Bug#798591: libreoffice: Long click closes Libre Office 5

[Chris Halls]

tags 798591 unreproducible moreinfo
thanks

Hi

On 10/09/15 22:30, Junior Polegato wrote:
> I updated some machines with the version 5 of Libre Office and now when user
> holds the left button of mouse to select some text or cells, the program
> closes, losing user data.

Oh dear, that doesn't sound good :(

> I noticed that this happened when the user holds and doesn't move the mouse,
> like a long click. Whatever, if he clicks, holds and moves the mouse, this
> doesn't happen.

I'm afraid I can't reproduce the problem here. I'm going to ask some
questions that will hopefully help us work out what is happening.

Is there a particular sequence of events that you need to trigger the
problem, or is it every time you hold the mouse button down?

You mentioned that you have upgraded some machines. Are you able to
reproduce the problem on different machines?

It's possible that there are some library incompatibilities caused by
the gcc 5 transition. Could you check if all packages are at their
newest version from unstable?

If you can easily reproduce the problem, could you try to get a stack
trace for us?

-
# Install debug symbols. These are quite big, so if you don't have
enough bandwidth or disk space, skip this step for now

apt-get install --no-install-recommends libreoffice-dbg

# We need gdb installed to generate the backtrace
apt-get install gdb

# Start libreoffice using '--backtrace'

loffice --backtrace

# Make it crash!

# Then send the file gdbtrace.log, which will be in your current working
directory where you started libreoffice from
-

Thanks!
Chris

Bug#798591: libreoffice: Long click closes Libre Office 5

[Chris Halls]

On 11/09/15 23:12, Nemo Inis wrote:
> Package: libreoffice
> Version: 1:5.0.1-1
> Followup-For: Bug #798591
> 
> I can reproduce this here. Backtrace is identical to Junior Polegato's (except
> different hex addresses.)
> All packages freshly updated from stretch.

Thanks to you both for confirming the bug, and for the backtrace from
Junior Polegato.

I've found that this bug has been reported upstream here:

 https://bugs.documentfoundation.org/show_bug.cgi?id=93778

I'm still unable to reproduce this on my system. Looking at the trace,
it looks like it is specific to the gtk integration. I don't use Gnome
so that's probably the problem.

I think you'll find that you can work around the bug by either:

* Uninstall package libreoffice-gtk
* Start LO like this: OOO_FORCE_DESKTOP=none loffice

Could you confirm this is the case please?

Thanks
Chris

Bug#798591: libreoffice: Long click closes Libre Office 5

[Chris Halls]

Hi John-Paul

On 13/09/15 00:27, John-Paul Durrieu wrote:
> uninstalling libreoffice-gtk, and/or starting LO with "OOO_FORCE_DESKTOP=none 
> loffice" does not
> work around the bug for me. Still crashes on long click ("long click" being 
> about 1 second
> without moving the mouse, by the way. Can click as long as I want while 
> moving the mouse).

Thanks for that. Could you try uninstalling libreoffice-gtk3 too please?
Also, which desktop do you use? Is it Gnome?

Chris

Bug#798591: Bug#795131: segfault in libvclplug_gtk3lo.so to use gcc5 libraries

[Chris Halls]

Hi JS

On 14/09/15 16:53, JS wrote:
> It was exactly reproduced by a long click, just as you said.
> 
> I hadn't noticed that before although the click only needs to last about
> 2 seconds so one barely notices it's long.

Ah, okay, that makes sense then. I'm CCing the long click bug report then.

> I was able to run under gdb but this is all I got (could retry if you
> suggest a group of breakpoints to set):

That doesn't look like an actual crash. /usr/bin/libreoffice and friends
are shell wrappers - maybe gdb was debugging the shell there?

You could try using the method Attaching to the soffice.bin process as
suggested in https://wiki.documentfoundation.org/Development/How_to_debug

$ gdb --pid=$(pidof soffice.bin)

Then type 'c' to continue at the gdb prompt, and reproduce the crash on
libreoffice.

