Currently the list is as follows:
{{{ pkgdb=# select distinct p.name from package as p, packagelisting as pl, grouppackagelisting as gpl, grouppackagelistingacl as gpla where gpla.grouppackagelistingid = gpl.id and p.id = pl.packageid and pl.id = gpl.packagelistingid and gpl.groupid= 107427 and acl = 'commit' and gpla.statuscode != 3; name
arj autoconf autoconf213 automake automake14 automake15 automake16 automake17 bacula beecrypt bind-libbind binutils bitlbee calamaris certmaster cobbler cryptsetup-luks device-mapper device-mapper-multipath digitemp dmraid dovecot dsniff duplicity eggdrop event-compat-sysv fet firefox func gcc gdal geos gkrellm-top glibc gnutls grass hdparm hwdata iftop initscripts irda-utils iverilog kdebase kdebase-runtime kdebase-workspace kdelibs kdepimlibs kernel koan ksh libcap libgcrypt1 libgeotiff libnids libopm librsync libtool lvm2 mapnik mimedefang mksh ogdi openlayers openssh openssl pam pciutils perl-ccom perl-CGI-SpeedyCGI perl-Net-LibIDN perl-Razor-Agent perl-Socket6 pexpect php-idn php-magickwand popt prctl prelink proj python-boto python-GnuPGInterface python-suds rubygem-hoe setuptool smartmontools squirrelmail tcpick thunderbird ucview unicap vim wget xulrunner (93 rows) }}}
Using the list above, here's the breakdown by number of packages and owners:
57 robert 17 rezso 15 karsten 8 mdehaan 5 than 5 mhlavink 5 lvm-team 4 tmraz 4 gecko-maint 3 jakub 2 notting 2 mcpierce 2 jortel 2 ixs 1 wtogami 1 sharkcz 1 nickc 1 mschwendt 1 kernel-maint 1 johnp
Note that some of these are teams 'kernel-maint', etc. Is our next step emailing these folks and asking for some justification from them? Or should we discuss what is valid justification and what is not before doing that?
binutils firefox func gcc glibc kernel xulrunner
These seem like acceptable use to me. That would eliminate having to talk to kernel-maint and jakub at least. If we think thunderbird and the one other app are fine, then we can eliminate talking to gecko-maint as well.
notting can answer for himself here, so that's one less person to chase :)
dovecot
It was an omission, opened now :-)
event-compat-sysv is a dead package on all branches (in fact, we could just delete the module.)
initscripts changes should go upstream (it's in Fedora hosted) - they shouldn't need to be added in package CVS. (And cvsadmin & arch teams have access if something does need emergency fixing.)
removing from meeting agenda for now
Adding back to agenda in light of provenpackager reseed, and request on f-devel.
I'm especially interested in these packages:
17 rezso
as discussed multiple times on f-d-l:
http://www.redhat.com/archives/fedora-devel-list/2009-March/msg00214.html
http://www.redhat.com/archives/fedora-devel-list/2009-March/msg01029.html
Current list: {{{ arj autoconf autoconf213 automake automake14 automake15 automake16 automake17 b bacula beecrypt bind-libbind binutils bitlbee calamaris certmaster cobbler Common documentation files for oVirt cryptsetup-luks device-mapper device-mapper-multipath digitemp dmraid dsniff duplicity eggdrop erlang-pgsql event-compat-sysv fet firefox fonts-sig func gcc gdal geos gkrellm-top glibc gnutls grass hdparm hwdata iftop initscripts irda-utils iverilog jabbperpy kdebase kdebase-runtime kdebase-workspace kdelibs kdepimlibs kernel koan ksh libcap libgcrypt1 libgeotiff libnids libopm librsync libtool lvm2 mapnik mimedefang mksh moon-buggy ogdi openlayers openssh openssl pam pciutils perl-ccom perl-CGI-SpeedyCGI Perl extension for portable daemons perl-Net-LibIDN perl-Razor-Agent perl-Socket6 pexpect php-idn php-magickwand popt prctl prelink proj python-boto python-GnuPGInterface python-suds rubygem-hoe salut saxon8 setuptool sip-redirect smartmontools squirrelmail tcpick thunderbird ucview unicap vim wget xulrunner (102 rows) }}}
Some of those have been opened in devel. Here's the list without the ones that have been opened in devel and with the typo-packages removed:
select distinct p.name from package as p, packagelisting as pl, grouppackagelisting as gpl, grouppackagelistingacl as gpla where gpla.grouppackagelistingid = gpl.id and p.id = pl.packageid and pl.id = gpl.packagelistingid and gpl.groupid= 107427 and acl = 'commit' and gpla.statuscode != 3 and pl.collectionid = 8;
{{{ arj bacula beecrypt bind-libbind binutils bitlbee calamaris certmaster cobbler cryptsetup-luks device-mapper device-mapper-multipath digitemp dmraid dsniff duplicity eggdrop erlang-pgsql fet firefox fonts-sig func gcc gdal geos gkrellm-top glibc gnutls grass hwdata iftop initscripts iverilog jabbperpy kdebase kdebase-runtime kdebase-workspace kdelibs kdepimlibs kernel koan ksh libcap libgcrypt1 libgeotiff libnids libopm librsync lvm2 mapnik mimedefang mksh moon-buggy ogdi openlayers openssh openssl pam pciutils perl-ccom perl-CGI-SpeedyCGI perl-Net-LibIDN perl-Razor-Agent pexpect php-idn php-magickwand popt prelink proj python-boto python-GnuPGInterface python-suds rubygem-hoe salut saxon8 setuptool sip-redirect smartmontools squirrelmail tcpick thunderbird ucview unicap vim xulrunner (85 rows) }}}
OK, I'm going to send mail this weekend to <packagename>-owner of each of these stating they have two weeks to file tickets justifying the existence of ACL's.
