On Mon, Dec 27, 2010 at 10:41:36PM -0600, Steve M. Robbins wrote:
> On Tue, Dec 28, 2010 at 09:38:30AM +1100, Matthew Palmer wrote:
> > On Mon, Dec 27, 2010 at 05:04:02PM +0100, Julien Cristau wrote:
> > > do you plan on NMUing pbuilder for this bug?
> >
> > No, I don't consider it appropriate to
On Tue, Dec 28, 2010 at 09:38:30AM +1100, Matthew Palmer wrote:
> On Mon, Dec 27, 2010 at 05:04:02PM +0100, Julien Cristau wrote:
> > do you plan on NMUing pbuilder for this bug?
>
> No, I don't consider it appropriate to NMU for an RC bug that I raised the
> severity on, withouth the acknowledgem
On Mon, Dec 27, 2010 at 05:04:02PM +0100, Julien Cristau wrote:
> do you plan on NMUing pbuilder for this bug?
No, I don't consider it appropriate to NMU for an RC bug that I raised the
severity on, withouth the acknowledgement of the maintainer that the
severity is justified. Someone else NMUing
On Wed, Dec 15, 2010 at 14:21:05 +1100, Matthew Palmer wrote:
> tag 542915 +patch
> usertags 542915 -in-progress +patch-in-git
> thanks
>
> Attached is my minimal patch solving the problem of data loss in
> bind-mounted directories. It provides a safety net that, in the event that
> *anything* i
On Thu, Dec 16, 2010 at 06:14:55PM +0100, Jakub Wilk wrote:
> * Matthew Palmer , 2010-12-15, 14:21:
>> +if mount |grep -q $(readlink -f $BUILDPLACE)/; then
>
> I think
>
> grep -q -F " $(readlink -f $BUILDPLACE)/"
>
> would be more robust.
Right you are. Patch 5 in a series:
>From 5
* Matthew Palmer , 2010-12-15, 14:21:
Jakub, I'd appreciate it if you would be able to apply the patch to
your local pbuilder installation (it should go onto pbuilder-modulers
in-situ, without a need for a rebuild) and see if you can find any ways
to break it.
I'm afraid I won't help here; I
tag 542915 +patch
usertags 542915 -in-progress +patch-in-git
thanks
Attached is my minimal patch solving the problem of data loss in
bind-mounted directories. It provides a safety net that, in the event that
*anything* is still mounted inside the chroot, no attempt to delete anything
will be made
On Sun, Dec 12, 2010 at 12:08:57PM +0100, Jakub Wilk wrote:
>> 3) cp /usr/share/pbuilder/examples/C10shell /tmp/hooks
>
> s,pbuilder,doc/pbuilder,
Ah, yes, sorry about that.
>> But if I ls /tmp/valuable, the file do-not-delete-please is still there.
>
> Are /tmp/valuable and /var/cache/pbuilder/b
submitter 542915 !
thanks
* Matthew Palmer , 2010-12-12, 14:07:
I'm trying to reproduce this bug, but I can't manage to get pbuilder to
behave in the manner you describe -- but I can't find any changes to
pbuilder in the meantime that would explain a change in behaviour.
Thanks for taking care
tag 542915 +moreinfo
thanks
Hi Jakub,
I'm trying to reproduce this bug, but I can't manage to get pbuilder to
behave in the manner you describe -- but I can't find any changes to
pbuilder in the meantime that would explain a change in behaviour.
Here is what I did, using pbuilder 0.199:
1) Crea
* Loïc Minier , 2009-09-21, 15:46:
First of all, you should not close a bug without a proper explanation.
mount and rm does exactly what they are meant to do (and what is
documented). This is not true for your package.
In fact, this bug is easily fixable:
- In most cases, `umount -a` called ins
On Fri, Sep 18, 2009, Jakub Wilk wrote:
> First of all, you should not close a bug without a proper explanation.
>
> mount and rm does exactly what they are meant to do (and what is
> documented). This is not true for your package.
>
> In fact, this bug is easily fixable:
> - In most cases, `umou
* Junichi Uekawa , 2009-09-19, 00:20:
Yeah, reassign this bug to 'mount' and 'rm'.
First of all, you should not close a bug without a proper explanation.
mount and rm does exactly what they are meant to do (and what is
documented). This is not true for your package.
In fact, this bug is eas
Package: pbuilder
Version: 0.189
Severity: grave
Justification: causes data loss
How to lose your data with pbuilder:
0. Install the C10shell hook.
1. Try to build a package which happens to FTBFS.
2. Realize that you need some files from your $HOME to investigate the
failure.
3. Bind-moun
14 matches
Mail list logo