Oh, and if you have the bandwidth, please will you install the
libreoffice-dbg package beforehand, so the backtrace is more meaningful.

Thanks for your help
Chris

Bug#349858: python-twisted-web: File conflict when upgrading from twisted 2,0

[Chris Halls]

Package: python-twisted-web
Version: 0.5.0-1
Severity: serious

Unpacking python-twisted-web (from .../python-twisted-web_0.5.0-1_all.deb) ...
dpkg: error processing 
/var/cache/apt/archives/python-twisted-web_0.5.0-1_all.deb (--unpack):
 trying to overwrite `/usr/bin/websetroot', which is also in package 
python-twisted

-- System Information:
Debian Release: testing/unstable
  APT prefers testing
  APT policy: (990, 'testing'), (500, 'unstable')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.15-1-k7
Locale: LANG=en_GB, [EMAIL PROTECTED] (charmap=ISO-8859-15)


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

Bug#350895: python2.3-twisted-runner: Conflict with python2.3-twisted-bin

[Chris Halls]

Package: python2.3-twisted-runner
Version: 0.1.0-1
Severity: serious
Justification: Policy 7.5.1

Unpacking python2.3-twisted-runner (from 
.../python2.3-twisted-runner_0.1.0-1_i386.deb) ...
dpkg: error processing 
/var/cache/apt/archives/python2.3-twisted-runner_0.1.0-1_i386.deb (--unpack):
 trying to overwrite 
`/usr/lib/python2.3/site-packages/twisted/runner/portmap.so', which is also in 
package python2.3-twisted-bin

$ dpkg -l python2.3-twisted-bin
||/ NameVersion Description
+++-===-===-==
ii  python2.3-twisted-bin   2.0.1-5 Event-based 
framework for internet applications


-- System Information:
Debian Release: testing/unstable
  APT prefers testing
  APT policy: (990, 'testing'), (500, 'unstable')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.15-1-k7
Locale: LANG=en_GB, [EMAIL PROTECTED] (charmap=ISO-8859-15)


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

Bug#346679: Fwd: Re: Bug#344679 acknowledged by developer (Bug#344679: fixed in ayttm 0.4.6+34-1)

[Chris Halls]

Hi Dato!

