Re: Bug#994750: RFS: mazeofgalious/0.63-1 [ITA] -- The Maze of Galious

2021-09-20 Thread Richard Hector

On 21/09/21 1:24 am, Parodper wrote:

  * URL : http://www.braingames.getput.com/mog/


No such site?

Cheers,
Richard



Re: Question about writing systemd unit for old package

2021-05-21 Thread Richard Hector

On 20/05/21 1:59 pm, Alec Leamas wrote:

Hi,

On 20/05/2021 03:35, Paul Wise wrote:

On Wed, May 19, 2021 at 8:51 AM Richard Hector wrote:


Does that not depend on whether it does anything before dropping
privileges? For example, a webserver can bind to low ports before
dropping privilege. I imagine if the systemd service unit specified
running as (eg) www-data, that wouldn't work.


I don't know the details, but I think systemd can open the ports and
transparently pass them to the unprivileged process when it is spawned
without any data loss, in a similar way to the inetd stuff used to
work.



http://0pointer.de/blog/projects/socket-activation.html


I confess I haven't read all that, and don't know the details of socket 
activation. But I think the service in question needs to be aware of it, 
doesn't it? It doesn't apply to wrapping a systemd service unit around 
an existing server. The nginx unit, for example, doesn't set a user, but 
a user is set in the nginx config file so it can drop privs.


I'm happy to be corrected :-)

Richard



Re: Question about writing systemd unit for old package

2021-05-19 Thread Richard Hector

On 18/05/21 11:58 am, Paul Wise wrote:

On Mon, May 17, 2021 at 12:51 PM Khoa Tran Minh wrote:




A related question: The binary itself can drop privilege and run as
non-root, then should I use that native feature or use systemd User= when
writing a default config/unit ?


I would suggest to use systemd features for this.


Does that not depend on whether it does anything before dropping 
privileges? For example, a webserver can bind to low ports before 
dropping privilege. I imagine if the systemd service unit specified 
running as (eg) www-data, that wouldn't work.


Cheers,
Richard



Re: Bug#979661: RFS: fonts-agave/37-1 -- monospaces programming font

2021-01-10 Thread Richard Hector

On 10/01/21 9:03 am, Gürkan Myczko wrote:

Package: sponsorship-requests
Severity: normal

Dear mentors,

I am looking for a sponsor for my package "fonts-agave":



   fonts-agave - monospaces programming font


That should probably be 'monospaced'.

Richard



Re: Bug#922197: RFS: worklog/2.0-1 -- Keep Track of Time worked on Projects

2019-02-12 Thread Richard Hector
On 13/02/19 6:57 PM, Adam Bilbrough wrote:
> I am looking for a sponsor for my package "worklog"
> 
> Package name: worklog
> Version: 2.0-1
> Upstream Author: Adam Bilbrough 
> URL: https://github.com/atsb/worklog

That site says the author is Truxton Fulton?

Richard



Re: debconf with locally derived options

2010-11-20 Thread Richard Hector
On Sat, 2010-11-20 at 17:30 +, Jonathan Wiltshire wrote:
> On Sat, Nov 20, 2010 at 08:36:35PM +1300, Richard Hector wrote:
> > One example would be configuring a daemon, and asking which of the
> > configured IP addresses of the machine it should listen on. Obviously
> > the packager has no idea what IP addresses are going to be on the target
> > machine, so can't list them in templates. I would therefore want to
> > generate the list on the fly, and present them as options to the user.
> 
> I recall that some packages do do this, and the first example that comes to
> mind is postfix. Take a look at the config, templates and postinst files in
> that package.

Thanks - there's a lot to go through there, but it's definitely pointing
me in the right direction :-)

Richard



-- 
To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: 
http://lists.debian.org/1290311031.17306.2.ca...@topaz.wgtn.cat-it.co.nz



debconf with locally derived options

2010-11-19 Thread Richard Hector
Hi all,

Is it possible for debconf, in the config file, to ask questions to
which the answer is one of several that make sense on the local system,
but couldn't be put in the templates file?

One example would be configuring a daemon, and asking which of the
configured IP addresses of the machine it should listen on. Obviously
the packager has no idea what IP addresses are going to be on the target
machine, so can't list them in templates. I would therefore want to
generate the list on the fly, and present them as options to the user.

Thanks,

Richard



-- 
To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: 
http://lists.debian.org/1290238595.27475.8.ca...@topaz.wgtn.cat-it.co.nz



Re: deploying package with NFS

2010-10-18 Thread Richard Hector
On Tue, 2010-10-19 at 15:50 +1100, Ben Finney wrote:
> Richard Hector  writes:
> 
> > We've run into an issue here, when we deploy a package (created
> > in-house) on a system that uses NFS for some filesystems. Due to
> > root-squashing, the postinst can't create or chmod/chown the files it
> > needs to.
> 
> You'll likely get better information if you say which paths are
> problematic, and which filesystems are NFS mounted.

True - we mount /var/lib/sitedata, and there is a subdirectory of that
for each website.

> 
> In the meantime, I would guess that either or both of the following are
> true:
> 
> * You're using Debian packaging to install files to paths that really
>   shouldn't be touched by the Debian packaging system.
> 
> * You're root-squashing filesystems that need root access.
> 
> > I was wondering though whether this issue is normally considered when
> > making a package? Or is root just assumed to be able to write
> > anywhere?
> 
> If the files are installed on a host from an OS package, it would be
> best for that host to have the filesystems mounted locally. That must be
> true for at least one host on your network, no?

The host that has it mounted locally does not have the package
installed. We have several webservers in an LVS cluster that need to
access shared data, on a dedicated fileserver.

> > The packages in question are web sites; we find that's the easiest way
> > to deploy them.
> 
> You might simply be using Debian's packaging system for a purpose
> contrary to its design.

True as well. We may well just have to use a local workaround.

I thought it was policy though that several paths are supposed to be
usable when remote mounted (like the whole of /usr, even?) - but perhaps
they can't be root-squashed. Or perhaps all updates are supposed to be
done on the NFS server.

Thanks,

Richard



-- 
To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: 
http://lists.debian.org/1287464274.8507.24.ca...@topaz.wgtn.cat-it.co.nz



deploying package with NFS

2010-10-18 Thread Richard Hector
Hi all,

We've run into an issue here, when we deploy a package (created
in-house) on a system that uses NFS for some filesystems. Due to
root-squashing, the postinst can't create or chmod/chown the files it
needs to.

We can probably avoid it by having the postinst su to the user that will
own the files, and creating them like that, and not using the usuall
installation modules.

I was wondering though whether this issue is normally considered when
making a package? Or is root just assumed to be able to write anywhere?

The packages in question are web sites; we find that's the easiest way
to deploy them. Regular debian packages don't go on NFS in our systems,
so they aren't an issue for us.

Thanks,

Richard



-- 
To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/1287459964.8507.5.ca...@topaz.wgtn.cat-it.co.nz