Package: ifupdown
Version: 0.8.35+b1
Followup-For: Bug #525315
Hey,
I'm pretty sure this should be filed as a bug to wpasupplicant.
In my case (same phenomena on display as OP reported), increasing the
hysteresis timeout to 10 seconds was enough to stop the cycling reliably.
For the record that
Package: prosody
Version: 0.9.11-1
Severity: wishlist
Dear Maintainer,
bcrypt is one of the hash algorithms looked upon more favourably.
Prosody supports it, yet the package makes no mention of it. Could we
recommend/suggest it?
Also: kudos for your on-the-button packaging!
Yours,
Hagen
Dear Johannes,
> No, not really - none of the demo programs in DOLFIN uses mshr.
Got me there. The project I'm contributing to (micromagnetics
software) switched to it and the only two examples I've found of any
interest were using it. I /might/ have extrapolated a bit. ;)
> mshr is already p
Package: fenics
Version: 2016.2.0.1
Severity: important
Hello,
Fenics/Dolfin contains the core module 'mshr' which deals in meshes.
All current examples seem to be employing 'mshr', eg.
https://fenicsproject.org/qa/9414/fenics-mesh-generation-mark-inner-region
The Fenics project page mentions th
Hey,
Alright, running kwrite with::
LANGUAGE=en LANG=en kwrite
works (with spotty highlighting for auto-check, but yeah).
My locale says::
LANG=en_GB.utf8
LANGUAGE=en_GB:en
however. I installed everything (on a clean slate Debian testing
machine) by saying::
aptitude install kwrite s
Hi,
> Do you have sonnet-plugins installed?
Yes. I also tried to provide different backends (aspell, hunspell,
myspell) exclusively and all together. No dice.
Note again, that after starting kwrite with eg.
strace -s 512 kwrite |& grep 'sonnet'
I'm still seeing the line
stat("/usr//usr/
Package: kwrite
Followup-For: Bug #793740
Hi,
Discovering that the bug is happily alive, I tinkered a bit with with
strace. One of the lines that appears after activating the
spell-checker reads::
stat("/usr//usr/lib/x86_64-linux-gnu/qt5/plugins/kf5/sonnet", 0x7ffe5537a0b0)
= -1 ENOENT (No s
Hi,
> Both these fixes will be in the first upload after Jessie releases.
Grand, thanks!
-Hagen
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Package: units
Version: 2.11-1
Followup-For: Bug #767864
Hello,
Thanks for maintaining units! Love it.
I have two suggestions,
1) Simply a refinement of Mike's proposal: Depend on or suggest
python-unidecode instead of python:all, it's compatible with Python
2 and 3, so this does t
Hello,
> > https://github.com/hfuchs/usbmount/tree/hfuchs.377548
>
> This version has a bug, though: [...]
> The problem is that $mountpoint is now set within a subshell[.]
Fixed. Using command groups now instead of that blasted subshell.
Updated on Github.
> (Shell scripting is never fun.)
Package: usbmount
Followup-For: Bug #676554
Hi,
Please close the bug report. There's nothing to do here. The problem
seems specific to the original poster's configuration (and might, of
course, be caused by what Edward described!).
Don't keep idle bug reports around, please close.
Thanks!
-H
Package: libc-bin
Version: 2.19-11
Followup-For: Bug #759530
Hello!
Something similar happened to me (after upgrade to 2.19-11). If this
report wasn't on top of libc-bin and absolutely recent, I would not have
said a word: read on!
Foundation-laying: the scenario. All packages that call ldconf
Hi there!
> > [flock(1)]
>
> I feel silly for not having been aware of that simple command. :)
Well, its handling is a bit weird (take a look at the 'third form'
description in the man page), but it does the job amiably. Also: I
spent a ludicrous amount of time trying to make usbmount lockless.
Package: usbmount
Followup-For: Bug #377548
Hi!
Frédéric's reply to this bugreport warmed my heart and I was inspired to
do what I promised Rogério a couple of years ago: solve all of
usbmount's problems. ;)
While Frédéric's approach is outstandingly creative, I think we can
address #377548 and
Package: usbmount
Followup-For: Bug #676554
Hi,
As nobody yelled during the last month, I strongly suggest we close this
bug report. If in future anything similar occurs, there's nothing
stopping anybody from filing a new bug or requesting to reopen this one.