This bug is now closed, thanks for the report

Chris

--  Forwarded Message  --

> From: Chris Halls <[EMAIL PROTECTED]>
> Date: Fri, 10 Feb 2006 08:02:06 -0800
>
> Source: ayttm
> Source-Version: 0.4.6+34-1
>
> We believe that the bug you reported is fixed in the latest version of
> ayttm, which is due to be installed in the Debian FTP archive:
>
> ayttm_0.4.6+34-1.diff.gz
>   to pool/main/a/ayttm/ayttm_0.4.6+34-1.diff.gz
> ayttm_0.4.6+34-1.dsc
>   to pool/main/a/ayttm/ayttm_0.4.6+34-1.dsc
> ayttm_0.4.6+34-1_i386.deb
>   to pool/main/a/ayttm/ayttm_0.4.6+34-1_i386.deb
> ayttm_0.4.6+34.orig.tar.gz
>   to pool/main/a/ayttm/ayttm_0.4.6+34.orig.tar.gz
>
>
>
> A summary of the changes between this version and the previous one is
> attached.
>
> Thank you for reporting the bug, which will now be closed.  If you
> have further comments please address them to [EMAIL PROTECTED],
> and the maintainer will reopen the bug report if appropriate.
>
> Debian distribution maintenance software
> pp.
> Chris Halls <[EMAIL PROTECTED]> (supplier of updated ayttm package)
>
> (This message was generated automatically at their request; if you
> believe that there is a problem with it please contact the archive
> administrators by mailing [EMAIL PROTECTED])
>
>
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
>
> Format: 1.7
> Date: Fri, 10 Feb 2006 15:08:27 +
> Source: ayttm
> Binary: ayttm
> Architecture: source i386
> Version: 0.4.6+34-1
> Distribution: unstable
> Urgency: low
> Maintainer: Chris Halls <[EMAIL PROTECTED]>
> Changed-By: Chris Halls <[EMAIL PROTECTED]>
> Description:
>  ayttm  - Universal Instant Messaging Client
> Closes: 285484 308160 314149 314742 344679
> Changes:
>  ayttm (0.4.6+34-1) unstable; urgency=low
>  .
>* New upstream CVS snapshot:
>  - Fix gcc4 compilation
>  - IRC commands patch
>  - Allow override of connect server for gtalk support
>* xlibs-dev transition: Change to libxpm-dev build dependency
>  (closes: #344679)
>* Fix FTBS on amd64, thanks Andreas Jochens for the patch
>  (closes: #285484)
>* Enable dependency generation for libjasper (closes: #308160)
>* German translation update from Jens Seidel (closes: #314149)
>* Do not note that modules have changed on first install. Thanks
>  Greg Olszewski for patch (closes: #314742)
> Files:
>  dcf07a800f2717f80569d2fa0a4a1577 786 net optional ayttm_0.4.6+34-1.dsc
>  6ed02913eb2ea44159eee0f942b2402f 2389087 net optional
> ayttm_0.4.6+34.orig.tar.gz 0a405356f338b474707276e6891000c4 75505 net
> optional ayttm_0.4.6+34-1.diff.gz 40da98db6e9563c97ee85f67c01fa320 1864690
> net optional ayttm_0.4.6+34-1_i386.deb
>
> -BEGIN PGP SIGNATURE-
> Version: GnuPG v1.4.2 (GNU/Linux)
>
> iD8DBQFD7LXCexmdExmX588RAtCaAJ9MlQM3HJxYmjbTQEEmZM9D7AllgQCeOPsB
> xr8fBB52S5ybbzMPCEOwLlM=
> =yfFY
> -END PGP SIGNATURE-

---


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

Bug#353174: openoffice.org-help: Non-free file in source

[Chris Halls]

Package: openoffice.org-help
Severity: serious
Justification: Policy 2.2.2

$ tar -ztvf openoffice.org-help_2.0.1.orig.tar.gz \*parser.jar
-rw-r--r-- rene/rene136133 2000-10-10 13:28:16 
openoffice.org-help-2.0.1/external/common/parser.jar

This file does not come with source and violates the DFSG

-- System Information:
Debian Release: testing/unstable
  APT prefers testing
  APT policy: (990, 'testing'), (500, 'unstable')
Architecture: i386 (i686)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.15-1-k7
Locale: LANG=en_GB, [EMAIL PROTECTED] (charmap=ISO-8859-15)


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

Bug#375677: A question about Jose's patch

[Chris Halls]

Bill

Thanks a lot for looking at these problems. The codebase has changed a
lot since the version in unstable, and I have fixed a lot of these
problems already. You can see the current code at 

http://svn.debian.org/wsvn/apt-proxy/people/halls/rework/apt_proxy/?rev=0&sc=0

I had already modified all calls to deferred.errback() with a Failure()
instance as you suggested. The code works, but the queueing strategy
needs to be changed before I can upload to unstable.

Sorry for the confusion
Chris


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

Bug#375677: Fixing this bug

[Chris Halls]

On Thursday 03 August 2006 16:36, Daniel 'DaB' Baur wrote:
> hello all,
>
> I have created a package of apt-proxy which contains the fixing of this
> bug and all other bugs for that was a patch avaible. For me it work very
> fine.
>
> It can be found at http://www.daniel.baur4.info/public/debian/apt-proxy/
>
> Please note that I'm not a debian-developper and using is on OWN RISC of
> corse.
>
> If somebody can tell me, how I can make this version the official one
> please write me a private eMail.

Hi Daniel

Thanks a lot for your work. I'm just in the process of preparing an upload 
that fixes many of these bugs. Maybe you could help test it? You can get the 
source using:

svn checkout svn://svn.debian.org/svn/apt-proxy/trunk apt-proxy-svn

I haven't looked at your changes in detail, but from the changelog I guess 
that most of your changes are already in the new version. If there is 
anything else you want to change, please send me a diff from 'svn diff', and 
I'll integrate it.

