_readonly_keyrings.dpatch new/gnupg-1.4.1/debian/patches/20_readonly_keyrings.dpatch --- orig/gnupg-1.4.1/debian/patches/20_readonly_keyrings.dpatch 1970-01-01 02:00:00.0 +0200 +++ new/gnupg-1.4.1/debian/patches/20_readonly_keyrings.dpatch 2005-07-18 19:53:13.0 +0300 @@ -0,0 +1,38 @@ +#! /bin/sh /usr/share/dpatch/dpatch-run +## 20_readonly_keyrings.dpatch by [EMAIL PROTECTED] +## +## All lines beginning with `## DP:' are a description of the patch. +## DP: Patch to handle readonly keyrings (see #38857). + [EMAIL PROTECTED]@ + +diff -rub orig/gnupg-1.4.1/g10/keydb.c gnupg-1.4.1/g10/keydb.c +--- orig/gnupg-1.4.1/g10/keydb.c 2004-10-14 10:11:56.0 +0300 gnupg-1.4.1/g10/keydb.c2005-07-18 19:34:53.0 +0300 +@@ -526,7 +526,17 @@ + rc = G10ERR_GENERAL; /* oops */ + break; + case KEYDB_RESOURCE_TYPE_KEYRING: ++ if(keyring_is_writable (hd-active[hd-found].token)) { + rc = keyring_update_keyblock (hd-active[hd-found].u.kr, kb); ++ } else { ++rc = keydb_locate_writable(hd,NULL); ++if (rc) { ++ log_error (_(no writable keyring found: %s\n), g10_errstr (rc)); ++ break; ++} ++rc = keyring_insert_keyblock (hd-active[hd-current].u.kr, kb); ++ ++ } + break; + } + +@@ -680,6 +690,8 @@ + case KEYDB_RESOURCE_TYPE_NONE: /* ignore */ + break; + case KEYDB_RESOURCE_TYPE_KEYRING: ++if(!keyring_is_writable(all_resources[i].token)) ++ break; + rc = keyring_rebuild_cache (all_resources[i].token,noisy); + if (rc) + log_error (_(failed to rebuild keyring cache: %s\n), signature.asc Description: Digital signature
Bug#312452: mencal: Support for default configuration file
Package: mencal Version: 2.3-4 Severity: wishlist Tags: patch upstream Hi, I wrote a patch for mencal to support having ~/.mencalrc read by default if such file exists. Patch attached. -- System Information: Debian Release: 3.1 APT prefers unstable APT policy: (500, 'unstable'), (500, 'testing'), (50, 'experimental') Architecture: i386 (i686) Shell: /bin/sh linked to /bin/bash Kernel: Linux 2.6.9 Locale: [EMAIL PROTECTED], [EMAIL PROTECTED] (charmap=ISO-8859-15) Versions of packages mencal depends on: ii perl 5.8.4-8Larry Wall's Practical Extraction -- no debconf information -- Toni Timonen toni dot timonen at iki dot fi NP Solutions Ltd Helsinki University of Technology Department of Engineering Physics and Mathematics diff -rub orig/mencal-2.3/mencal mencal-2.3/mencal --- orig/mencal-2.3/mencal 2003-07-17 08:53:15.0 +0300 +++ mencal-2.3/mencal 2005-06-08 11:23:39.0 +0300 @@ -229,6 +229,13 @@ ); my @confs=(); + #There exists a config file + $_ = $ENV{HOME}/.mencalrc; + if (-f $_ -r $_) { +push(@confs,Conf-new()); +$confs[$#confs]-parsefile($_); + } + for(my $i=0; $i=$#ARGV; $i++) { foreach($ARGV[$i]) { /^(-h|--help)$/ Main::usage; signature.asc Description: Digital signature
Bug#291180: pidentd: patch for the path of the pidfile
tags 291180 + patch sid thanks Hi, the patch for correct pidfile location was in the previous version, but was lost with the new upstream release. It can be fetched back with: $ zcat pidentd_3.0.16-7.diff.gz | filterdiff -i '*/pidentd.h' That patch is also attached. -- Toni Timonen toni dot timonen at iki dot fi NP Solutions Ltd Helsinki University of Technology Department of Engineering Physics and Mathematics --- pidentd-3.0.16.orig/src/pidentd.h +++ pidentd-3.0.16/src/pidentd.h @@ -42,7 +42,7 @@ #endif #ifndef PATH_PIDFILE -#define PATH_PIDFILE /etc/identd.pid +#define PATH_PIDFILE /var/run/identd/identd.pid #endif #ifndef PATH_CFGFILE signature.asc Description: Digital signature
