FC5 anaconda
by Gene C.
There have been a number of characteristics/functionality for anaconda in FC5
mentioned on this list. Is there some place where the plan is
described/discussed?
I currently subscribe to the anaconda-devel-list but only started in late
January so if there was discussion earlier, I missed it.
--
Gene
19 years, 1 month
I want more CDs in the distribution
by Daryll Strauss
Here's a counter idea for people to chew on. I'd like to see more CDs
in the distribution, I'd just like them all to be less full.
One of the problems we seem to be bumping up against is that although
a personal desktop, a workstation, and a server share a lot of
packages, there's also a lot of packages that are specific to one type
of use or another. Server's typically don't install OpenOffice, and
desktops don't install NetNews servers in most cases.
What if we break the distribution in to more CD's each with fewer
packages more tightly organized around usage. For a made up example:
CD1 & CD2 - Base OS (practically everyone uses)
CD3 - Core Server packages
CD4 - Core Dekstop packages
CD5 - Additional Server packages
CD6 - Additional Desktop packages
CD7 - Really bizarre packages (anything else we include for some reason)
ADVANTAGES:
If I'm installing a basic desktop or basic server I'm downloading
fewer packages.
The installer already seems to handle multiple discs and notifying
the user which ones are needed.
If you're downloading everything the total size really isn't
substantially bigger (just a smidge for ISO overhead)
If you install a minimal version and want to add packages later you
can do that via the Internet anyway. (We should encourage people to do
this if they have broadband)
DISADVANTAGES:
More files to download - I don't think this is an issue
More CDs required for the full distribution - Media costs are small
and you can put it all on one DVD if you really care. This probaby is
an issue for publishers.
More disk swapping to install everything - That's a real drawback,
but since you have to swap disks anyway I'm not sure it's a big deal.
More work categorizing end user usage and arguing over which
packages are in which category. - yes :)
Users need to figure out which disks to download.
>From my point of view that last one is the only one that a big
usability issue. The installer already tells me which disks are
required early on in the process but that's not when you want to find
out. It's just a failsafe. Documentation would help, but we know
people often don't read it. A web script that let a user pick what
packages they want to install and it tells them what discs to get
would be helpful and fairly easy. In reality we want the installer to
get stuff from repositories after the fact, but that's a more major
change.
What do you think?
19 years, 1 month
writing zero bytes in bash
by Russell Coker
This message is intentionally sent to the fedora-devel-list and the SE Linux
list not the fedora-selinux-list. It is not related to Fedora specific SE
Linux functionality, but it is related to bash (a core part of Fedora) and SE
Linux kernel code.
To unset the fscreate or exec context you have to write zero bytes
to /proc/self/attr/fscreate or /proc/self/attr/exec respectively.
If you want to do this in a shell script you would do something like:
echo -n "" > /proc/self/attr/fscreate
However that shell command results in bash (both the version in rawhide and
the version in RHEL4) performing the following system calls:
open("/proc/self/attr/exec", O_WRONLY|O_CREAT|O_TRUNC|O_LARGEFILE, 0666) = 3
fcntl64(1, F_GETFD) = 0
fcntl64(1, F_DUPFD, 10) = 10
fcntl64(1, F_GETFD) = 0
fcntl64(10, F_SETFD, FD_CLOEXEC) = 0
dup2(3, 1) = 1
close(3) = 0
dup2(10, 1) = 1
fcntl64(10, F_GETFD) = 0x1 (flags FD_CLOEXEC)
close(10) = 0
It opens the file with O_CREAT (so if you were to do `echo -n "" > /tmp/flag`
to create a flag file then it would work as expected), but never calls the
write(2) system call.
To unset the fscreate or exec context you have to call write(fd, X, 0) (the
value of X doesn't seem to matter as the kernel code doesn't dereference it).
It seems likely to me that there may be other bits of kernel code exporting an
interface under /proc where a write of 0 bytes may have a different meaning
to merely opening a file with write access and then closing it.
So it seems reasonable to me to consider this to be a bug in bash where it's
optimisation of shell code results in the action requested by the user not
being performed correctly.
It took a surprising amount of time for me to realise that bash wasn't doing
what I expected of it and ran strace to prove it. I expect that most
sys-admins would take even longer to work it out (if they ever did).