Regards,
Hagen
--
To UNSUBSCRI
Package: clusterssh
Followup-For: Bug #666059
Hello,
Please close this report - clusterssh does not exhibit the bug any more.
I wrote the following over two years ago, but for some reason it never
reached the server (probably missing pseudo-header):
> Upstream fixed the problem in question (emp
Package: usbmount
Followup-For: Bug #658028
This is indeed a very valid concern (not that I have been affected
personally). Quite a few scripts in /lib/udev/rules.d seem to be doing
things that might take a while or even hang. I'm surprised, honestly.
In any case, I imagine usbmount's udev comp
Package: usbmount
Followup-For: Bug #676554
Hello,
> All my hotplugged usb devices are owned by root and only
> writable by the superuser. This can't possibly be the way this is
> supposed to work.
It's not. Rather, I'd suggest that somewhere along the line something
got misconfigured.
Do repl
Package: logcheck-database
Version: 1.3.16
Severity: normal
Tags: patch
Hello,
Attached you'll find a trivial patch that enables
'ignore.d.workstation/kernel' to capture the recently reworded message
"No Caching mode page present" (nowadays its "present" -> "found").
Thanks,
Hagen
diff --git a
Hello,
For posteriority and vanity's sake: I did indeed write a script for the
transition period from long to short Bonjour names as advertised by
CUPS[0].
It turned out surprisingly effective and stable, so I decided to keep
the script around instead of just throwing it away. Perhaps somebody
f
Hi!
> Sorry for not mentioning this in my first reply.
More embarrassing hand-holding? :)
Alright, as you guys have this thing covered on all angles, I'll just
say this: Awesome work! (I need new vocabulary.)
I'm actually looking forward to my next bug report. Sadly enough, I
really do.
You
> [package tracking system and git repository]
I see, thanks!
> Wondering if we should clone it with lowered severity for a more
> elegant fix. Comments?
Executive Sum.: Upstream is cleared of any wrongdoing, common::sense
needs to be built for every Perl version anew.
Longer:
I've been delving
First of all: Nice response time! I'm always in awe of Debian maintainers'
work ethic.
In response to your last comment,
> I've committed the changes for the first point; reviews welcome!
I'm not sure how I can review those changes? I `dpkg -i`'d the unstable
version of libcommon-sense-perl, b
Package: libcommon-sense-perl
Version: 3.72-1
Severity: grave
Justification: renders package unusable
Dear maintainer,
I just noticed that, upon upgrading to Perl 5.18.1-3 (Debian testing),
common::sense stopped having some:
$ perl -Mcommon::sense -e 'say "hi";'
syntax error at -e line 1
Hello again,
That's me back-tracking. Otmar was absolutely right: The load
excursions started again (after three months, quite suddenly); reliably,
usually at least twice a day with a production print server on Debian
Wheezy. Again, Till Kamppeter's patch referenced earlier fixed it (as
it did f
Hello again,
Thanks for the effort, appreciate it!
I have now pretty much exhausted my options by playing with Avahi's
settings (and CUPSens, for all I know). For example, I set
browse-domain=domain.local
in Avahi's config and got at least my Ubuntu test client to go along
with that. Not
Hi Brian,
Thanks for the quick reply!
> The only way I can get the second type of record permanently is by
> putting it in /etc/hostname, which I believe is read by avahi-daemon.
Ah, well now we're going into interesting territory. My /etc/hostname
actually says 'cups4dhcp'[0]. I have set CUPS
Package: cups-daemon
Version: 1.6.2-10
Severity: important
Hello!
I am using CUPS in Debian testing to provide access to a pool of 30+
printers to all sorts of OSes (research institute).
Switching from CUPS 1.5.3 to 1.6.2, I noticed that it registers its
Bonjour records differently with Avahi:
Hello,
About two weeks ago I replaced all my locally modified packages with
those in testing (which are currently the same version as in stable:
0.6.31-2).
In that period approx. 4000 print jobs have gone over our four print
servers and none of them showed any signs of the aforementioned load
exc
Package: avahi-daemon
Followup-For: Bug #693604
Hello,
I can confirm that applying the patch mentioned above[0] directly on the
Debian avahi-sources stopped these occurrences. To be clear: Avahi
/never/ recovers from these load excursions and stays at ~100% until
killed.
