Yes, I had to set use_from=yes in my ~/.muttrc. This directly
contradicts the documentation, which says:
1) The value of the from setting is used unless use_from is disabled
2) use_from is enabled by default
It should be removed from /etc/Muttrc, because this change is not a good
idea for many
Package: mutt
Version: 1.5.13-1
Severity: normal
I wanted to set my From address to [EMAIL PROTECTED] rather than [EMAIL
PROTECTED]
According to the muttrc(5) man page, I could do this by setting the from
variable in my ~/.muttrc, but for some reason this wasn't working.
It turns out that
Package: apache2-common
Version: 2.0.54-5
When I issue /etc/init.d/apache2 stop, I get a message:
Stopping web server: Apache2 ... no pidfile found! not running?.
This appears to be caused by a binary file existing in my /etc/apache2
directory, which causes the grep command in
Hi James
Upgrading to version 1.18.1 caused the problem to go away. I think
1.18.1 is now in etch.
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Package: postfix
Version: 2.2.4-1.0.1
Severity: normal
When sending mail from a private host (with a host name that
doesn't resolve via public DNS in the From and probably MAIL FROM)
via a relay host to my Debian system, the message is rejected with
this log message:
NOQUEUE: reject: RCPT from
This report is incorrect.
The reject_unknown_sender_domain option was enabled, however it was
specified in the smtpd_recipient_restrictions. This usage is a bit
weird and made it harder to notice this.
Sorry for the bug.
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of
Yes, my original test case now appears to be fixed.
PCRE support seems to have been since added and then removed, so in
theory we're back where we started. I wonder what change actually fixed
it. :-o
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble?
Package: fish
Version: 1.16.2-1
Severity: important
When running fish from within my login shell (zsh), it hangs
indefinitely before even showing a prompt. The only way to use it is to
press Ctrl-C which interrupts it. I then get a prompt that is not fully
usable (e.g. online help from the help
fd 3 appears to be a socket to communicate with fishd. I don't think
this has always been there, but don't know.
I now also get some error messages about common key bindings such as
Ctrl-U and Ctrl-D, but I suspect this is because readline is not fully
initialized or the readline configuration
Package: postfix
Version: 2.2.4-1.0.1
Severity: normal
Trying to run sudo newaliases fails with a permission denied error,
despite the destination file, /etc/aliases.db, being owned by root and
being user writeable.
ls -ld /etc
drwxrwsr-x 130 root root 8192 2005-11-29 20:50 /etc
/etc
ls -l
It appears to have been doing this because /etc/aliases was owned by
user number 500.
I still don't see why this is a good idea. It causes problems if
/etc/aliases is kept under RCS since committing a change to a file
stored under RCS changes the file's owner to the user committing that
Package: dovecot-imapd
Version: 1.0.alpha3-1
Severity: normal
Since upgrading to 1.0.alpha* from 0.99.*, my IMAP mail client can't
store mail in my Sent folder.
The client logs don't show anything meaningful, but the mail log shows:
Lost transaction log file
Package: dovecot
Version: 0.99.14-1
Severity: important
After performing an upgrade on my Etch system this week (from dovecot 0.99*
to the new 1.0 split packages), I get the following error when trying to
reconfigure or start Dovecot:
Unknown setting: passdb
The relevant section of the
It turns out the configuration file syntax has changed between 0.99.x
and 1.0. I had to migrate to a new format last month and already it's
changed again.
The following section appears to be equivalent to the old one.
Note that:
1. the braces are mandatory, even for an empty block
2. the end
Matthias Klose wrote:
Michael Wardle writes:
Package: bash3
Version: 3.0-12
Severity: minor
In bash 2, ALT TAB (M-\C-i) performed dynamic-complete-history.
In bash 3, it now does tab-insert.
The info node Commands for Completion still lists the old binding.
This node doesn't exist. please
Matthias Klose wrote:
Michael Wardle writes:
Package: bash3
Version: 3.0-12
Severity: minor
In bash 2, ALT TAB (M-\C-i) performed dynamic-complete-history.
In bash 3, it now does tab-insert.
The info node Commands for Completion still lists the old binding.
This node doesn't exist. please
Hi
It seems that the binding for Alt-Tab changed from dynamic-complate-history
to tab-insert around Bash 3.0.
The bashref.info file as found in Debian and the current version 3.0 Bash
source tarball on http://ftp.gnu.org/gnu/bash/ both still state the default
binding for
Matthias Klose wrote:
Michael Wardle writes:
I'll report this upstream.
It does exist in the generated .info file, which is what I was
originally looking at.
no, it exists in the bash tarball as well. which info file do you
read. bashref, or rluserman?
The version generated by upstream
I don't think it's to do with texinfo, the original texinfo source also
has the extra spacing. (See the attached sample patch).
I'm also not sure why the source wraps at about the 72nd column. This
makes the documentation artificially difficult to read on a large
terminal. Surely the
I was just hit by this in Etch after an update, upgrade.
mysql-server 4.0.24-10 seems to be the current version and
mysql-server-4.1 does not yet replace mysql-server 4.0.
Is this something that was not caught when this issue was resolved, or
is there something still coming down the line for
On Sun, 2005-08-14 at 22:41 +0100, Scott James Remnant wrote:
On Sat, 2005-07-16 at 12:49 +1000, Michael Wardle wrote:
Would it not be more efficient if this information were stored
in /var/lib/dpkg/status?
I would find this feature very useful.