I plan to make an upload tomorrow, when I have done some more testing on the 
packaging.

Chris


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

Bug#386344: apt-proxy: APT-Proxy has Problems with ~ in package-names

[Chris Halls]

package apt-proxy
tags 386344 + moreinfo
thanks

On Wednesday 06 September 2006 23:29, Daniel 'DaB.' Baur wrote:
> Package: apt-proxy
> Version: 1.9.35
> Severity: grave
> Tags: security
> Justification: renders package unusable
>
> Apt-Proxy can not handle Package, which have the new ~ in there names.
> A example is libkrb5-dev_1.4.4~beta1-1_i386.deb which gives a "404 file not
> found on backend" error. Please fix it, because it is unpossible to play in
> any packages or security-updates.

$ sudo apt-get install brazilian-conjugate
[...]
Get: 1 http://aptproxy testing/main brazilian-conjugate 3.0~beta4-1 [104kB]
Fetched 104kB in 0s (153kB/s)

It doesn't seem to be purely down to a '~' in the filename. It might be 
something in your configuration. For example, I am using an http backend but 
you could be using another type. The best thing is to turn on debug info in 
your apt-proxy-v2.conf:

debug = all:9 db:7

Then restart apt-proxy, and send a copy of /var/log/apt-proxy.log for the 
session where you try to get the file. When you do the test, can you also 
please check that the backend server you are using really has the file - 
sometimes mirrors can get out of sync.

Thanks
Chris


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

Bug#272206: apt-proxy/debootstrap keepalive bug workaround

[Chris Halls]

On Tuesday 22 Feb 2005 15:53, Isaac Wilcox wrote:
> I find this bites me when debootstrapping a new machine.  For the
> benefit of anyone else having problems with this setup (I'm using
> apt-proxy 1.9.25) and determined to use a local proxy, the following
> workaround solved it for me...

Hi, please can you test with 1.9.27 please?  With pipelining switched off 
(which it is by default since 1.9.19), there should not be any problems.  But 
maybe you have a different problem.

Thanks
Chris


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

Bug#297743: openoffice.org: oowriter crashes short after startup

[Chris Halls]

On Wednesday 02 Mar 2005 19:41, Andreas Pakulat wrote:
> On 02.Mär 2005 - 19:40:29, Rene Engelhard wrote:
> > Andreas Pakulat wrote:
> > > Xlib: unexpected async reply (sequence 0x7ada)!
> >
> > Since when does that happen? Did you change something except OOo? If
> > yes, what? May it be a xlib bug?
>
> Since this weekend I think. I wrote a document last week and it did
> not happen. This week oowriter crashes when I only start it and then
> look at my mails or something. nearly-daily updates of sid and yes
> there where other things changed. Luckily I collect the changelogs ;-)
>
> So here's the list:
>
> openoffice.org 1.1.3-5 on 19.Feb
> xfree86 4.3.0.dfsg.1-12 on 20.Feb
> openoffice.org 1.1.3-6 on 21.Feb
> openoffice.ogr 1.1.3-7 on 26.Feb
> xfree86 4.3.0.dfsg.1-12.0.1 on 26.Feb

Hmm, this thread started today on the upstream users list and talks about the 
same problem, but with 1.9:

http://www.openoffice.org/servlets/ReadMsg?listName=users&msgNo=86626

> Date: Wed, 02 Mar 2005 16:20:20 +0200
> From: Rich
> Content-Type: text/plain; charset=ISO-8859-13; format=flowed
> Subject: [users] Xlib: unexpected async reply
> 
> 
> unfortunately i haven't found a reliable way to reproduce this problem. 
> usually it happens when i'm using generic widgets and several times 
> switch between oo.org windows - although that's not a requirement, last 
> time it hang when i had a single drawing open. i'll try do some more 
> testing, so far i've seen it approx 5 times. hopefully ;) it'll surface 
> again.
> 
> Philipp Lohmann - Sun Germany wrote:
> > This would be a severe problem within OOo. Can you describe a method to 
> > reproduce this problem ? If so, please file an issue.
> > 
> > Kind regards, pl
> > 
> > Rich wrote:
> > 
> >> lately 1.9 beta candidate hangs now and then with an error message (i 
> >> run it from console) :
> >> Xlib: unexpected async reply
> >>
> >> could this be a problem in oo.org (i found a message in a debian list 
> >> that mentioned similar problem with staroffice, but it was pretty old 
> >> - so 5.2) ?
> 

Bug#272206: apt-proxy - supports keep-alive on http 1.0 requests

[Chris Halls]

severity 272206 normal
retitle 272206 apt-proxy - keep-alive behaviour broken if pipeling is enabled
tags 272206 + confirmed
thanks

On Saturday 18 Sep 2004 11:27, Bastian Blank wrote:
> Package: apt-proxy
> Version: 1.9.18
> Severity: important
>
> apt-proxy uses keep-alive on http 1.0 requests if it get the not
> specified "Connection: Keep-Alive" header but responds in a way which is
> not understood by the broken clients which sends this header.

pipelining is disabled by default as of 1.9.19, and with 1.9.27 I can only 
reproduce this when I have pipelining enabled.  So I am downgrading the 
report since I think we no longer need to regard it as release critical.

Steps to reproduce:

1. add or uncomment in apt-proxy.conf:

disable_pipelining=0

2. run:

 wget -S http://localhost:/debian/dists/sid/Release \ 
 http://localhost:/debian/dists/testing/Release

3. Examine Keep-Alive headers and wget behaviour 

Thanks a lot for your report
Chris


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

Bug#267880: Shouldn't be included on sarge

[Chris Halls]

On Tuesday 24 Aug 2004 22:01, Otavio Salvador wrote:
> Package: apt-proxy
> Version: 1.9.17
> Severity: serious
>
> This package is really usable for most users but have a lot of already
> know and serious bugs so I think is better leave it out of sarge.

Well, I think we are ready to remove this now.  If version 1.3 had been 
allowed into sarge I would not be saying this, but I think apt-proxy v2 is 
now stable enough to include in sarge now without being regarded as 
completely broken.

I am still aware of some fairly serious problems, which are top priority for 
the bugfixing effort now:

1. huge virtual memory usage.  Maybe in the Packages databases, since this 
only happens after a while.

  PID USER  PR  NI  VIRT  RES  SHR S %CPU %MEMTIME+  COMMAND
 5775 aptproxy  16   0  151m 130m 4356 S  0.0 13.4   0:30.56 twistd

2. apt-proxy-import seems to be broken.

What do you think Otavio?

Chris


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

Bug#284096: problem depends on Window manager?

[Chris Halls]

On Thursday 13 Jan 2005 20:02, Eric Valette wrote:
> Maintainers could try to reproduce the bugs on KDE if they use gnome as
> their favorite window manager...

I use KDE and can't reproduce this either, I'm afraid

Chris


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

Bug#284096: openoffice.org: writer crash in 'modify style' when selecting another tab

[Chris Halls]

On Friday 03 Dec 2004 16:10, Pontus Lidman wrote:
> Package: openoffice.org
> Version: 1.1.3-3
> Severity: grave
> Justification: renders package unusable
>
> Writer crashes in the style edit dialog. How to reproduce:
> 1) start oowriter
> 2) select any style in the stylist, right-click, select 'modify'
>from drop-down menu
> 3) Use left mouse button to click on any tab in the dialog. writer crashes.

I still can't reproduce this problem on my system, so I need your help.  
Please download this file:

http://people.debian.org/~halls/openoffice/test/libsvt645li.so.bz2

It is a version of the library that crashes with debugging symbols included.  
bunzip the file and copy into /usr/lib/openoffice/program, replacing the file 
that is there from the package.