In our environment of 3
Hi,
> I already uploaded it a while ago, so if it doesn't tonight, it surely will be
> tomorrow :)
I was actually looking forward to digging around in the code a bit
tonight, but you were simply too fast - I didn't realize that you not
only fixed the thing but also packaged it by now.
Impressive
Hello David,
> Unfortunately, I'm the only upstream developer, so I can't really test
> every possible scenario.
Well, you're doing one hell of a job. wicd is my prime architectural
role models for UI server-client architecture. One of those "keep in
mind for the big thing" projects.
> I'm rea
the bug report :).
I suggest you get this to upstream's attentation as quickly as possible
as quite a few of wicd's own templates are now invalid. Hence the
elevated priority (I *do* hope, I'm not wrong on this ... ).
Thanks!
Regards,
Hagen Fuchs
Trivial proof-of-concep
# 2011-11-04, Created by Hagen Fuchs
#
# Purpose: ClusterSSH does not deal well with compound commands as
# handed to the '-a' flag, eg.
#
# cssh [remote_server] -a "cd /; ls; sleep 10"
#
# The directory listed will be the local working directory, /not/ the
# remote root di
Package: usbmount
Severity: normal
Hello,
I tested Jan's patch with a flash drive that has 12 partitions on it.
The original position of the locking code just at the beginning of the
"add" handler will usually only be able to mount 3 to 4 partitions on my
machine before lockfile-create will give
Hello again,
I was pondering a lock-less solution and your patches seem to provide a
good upgrade path towards that. Unfortunately:
$ git clone git://git.katzien.de/usbmount.git
Cloning into usbmount...
git.katzien.de[0: 82.165.99.49]: errno=Connection timed out
fatal: unable to
tting logically related code in a
separate function is almost always a good idea. So, yeah, I think we
should use that.
Best Regards,
Hagen Fuchs
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
ctive/ devices,
you'll probably overmount. ;)
So I think we'll have to find another way to speed usbmount up a little.
Best Regards,
Hagen Fuchs
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Package: usbmount
Severity: normal
Hello,
I just now perused the changelog and saw the message for version
0.0.17.1:
revert patch from #403209 (seems to introduce race conditions)
Seems, I should familiarize myself with the code /and/ its history,
before going off and analysing code from my
to do?
Please elaborate on those points, if at all possible. If you're not
sure whether those problems persist or you don't care anymore, please
tell us regardless.
Many thanks!
Best Regards,
Hagen Fuchs
[0] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=321694
--
To UNSUBS
Package: ding
Version: 1.7-1
Severity: normal
Hi there,
I ran into considerable difficulties trying to make good on the promise
of the package openthesaurus-de-text that it would "work with ding".
The format of their file is
bla;foo;bar
so that I cannot simply use ding's dictionary mode (us
openthesaurus.txt file), I would like to request a short description of
the process in the README.Debian file.
Best Regards,
Hagen Fuchs
-- System Information:
Debian Release: 6.0.2
APT prefers stable
APT policy: (500, 'stable')
Architecture: i386 (i586)
Kernel: Linux 2.6.32-5-486
Lo
min's suggestion (I'd written this report before
spotting his - I'd probably have shut up otherwise :).
Best Regards,
Hagen Fuchs
--- lxc-debian 2011-05-16 10:19:09.0 +0200
+++ lxc-debian.fixed2011-05-16 10:19:03.0 +0200
@@ -151,7 +151,7 @@
arch=amd64
Package: apt
Version: 0.8.14.1
Severity: normal
I recently discovered the /etc/cron.daily/apt script and wanted to use
it to download upgradeable packages and then immediately afterwards
delete them. Point is: There's a caching daemon (apt-cacher-ng) in
between that will keep requested packages
#x27; will be
appended implicitly.
Thanks!
Best Regards,
Hagen Fuchs
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Hi,
I realise that both versions posted earlier (1: downgrade
system()-string; 2: use Encode, use open) break character representation
in e-mail notifications. So the royal way would probably be to properly
decode()/encode() everywhere.
Or am I going completely astray here?
Regards,
HF
--
interested - I
still haven't quite grasped what I actually did.
