[arch-dev-public] [tur-users] UID/GID db sync

[Roman Kyrylych]

2007/11/8, Firmicus <Firmicus at gmx.net>:
> Roman Kyrylych a écrit :
> > Hi!
> >
> > Just FYI - I've synced both databases on our devwiki and public wiki.
> > Please everyone adding new UID/GID to database - do this as well.
> >
> > Users maintaining packages in AUR should check the db on wiki [1]
> > to make sure there are no conflicts with already existing UIDs/GIDs.
> > It is adviced to post a request first before adding UID/GID. Do not
> > add them without a need.
> >
> > [1] http://wiki.archlinux.org/index.php/UID_and_GID_List
> > P.S.: I'm watching this page so I'll do syncs when needed.
> >
> >
> I notice one omission:
>
> * davfs2 from extra creates user/group davfs2 with UID/GID = 88 (!
> conflicts with postgres !)

That's bad. davfs2 should change UID/GID.

>
> Some observations (probably irrelevant):
>
> * texlive-core creates group tex (no particular GID).
>
> * vmware-workstation creates group vmware (no particular GID) -- but
> this is in unsupported, so you probably won't put it in the db.

Yep, I don't know if it's worth to assign a defined GID to it.

About changing UID/GID for davfs2, texlive-core & vmware-*:
This could break already existing user setups (in some cases even
leading to files/dirs with old GID unexistent in new /etc/groups).
An example how to handle such situations can be seen in mlocate's
.install file, but in case of VMware it would be hard (or impossible?)
to search for files/dirs with vmware group.
For texlive-core and davfs2 it should be done, I think.

>
> * On my machine I also have group "testing" with GID 105, but I don't
> know where that comes from. Should I remove it?

I don't think such GID exists in Core&Extra, probably it's some crap.
You can use find to search for files with that GID to be sure.

-- 
Roman Kyrylych (Роман Кирилич)