Bug#291636: pidentd: patch for the load issue
tags 291636 + sid patch security thanks Hi, I managed to fix the issue. The original one I had on other(non-SMP) machine that some java processes doesn't authenticate correctly, and the problems arised when hammering a port(on a SMP machine; forgot to mention explicitly, sorry). The fix for the actual problem was already in 3.0.16-7 (assumably written by Herbert Xu, basing on the patch made by Paul Szabo. See #208290 for details), but didn't apply cleanly to the new upstream version. I rewrote/merged few parts that had problems applying cleanly. A patch attached. PS. As it seems (eg. according to #291180 ) there are few parts missing from the previous debian diff. You might consider applying them too. So, for convenience, also a cleanly applying patch containing those missing parts not directly related to #291636 or #291180 is also attached. -- Toni Timonen toni dot timonen at iki dot fi NP Solutions Ltd Helsinki University of Technology Department of Engineering Physics and Mathematics diff -rub us/pidentd-3.0.18/debian/changelog try3/pidentd-3.0.18/debian/changelog --- us/pidentd-3.0.18/debian/changelog 2005-01-21 19:41:09.0 +0200 +++ try3/pidentd-3.0.18/debian/changelog2005-01-22 16:56:25.0 +0200 @@ -1,3 +1,10 @@ +pidentd (3.0.18-1.1) unstable; urgency=low + + * Non-maintainer upload. + * Reapplied the TCPDIAG support patch. + + -- Toni Timonen [EMAIL PROTECTED] Sat, 22 Jan 2005 16:56:08 +0200 + pidentd (3.0.18-1) unstable; urgency=low * New upstream release. diff -rub us/pidentd-3.0.18/src/kernel.c try3/pidentd-3.0.18/src/kernel.c --- us/pidentd-3.0.18/src/kernel.c 2004-06-13 10:47:52.0 +0300 +++ try3/pidentd-3.0.18/src/kernel.c2005-01-22 16:55:36.0 +0200 @@ -107,7 +107,7 @@ if (debug) fprintf(stderr, ka_lookup(), attempt = %d, status = %d\n, attempt, kp-status); - if (kp-status 0) + if (kp-status = 0) break; if (attempt 2 (attempt 1) == 1) diff -rub us/pidentd-3.0.18/src/k_linux.c try3/pidentd-3.0.18/src/k_linux.c --- us/pidentd-3.0.18/src/k_linux.c 2003-01-09 01:02:26.0 +0200 +++ try3/pidentd-3.0.18/src/k_linux.c 2005-01-22 16:54:45.0 +0200 @@ -17,15 +17,28 @@ #include stdio.h #include syslog.h #include errno.h +#include unistd.h #include sys/types.h #include sys/param.h #include sys/socket.h #include netinet/in.h +#include netinet/tcp.h #include arpa/inet.h +#include linux/types.h +#include linux/netlink.h +#include linux/tcp_diag.h + #include pidentd.h +struct kainfo +{ +int nlfd; +__u32 seq; +FILE *proc_net_tcp; +}; + /* ** Make sure we are running on a supported OS version */ @@ -39,23 +52,85 @@ int ka_open(void **misc) { -FILE *fp; +struct kainfo *kp; +kp = s_malloc(sizeof(*kp)); -while ((fp = fopen(/proc/net/tcp, r)) == NULL errno == EINTR) - ; +kp-seq = 0; +kp-nlfd = socket(AF_NETLINK, SOCK_DGRAM, NETLINK_TCPDIAG); +if (kp-nlfd = 0) +goto out; -if (fp == NULL) +syslog(LOG_INFO, netlink failed, fallback to /proc/net/tcp: %m); +kp-proc_net_tcp = fopen(/proc/net/tcp, r); + +if (kp-proc_net_tcp == NULL) { syslog(LOG_ERR, fopen(\/proc/net/tcp\, \r\): %m); return -1; } -*misc = (void *) fp; +out: +*misc = (void *) kp; return 0; } +static int +netlink_lookup(struct kainfo *kip, struct kernel *kp) +{ +int status; +struct { +struct nlmsghdr nlh; +union { +struct tcpdiagreq req; +struct tcpdiagmsg rsp; +} u; +} buf; +struct sockaddr_nl addr; + +memset(buf, 0, sizeof(buf)); +buf.nlh.nlmsg_len = NLMSG_ALIGN(NLMSG_LENGTH(sizeof(buf.u.req))); +buf.nlh.nlmsg_type = TCPDIAG_GETSOCK; +buf.nlh.nlmsg_flags = NLM_F_REQUEST; +buf.nlh.nlmsg_seq = ++kip-seq; +buf.u.req.tcpdiag_family = AF_INET; + +buf.u.req.id.tcpdiag_dport = kp-remote.sin_port; +buf.u.req.id.tcpdiag_sport = kp-local.sin_port; +buf.u.req.id.tcpdiag_dst[0] = kp-remote.sin_addr.s_addr; +buf.u.req.id.tcpdiag_src[0] = kp-local.sin_addr.s_addr; +buf.u.req.id.tcpdiag_cookie[0] = TCPDIAG_NOCOOKIE; +buf.u.req.id.tcpdiag_cookie[1] = TCPDIAG_NOCOOKIE; + +status = write(kip-nlfd, buf, buf.nlh.nlmsg_len); +if (status 0) { +syslog(LOG_ERR, netlink_lookup: write failed: %m); +return 3; +} + +do { +socklen_t alen = sizeof(addr); +status = recvfrom(kip-nlfd, buf, sizeof(buf), 0, + (void *)addr, alen); +if (status 0) { +if (errno == ENOBUFS) +return -1; +syslog(LOG_ERR, netlink_lookup: recvfrom failed: %m); +return 3; +} +} while (addr.nl_pid || buf.nlh.nlmsg_seq != kip-seq); + +if (buf.nlh.nlmsg_type != TCPDIAG_GETSOCK) +return 0
Bug#291636: pidentd: New version is DoS-able and doesn't work well on load
Package: pidentd Version: 3.0.18-1 Severity: important Hi, the new version 3.0.18-1 has some problems when ran on machines having enough network traffic. First symptom that caused me to start looking was that sometimes java processes don't get authenticated correctly (very annoying when used with tomcat4 and ident authentication of postgresql). One way to reproduce the problem: -Run an irc server. -Put a torture script running: $ while [[ 1 ]]; do nc -z -v localhost 6667; done -Wait a while (15-30 seconds). -Try eg. indent-authenticate to postgresql server. Expected result (and behaviour of 3.0.16-7 in the same test): -Some system load -Running 7 identd processes -Some (one or two) zombies can be seen when repeatingly run 'ps -fA | grep ident'. -ident authentication works for other processes. Actual result: -Heavy system load. Some identd processes are running and eating all available cpu even after 5 minutes of actual port-hammering. -More than 100 identd -processes. -And the most important: authentication stops working. -- System Information: Debian Release: 3.1 APT prefers testing APT policy: (500, 'testing') Architecture: i386 (i686) Kernel: Linux 2.4.27-SMP Locale: [EMAIL PROTECTED], [EMAIL PROTECTED] (charmap=ISO-8859-15) Versions of packages pidentd depends on: ii adduser 3.59 Add and remove users and groups ii libc6 2.3.2.ds1-20 GNU C Library: Shared libraries an ii libssl0.9.7 0.9.7e-2 SSL shared libraries ii netbase 4.19 Basic TCP/IP networking system ii passwd 1:4.0.3-30.7 Change and administer password and -- no debconf information -- Toni Timonen toni dot timonen at iki dot fi NP Solutions Ltd Helsinki University of Technology Department of Engineering Physics and Mathematics signature.asc Description: Digital signature