Bug#391698: fetchmail: try-restart section in init-script should call $0 restart, not $0 awaken

2006-10-08 Thread Thomas Schmidt
Package: fetchmail
Version: 6.3.4-7
Severity: normal

The try-restart section in the init-script should call $0 restart, not
$0 awaken, because otherwise the resolvconf-script 
/etc/resolvconf/update-libc.d/fetchmail, which is called when the
ipaddress of the dns-server changes will have no effect, fetchmail will
still try to query the old dns-server, which is not reachable anymore.

The reason for this seems to be that the fetchmail-daemon does not 
recognize new dns-servers when it is just awakened, it must be really 
restarted. (So the cause for this is more or less an upstream problem.)

The problem was introduced with the fix for #268346 and i only noticed
it on my laptop, it allways occurs when the laptop switches to a 
different network after coming out of suspend.

It might also be possible that #369270 is more or less the same problem,
because at least the symptoms are exactly the same.


Regards,
Thomas


-- System Information:
Debian Release: testing/unstable
  APT prefers testing
  APT policy: (500, 'testing'), (190, 'unstable'), (1, 'experimental')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.18-w3n
Locale: LANG=de_DE.UTF-8, LC_CTYPE=de_DE.UTF-8 (charmap=UTF-8)

Versions of packages fetchmail depends on:
ii  adduser  3.97Add and remove users and groups
ii  debianutils  2.17Miscellaneous utilities specific t
ii  gettext  0.14.6-1GNU Internationalization utilities
ii  libc62.3.6.ds1-4 GNU C Library: Shared libraries
ii  libssl0.9.8  0.9.8c-1SSL shared libraries
ii  lsb-base 3.1-15  Linux Standard Base 3.1 init scrip

Versions of packages fetchmail recommends:
ii  ca-certificates   20060816   Common CA Certificates PEM files

-- no debconf information


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Bug#391698: fetchmail: try-restart section in init-script should call $0 restart, not $0 awaken

2006-10-08 Thread Nico Golde
Hi,
* Thomas Schmidt [EMAIL PROTECTED] [2006-10-08 12:58]:
 Package: fetchmail
 Version: 6.3.4-7
 Severity: normal
 
 The try-restart section in the init-script should call $0 restart, not
 $0 awaken, because otherwise the resolvconf-script 
 /etc/resolvconf/update-libc.d/fetchmail, which is called when the
 ipaddress of the dns-server changes will have no effect, fetchmail will
 still try to query the old dns-server, which is not reachable anymore.
 
 The reason for this seems to be that the fetchmail-daemon does not 
 recognize new dns-servers when it is just awakened, it must be really 
 restarted. (So the cause for this is more or less an upstream problem.)
[...] 
I'm sorry but I will not fix this since it really is an 
upstream problem, look #389270 for example. I hope this will 
be fixed soon. If not I will fix this as a workaround for 
etch.
Kind regards
Nico
-- 
Nico Golde - http://www.ngolde.de
JAB: [EMAIL PROTECTED] - GPG: 0x73647CFF
Forget about that mouse with 3/4/5 buttons,
gimme a keyboard with 103/104/105 keys!


pgp2Q7nqU5gxB.pgp
Description: PGP signature


Bug#391698: fetchmail: try-restart section in init-script should call $0 restart, not $0 awaken

2006-10-08 Thread Thomas Schmidt
* Nico Golde schrieb am 08.10.06, um 13:18 Uhr:
  The reason for this seems to be that the fetchmail-daemon does not 
  recognize new dns-servers when it is just awakened, it must be really 
  restarted. (So the cause for this is more or less an upstream problem.)
 [...] 
 I'm sorry but I will not fix this since it really is an 
 upstream problem, look #389270 for example. I hope this will 
 be fixed soon. If not I will fix this as a workaround for 
 etch.

Ok, i hope to see a real fix for this issue very soon too, in the
meantime i will have to change the init-script manually. ;-)

Regards,
Thomas

-- 
Thomas Schmidt, Debian VDR Team
http://pkg-vdr-dvb.alioth.debian.org/


signature.asc
Description: Digital signature


Bug#391698: [pkg-fetchmail-maint] Bug#391698: fetchmail: try-restart section in init-script should call $0 restart, not $0 awaken

2006-10-08 Thread Nico Golde
Hi,
* Thomas Schmidt [EMAIL PROTECTED] [2006-10-08 15:20]:
 * Nico Golde schrieb am 08.10.06, um 13:18 Uhr:
   The reason for this seems to be that the fetchmail-daemon does not 
   recognize new dns-servers when it is just awakened, it must be really 
   restarted. (So the cause for this is more or less an upstream problem.)
  [...] 
  I'm sorry but I will not fix this since it really is an 
  upstream problem, look #389270 for example. I hope this will 
  be fixed soon. If not I will fix this as a workaround for 
  etch.
 
 Ok, i hope to see a real fix for this issue very soon too, in the
 meantime i will have to change the init-script manually. ;-)

This DNS stuff is really annoying, I already contacted 
upstream, maybe we'll find a solution, maybe not, hopefully 
we do :)
Kind regards
Nico
-- 
Nico Golde - http://www.ngolde.de
JAB: [EMAIL PROTECTED] - GPG: 0x73647CFF
Forget about that mouse with 3/4/5 buttons,
gimme a keyboard with 103/104/105 keys!


pgpD1OWtE3jBD.pgp
Description: PGP signature


Bug#391698: [pkg-fetchmail-maint] Bug#391698: fetchmail: try-restart section in init-script should call $0 restart, not $0 awaken

2006-10-08 Thread Matthias Andree
Nico Golde [EMAIL PROTECTED] writes:

 This DNS stuff is really annoying, I already contacted 
 upstream, maybe we'll find a solution, maybe not, hopefully 
 we do :)

Well, the problem is: there is no explicit caching of resolver, cached
results etc. in fetchmail -- and the upstream (that is me) cannot
reproduce the problem. So where do we start looking for a problem in
fetchmail's code? getaddrinfo()/freeaddrinfo() is what we're currently
doing, and once again for every run...

What I *can* say is that I have seen many a standard-violating
getaddrinfo() implementation... few are thread-safe, although IEEE Std
1003.1 mandates so.

-- 
Matthias Andree


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]