Out of interest, what would you find
I noticed I only had logs since July, and didn't make the effort to
discover whether it was due to a log rotation policy or the log file
only being recently introduced. Clearly it's the latter. Thanks for
pointing it out.
My main use case was to be able to tell which packages were upgraded or
Would it not be more efficient if this information were stored
in /var/lib/dpkg/status?
I would find this feature very useful.
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Package: bash3
Version: 3.0-12
Severity: minor
In bash 2, ALT TAB (M-\C-i) performed dynamic-complete-history.
In bash 3, it now does tab-insert.
The info node Commands for Completion still lists the old binding.
-- System Information:
Debian Release: 3.1
APT prefers testing
APT policy:
Package: emacs21-el
Version: 21.4a-1
Severity: normal
If you select Save for Future Settings in the customize mode
(e.g. M-x describe-variable; indent-tabs-mode; /cust/; RET)
and the Emacs custom-file is a symbolic link, the symbolic link is
overwritten with a new file.
Reportedly there is an
My locale is en_AU.UTF-8 as you'll see hidden in the initial report. I
have also tried en_AU.utf8, but the behavior is the same (as you would
expect).
My terminal emulator is PuTTY 0.57 on Windows XP from
http://www.chiark.greenend.org.uk/~sgtatham/putty/.
PuTTY is configured to use UTF-8
Package: subversion
Version: 1.1.4-1
Followup-For: Bug #223303
The two bugs originally suggested as blockers for this one have been
fixed.
After discovering this emacs lisp module and uncommenting the invocation
in /etc/emacs to make it autoload, I gave it a brief spin and haven't
noticed any
Package: apache-dev
Version: 1.3.33-4
Severity: normal
apxs from apache-dev cannot be used to compile Apache 2 modules without
one of the apache2-worker-dev modules.
There is no obvious reference to this fact.
Please consider adding the apache2-worker-dev packages to the Suggests
header of
Package: apache2-prefork-dev
Version: 2.0.54-2
Severity: wishlist
The apache2-worker-dev packages only state that they provide Apache
development headers in the package description, not that they provide
any binaries.
Compare this to the apache-dev package, which seems to provide a more
useful
From what I have read, the wide character version of ncurses, ncursesw
(Debian package: libncursesw5), is supposed to provide UTF-8 encoded
line drawing characters using the same API, yet relinking the CVS
version of aptitude against -lncursesw exhibits the same behavior.
Indeed, testvscreen
Package: aptitude
Version: 0.2.15.9-2
Severity: normal
When using aptitude under a UTF-8 character set, what should be box
borders are appearing as lqqq...k under PuTTY.
My guess is that PuTTY expects the Unicode line-drawing characters
outlined here
Package: cvs
Version: 1:1.12.9-13
Severity: minor
The cvs(1) manual page contains far too many redundant new lines.
Paragraphs are mostly separated by two blank lines, rather than the
customary one, and as many as six blank lines appear in places.
This differs from the usual manual format and
Package: bash
Version: 2.05b-26
Severity: minor
Each link (text preceeded by * indicating a link to another document) is
separated by a blank line, giving the appearance of double-spaced lines.
This is unnecessary, causes less information to appear on the screen at
once, is unlike many other
Package: cvs
Version: 1:1.12.9-13
Severity: normal
The cvs(1) manual page has no section referring to the status command
whereas most of the other commands such as release and update do.
See also bugs #273872 and #282735.
-- System Information:
Debian Release: 3.1
APT prefers testing
APT
I misunderstood the problem. There was a more serious misconfiguration
relating to GNOME and automount on the client side causing it to
repeatedly and frequently *mount* /pub/.hidden.
A warning for a failed mount attempt (rather than a failed file access)
seems reasonable.
--
To
Package: nfs-kernel-server
Version: 1:1.0.6-3.1
Severity: normal
It seems that a failed stat() call on an NFS mounted directory causes
the following error message to be sent to syslog with facility user and
priority warning (example file is /pub/.hidden):
rpc.mountd: can't stat exported dir
Package: nis
Version: 3.13-1
Severity: normal
With a ksh-like shell (such as pdksh) as /bin/sh, /etc/init.d/nis stop
fails with a syntax error on line 160.
This is because the nis init script attempts to define a function
stop(), which is a predefined function for handling the STOP signal.
The
and consideration of this bug. :-)
--
Michael Wardle [EMAIL PROTECTED]
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Package: aspell-bin
Version: 0.60.2+20050121-1
Severity: minor
The aspell(1) manual page states that -l lists misspelled words
from standard input (which is the way that ispell works), however
this is not the case. The l flag in fact specifies the language
to use. This can be confirmed by
Package: grep
Version: 2.5.1.ds1-4
Severity: important
egrep is supposed to support beginning of word (\) and end of word (\)
anchors. For some reason, when invoked as egrep, the end of word anchor
does not work.
Contrast:
$ grep 'bug\' /usr/share/dict/words
bedbug
bedbug's
bug
bug's
debug
I don't have this problem on Red Hat Linux with the same upstream version.
I notice that Red Hat's grep is linked against PCRE and Debian's isn't,
so the bug could be anywhere (except PCRE ;-)).
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL
I was lead to believe this syntax should be supported by:
- Info:Grep:Regular Expressions
- regexp(5) on SunOS
On the other hand, none of these mention the word anchoring syntax:
- regex(3) on Debian
- regex(7) on Debian
- grep(1) on Debian
- POSIX standard
42 matches
Mail list logo