Replying to [comment:10 jstanley]:
This is too long a timeframe to wait in the case of the gdal/grass stack of packages which is causing broken deps in Beta now. These need to be rebuilt and tagged ASAP. Please consider opening up these packages now, broken deps should override packager niceties IMHO.
Updated list with usernames:
Two step process,
select pl.owner, p.name from package as p, packagelisting as pl, grouppackagelisting as gpl, grouppackagelistingacl as gpla where gpla.grouppackagelistingid = gpl.id and p.id = pl.packageid and pl.id = gpl.packagelistingid and gpl.groupid= 107427 and acl = 'commit' and gpla.statuscode != 3 and pl.collectionid = 8 order by pl.owner;
Then get the userid/username mapping from fas.
{{{ orphan | fonts-sig kernel-maint | kernel lvm-team | lvm2 lvm-team | device-mapper lvm-team | cryptsetup-luks lvm-team | device-mapper-multipath lvm-team | dmraid gecko-maint | thunderbird gecko-maint | firefox gecko-maint | xulrunner karsten | libcap karsten | vim karsten | hwdata mschwendt | libgcrypt1 notting | initscripts robert | popt tmraz | openssl tmraz | pam tmraz | gnutls ixs | bacula than | kdebase than | kdelibs than | kdepimlibs than | kdebase-runtime than | kdebase-workspace mdehaan | func mdehaan | koan mdehaan | certmaster mdehaan | cobbler devrim | saxon8 jakub | prelink jakub | glibc jakub | gcc stahnma | jabbperpy rezso | geos rezso | grass rezso | proj rezso | libgeotiff rezso | mapnik rezso | openlayers rezso | gdal rezso | ogdi rezso | iverilog rezso | fet morgan | salut jortel | python-suds nickc | binutils mcpierce | rubygem-hoe mhlavink | setuptool mhlavink | pciutils mhlavink | squirrelmail mhlavink | ksh mhlavink | smartmontools jfch2222 | openssh (54 rows)
}}}
The exception for initscripts was declined. Others were considered in individual tickets.
What is the timetable for actually doing the ACL opening? This clearly hasn't yet been enacted as gdal (which didn't get an exception ticket files) is still not yet re-opened:
https://admin.fedoraproject.org/pkgdb/packages/name/gdal
Re-opening this ticket until the ACLs are actually changed.
This has been done except for the following packages:
pkgdb=# select pl.owner, p.name from package as p, packagelisting as pl, grouppackagelisting as gpl, grouppackagelistingacl as gpla where gpla.grouppackagelistingid = gpl.id and p.id = pl.packageid and pl.id = gpl.packagelistingid and gpl.groupid= 107427 and acl = 'commit' and gpla.statuscode != 3 and pl.collectionid = 8 and pl.statuscode in (3, 14) order by pl.owner;
{{{ owner | name -------+------------------------- 9905 | device-mapper-multipath 9905 | device-mapper 9905 | lvm2 9905 | dmraid 9905 | cryptsetup-luks 9910 | thunderbird 9910 | xulrunner 9910 | firefox (8 rows) }}}
As this is complete, except for the LVM stuff being tracked in ticket 124, closing this.
Log in to comment on this ticket.