Then, reproduce the crash.  When the crash dialog comes up, do not click it 
away but instead execute this from the command line (make sure you have gdb 
installed)

echo thread apply all bt > gdb.script
gdb -p $(pgrep -u $USER soffice.bin| head -1) < gdb.script | tee oo.log 

Please send oo.log to the bug report.

Thanks
Chris


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

Bug#289715: Same already fixed bugs are back rendering Impress almost unusable

[Chris Halls]

merge 289715 284096
tags 289715 + moreinfo unreproducible
thanks

On Monday 10 Jan 2005 18:07, Eric Valette wrote:
> Well done guys. Could you backup your erroneous patches once again...

I'm afraid that backing up will only delay things since these changes are 
backported from 2.0 so any bugs are likely to show up again then.

> 

Looking at the stack trace, this is the same crash as reported in 289715 so 
I'm merging the bug.  Looking at the code, OOo is dealing with font lists 
which are likely to be different on your machine to mine, which may be why I 
can't reproduce this here.  Please can you do the same as I asked the 
submitter of #284096 to do:
---
Please download this file:

http://people.debian.org/~halls/openoffice/test/libsvt645li.so.bz2

It is a version of the library that crashes with debugging symbols included.  
bunzip the file and copy into /usr/lib/openoffice/program, replacing the file 
that is there from the package.

Then, reproduce the crash.  When the crash dialog comes up, do not click it 
away but instead execute this from the command line (make sure you have gdb 
installed)

echo thread apply all bt > gdb.script
gdb -p $(pgrep -u $USER soffice.bin| head -1) < gdb.script | tee oo.log 

Please send oo.log to the bug report.

Thanks for your help
Chris

Bug#284096: debugging informatino with ooimpress (oo.log)

[Chris Halls]

On Monday 24 Jan 2005 14:16, Alexandre wrote:
> Attached to this mail is the oo.log file I got from the instructions in
> your mail from last friday.

Thanks a lot, that helped a lot. I have updated that library.  Could you 
please try the new one?  It won't fix the problem but should give me more 
information about the problem.  Sorry it is much larger than the previous 
library because it includes full debugging information for gdb.  Please send 
me an oo.log, and also the console output from the crash (you will see some 
extra messages beginning FontList::Get just before the crash)

Thanks
Chris


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

Bug#284096: debugging informatino with ooimpress (oo.log)

[Chris Halls]

tags 284096 - moreinfo unreproducible
tags 284096 + confirmed
thanks

For the record:

The key to reproudcing this are the following conditions:

- Not using KDE or Gnome plugin (or using OOO_FORCE_DESKTOP=none)
- No CUPS server running (maybe no printer queues at all on the system)

Finally I can reproduce this crash.

Thanks to Alexandre for giving me access to his machine to find this

To work around this bug, install the KDE or Gnome interface and configure OOo 
to use it.

Chris


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

Bug#284096: Test library for OOo impress fixes

[Chris Halls]

Hi all,

Thanks to Alexandre's help I was able to reproduce these problems on my 
machine and, after much digging around in the code, was able to find a 
solution that fixes the bugs on my system.

I have uploaded a library here for you to test:

http://people.debian.org/~halls/openoffice/test/libpsp645li.so.bz2

Please bunzip into /usr/lib/openoffice/program, replacing the existing library 
there.

For the record, here is a short explaination of the problem:  When reproducing 
the crasher in impress->format->character->change tab, valgrind reports that 
memory is being read that has already been deleted:

Address 0x1CC6F654 is 20 bytes inside a block of size 92 free'd
[...]
by 0x8086B16: operator delete(void*) (in soffice.bin)
by 0x21811850: SdDrawDocShell::SetPrinter(SfxPrinter*) (in libsd645li.so)
by 0x216FD206: SdWindow::DataChanged(DataChangedEvent const&) (in 
libsd645li.so)

The SetPrinter event is triggered by a SALEVENT_PRINTERCHANGED event which is 
generated whenever CUPSManager::checkPrintersChanged is called, which does 
not appear to be correct.  Since the 1.1 code does not seem to have been 
tested with such dynamic printer updates, I have disabled them, and now 
printer queues are only enumerated on startup.