Best Regards,
Hagen Fuchs
-- System Information:
Debian Release: 5.0.7
Architecture: i386 (i686)
Kernel: Linux 2.6.26-2-686 (SMP w/2 CPU cores)
Locale: LANG=en_GB.utf8, LC_CTYPE=en_GB.utf8 (charmap=UTF-8) (ignored: LC_ALL
set to e
Package: cil
Version: 0.07.00-1
Severity: important
The upstream version of cil has the dependency on libfile-homedir-perl
in 'debian-lenny/control'. But, even tough the debian source-archive
(acquired with `apt-get source cil`) contains the dependency
specification, the actual package does not l
Package: gitosis
Version: 0.2+20090917-2
Severity: important
Following the setup procedures in /usr/share/doc/gitosis/README.Debian,
I issued (as root):
sudo -H -u gitosis gitosis-init < gitosis_rsa.pub
The resulting Python traceback essentially says:
ImportError: No module named peak
There seems to be a duplication as well:
cd /usr/share/games/fortunes
$ grep encroachment *
cookie:"The greatest dangers to liberty lurk in insidious encroachment by
mean of zeal,
cookie.u8:"The greatest dangers to liberty lurk in insidious encroachment by
mean of zeal,
politics:The gr
Package: nagios3-common
Version: 3.0.6-3
Severity: normal
Trying to purge nagios3 from my system, I encountered this message:
Removing nagios3-common ...
/var/lib/dpkg/info/nagios3-common.prerm: line 23: syntax error near
unexpected token `fi'
dpkg: error processing nagios3-common (--remo
You just made my day. I'm halfway into the new one already? Nevermind,
you made that, too.
Thanks,
Hagen
signature.asc
Description: Digital signature
Package: iwatch
Version: 0.2.1-2
Severity: minor
In /etc/iwatch/iwatch.xml the guard email entry is
[EMAIL PROTECTED]
This should probably be
[EMAIL PROTECTED]
For reference, the whole entry (I'm too tired to diff properly, it's
trivial anyway):
-- System Information:
Debian Release: 4
Package: texlive-latex-extra
Version: 2007.dfsg.4-1
Severity: normal
The 'flashcards' package, residing in
/usr/share/texmf-texlive/tex/latex/flashcards/, does not properly built
the example flash card document 'samplecards.tex'
http://www.tug.org/texlive/Contents/live/texmf-dist/doc/latex/flas
Package: rsnapshot
Version: 1.2.9-1
Severity: important
As the cron-file for rsnapshot is in cron.d/, systems which rely on
anacron for execution of cron-jobs may not be doing backups as expected!
Anacron does *not* execute jobs in cron.d/.
I suggest splitting the rsnapshot cron-files into cron.d
Package: rsnapshot
Version: 1.2.9-1
Severity: normal
rsnapshot produces warnings like this
WARNING: Could not lchown() symlink
when the Perl Lchwon.pm-module is not installed. It is, however, not
possible to install this package in a clean, Debian-ized way, as this
module (file) is nowhere to
After a substantial amount of hacking, I think I have actually succeded
in implementing this function. I would not consider my changes to be
hacks, although I had my quarrels with the design (which I find quite
pleasing now).
Anyway, my patch makes it possible to press 'g' (goto) on an search
res
> I'm not sure I did understand you correctly. What do you mean by
> "entering some album"? Do you mean something like "add the corresponding
> album to the playlist"?
That would be a consequence of the feature I'm interested in:
Enter the browse-screen (3) from one of the entries of the search-sc
Package: ncmpc
Version: 0.11.1+svn-r3965-2
Severity: wishlist
It would be very logical to implement a way to enter the album that
corresponds to one of the resulting entries in a search list; á la:
"I'd really like to hear that album where 'foo' sang 'bar'!"
Thanks,
HF
-- System Information
> can you please tell me, which aspell dictionary you are using? may be
> if it is too big, freetalk is not handling it..
Sorry, I tried different word-lists (wngerman was the one with 300.000+
entries) - english, british and even cleared the wordlist once:
Loading dictionary [/usr/share/dict/w
Hi,
> Try latest version from Lenny or even more latest version from Sid.
> I have never seen such issues before with freetalk.
Already tried that - sorry to say: same outcome.
I can wait infinitely long (so it seems) after invoking freetalk.
Then I Ctrl-C and all is well. Is there some way to
Package: freetalk
Version: 0.5-2
Severity: normal
After invoking freetalk 0.5, I need to press CTRL-C after waiting one moment:
$ freetalk
[ 1 sec ]
CTRL-C
Loading dictionary [/usr/share/dict/words]... [320577] words
and normal operation begins.
This even happens, when I remove my '.free
62 matches
Mail list logo