Thanks
Chris


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

Bug#272206: apt-proxy - supports keep-alive on http 1.0 requests

[Chris Halls]

severity 272206 normal
retitle 272206 apt-proxy - keep-alive behaviour broken if pipeling is enabled
tags 272206 + confirmed
thanks

[I sent this message on 7 March but there was a problem with the mail itself, 
resending.]

On Saturday 18 Sep 2004 11:27, Bastian Blank wrote:
> Package: apt-proxy
> Version: 1.9.18
> Severity: important
>
> apt-proxy uses keep-alive on http 1.0 requests if it get the not
> specified "Connection: Keep-Alive" header but responds in a way which is
> not understood by the broken clients which sends this header.

pipelining is disabled by default as of 1.9.19, and with 1.9.27 I can only 
reproduce this when I have pipelining enabled.  So I am downgrading the 
report since I think we no longer need to regard it as release critical.

Steps to reproduce:

1. add or uncomment in apt-proxy.conf:

disable_pipelining=0

2. run:

 wget -S http://localhost:/debian/dists/sid/Release \ 
 http://localhost:/debian/dists/testing/Release

3. Examine Keep-Alive headers and wget behaviour 

Thanks a lot for your report
Chris


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

Bug#314236: apt-proxy: breaks with twisted > v2

[Chris Halls]

On Wednesday 15 Jun 2005 11:25, Mark De Souza wrote:
> Current version does not enforce python2.3-twisted < v2 yet when
> python2.3-twisted >= v2 is installed apt-proxy breaks

In what way?

> "/etc/init.d/apt-proxy start" breaks silently

Is there any output in the logfile?

> running /usr/sbin/apt-proxy identifies the problem correctly (ie the
> program states it will not run on versions > 2)

It says it has not been tested, but for me it works ok.  I uploaded 1.9.31 
that adds 2.0.1 to the list of known working versions.

Chris


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

Bug#462410: More info on ABI breakage

[Chris Halls]

I have established that the main ABI breakages were due to the Bristuff
patches.

1.4.2  - non-Bristuffed, same ABI as upstream
1.4.10 - Bristuff addeded, ABI change
1.4.15 - incompatible change to chan struct, ABI change

The asterisk-chan-capi version that this bug was reported against was
built against 1.4.13, which is the cause of the crashes.

I've had a discussion with Faidon on IRC, so just to document his
solution here: the 1.4.19 packages will provide a virtual package name
(currently asterisk-1.4), which should be used by dependent packages in
the future. If Bristuff breaks the ABI again, it will be possible to
change the name of the virtual provides so that dependent modules will
no longer be installable until compiled against the new ABI.

Faidon asked not to NMU as he plans an upload soon.

Chris




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

Bug#402481: apt-proxy: '500'-error still exists/has returned in version 1.9.35

[Chris Halls]

On Sunday 10 December 2006 19:47 pm, Folkert van Heusden wrote:
> When doing something like apt-get update && apt-get upgrade I get
> frequent 500-errors:
> Get: 288 http://192.168.64.99 testing/contrib msttcorefonts 1.5 [29.1kB]
> Errhttp://192.168.64.99 testing/main myspell-en-gb 1:2.0.4~rc1-3
>   500 Could not list directory

Hi

Could you attach your /var/log/apt-proxy.log please?

Thanks
Chris


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

Bug#436011: apt-proxy: 'module' object has no attribute 'CPU'

[Chris Halls]

Hiya

On Saturday 04 Aug 2007, Mykola Nikishov wrote:
> With this it works fine.

Thanks a lot for testing this and the suggested patch. Unfortunately I found 
that the patch breaks the package database handling. (It is possible to see 
this by comparing the unit tests before and after the patch.)

I think the package database code is probably using the apt library in the 
wrong way - see bug #285360. The original author of the code is no longer 
available so it will take me some time to fix this properly. However it 
passes the unit tests if I set the architecture to 'i386', even if it is not 
going to fix #285360.

Chris


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

Bug#436011: setting package to apt-proxy, forcibly merging 436011 436319

[Chris Halls]

# Automatically generated email from bts, devscripts version 2.10.6
package apt-proxy
# This FTBFS caused by python-apt ABI change
forcemerge 436011 436319



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

Bug#436112: apt-proxy: Failed to load application: /usr/lib/python2.4/site-packages/apt_pkg.so: undefined symbol: pkgCPU

[Chris Halls]

On Sun, Aug 05, 2007 at 06:45:28PM +0200, Folkert van Heusden wrote:
> Nope, problem persists.

> Failed to load application: /usr/lib/python2.4/site-packages/apt_pkg.so: 
> undefined symbol: pkgCPU

That's a problem with python-apt loading the apt library and would be a
bug in python-apt so I'm surprised its not been spotted already. Maybe
you forced a particular package version or have upgraded apt without
upgrading python-apt? In any case, upgrading your python-apt version
should fix this.

Chris


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

Bug#425535: ca-certificates: install failure: missing mktemp dependency

[Chris Halls]

Package: ca-certificates
Version: 20070303
Severity: serious
Justification: Policy 7.2

# apt-get install ca-certificates
[...]
Setting up ca-certificates (20070303) ...
Updating certificates in
/etc/ssl/certs/usr/sbin/update-ca-certificates: line 58: mktemp:
command not found
dpkg: error processing ca-certificates (--configure):
 subprocess post-installation script returned error exit status 127
 Errors were encountered while processing:
  ca-certificates
  E: Sub-process /usr/bin/dpkg returned an error code (1)

This is in a fairly small chroot. There is no mktemp package installed,
which provides /bin/mktemp and fixes the problem:

# apt-get install mktemp ca-certificates
[...]
The following NEW packages will be installed
  mktemp
  0 upgraded, 1 newly installed, 0 to remove and 0 not upgraded.
  1 not fully installed or removed.
[...]
  Unpacking mktemp (from .../mktemp_1.5-3_amd64.deb) ...
  Setting up mktemp (1.5-3) ...
  Setting up ca-certificates (20070303) ...
  Updating certificates in /etc/ssl/certsdone.

Chris

-- System Information:
Debian Release: lenny/sid
  APT prefers unstable
  APT policy: (500, 'unstable')
Architecture: amd64 (x86_64)

Kernel: Linux 2.6.20-1-amd64 (SMP w/2 CPU cores)
Locale: LANG=C, LC_CTYPE=C (charmap=ANSI_X3.4-1968)
Shell: /bin/sh linked to /bin/bash

Versions of packages ca-certificates depends on:
ii  debconf [debconf-2.0] 1.5.13 Debian configuration management sy
ii  openssl   0.9.8e-5   Secure Socket Layer (SSL) binary a

ca-certificates recommends no packages.

-- debconf information excluded


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

Bug#507865: [sw-discussion] OOo 3.0, 2.x and hidden text

[Chris Halls]

Hi Andreas

On Monday 08 Dec 2008, Andreas Martens wrote:
> On 12/05/08 16:48, Chris Halls wrote:
> > We've discovered that a bug fix to OOo's document export to ODF has made
> > certain writer documents saved with version 3.0 to appear to loose text
> > when you load the document in 2.x, even if you change the ODF format
> > version to "1.0/1.1 (Openoffice 2.x)".
> > http://www.openoffice.org/issues/show_bug.cgi?id=96878
> >
> > The issue has been closed because the problem with Writer 2.x loading the
> > document would need to be fixed in OOo 2.x, which is no longer supported
>
> We reopened this issue. I will try to fix it for a release of OOo2.x.
> Then OOo2.4.3 will be able to handle the "hidden-text"-attribute of
> documents from older 2.x versions as well as documents from OOo3.0 in a
> way to meet user expectations.

That's great, thanks Andreas!

I'm CCing the Debian bug which also applies to the 2.x branch.

Chris



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