I expect that the bash developers may disagree with this assessment so I would
like some more input on the lists before I file a bug report.
--
http://www.coker.com.au/selinux/ My NSA Security Enhanced Linux packages
http://www.coker.com.au/bonnie++/ Bonnie++ hard drive benchmark
http://www.coker.com.au/postal/ Postal SMTP/POP benchmark
http://www.coker.com.au/~russell/ My home page
19 years, 1 month
ITE 8212 and pwc drivers absent in Rawhide
by Peter Backlund
Hello.
I'd like to see the ITE 8212 IDE/RAID pci card driver
( CONFIG_BLK_DEV_IT8212) and the pwc webcam driver back in the kernel.
Is there any reason why they were removed?
/Peter
19 years, 1 month
debug about NetworkManager & resolv.conf being erased
by Arnaud Abélard
Dan, you've been asking for debug infos about NetworkManager & the
resolv.conf problem, so here it comes:
I've been trying to use NetworkManager to manage my ethernet and
wireless devices on FC3. As it was previously said on the list,
NetworkManager sometimes erase the current resolv.conf file with en
empty one.
I did some tests:
NetworkManager will only do the dhcp request when detecting a new
network connection. Which means that if you run NetworkManager when the
network cable is already plugged on the NIC, NetworkManager will not do
the dhcp request and will erase the resolv.conf file with an empty one.
I ran NetworkManager after booting. the network cable is already plugged
on the NIC:
[root@gallilee ~]# NetworkManager --no-daemon
NetworkManager: starting...
NetworkManager: eth0: Driver support level is fully-supported
NetworkManager: nm_create_device_and_add_to_list(): adding device 'eth0'
(wired)
NetworkManager: eth1: Driver support level is fully-supported
NetworkManager: nm_create_device_and_add_to_list(): adding device 'eth1'
(wireless)
NetworkManager: running mainloop...
NetworkManager: nm_dbus_get_networks():
org.freedesktop.NetworkManagerInfo.NoNetworks raised There were are no
wireless networks stored.
NetworkManager: AUTO: Best wired device = eth0, best wireless device =
eth1 ()
NetworkManager: SWITCH: best device changed
NetworkManager: nm_state_modification_monitor(): beginning activation
for device 'eth0'
NetworkManager: nm_state_modification_monitor() activated device eth0
Result: resolv.conf has been erased.
Now, i stop NetworkManager and unplug the network cable then run
NetworkManager once more before replugging the cable:
[root@gallilee ~]# NetworkManager --no-daemon
NetworkManager: starting...
NetworkManager: eth0: Driver support level is fully-supported
NetworkManager: nm_create_device_and_add_to_list(): adding device 'eth0'
(wired)
NetworkManager: eth1: Driver support level is fully-supported
NetworkManager: nm_create_device_and_add_to_list(): adding device 'eth1'
(wireless)
NetworkManager: running mainloop...
NetworkManager: nm_dbus_get_networks():
org.freedesktop.NetworkManagerInfo.NoNetworks raised There were are no
wireless networks stored.
NetworkManager: AUTO: Best wired device = (null), best wireless device =
eth1 ()
NetworkManager: SWITCH: best device changed
NetworkManager: nm_state_modification_monitor(): beginning activation
for device 'eth1'
NetworkManager: nm_device_activation_worker (eth1) started...
NetworkManager: nm_device_activate_wireless(eth1): waiting for an access
point.
NetworkManager: nm_device_activate_wireless(eth1): waiting for an access
point.
NetworkManager: nm_device_activate_wireless(eth1): waiting for an access
point.
NetworkManager: nm_device_activate_wireless(eth1): waiting for an access
point.
NetworkManager: nm_device_activate_wireless(eth1): waiting for an access
point.
NetworkManager: nm_device_activate_wireless(eth1): waiting for an access
point.
NetworkManager: nm_device_activate_wireless(eth1): waiting for an access
point.
NetworkManager: nm_device_activate_wireless(eth1): waiting for an access
point.
NetworkManager: HAL signaled link state change for device eth0.
NetworkManager: AUTO: Best wired device = eth0, best wireless device =
eth1 ()
NetworkManager: SWITCH: best device changed
NetworkManager: nm_device_activation_cancel(eth1): cancelling...
NetworkManager: nm_device_activation_worker(eth1): activation canceled.
NetworkManager: Activation (eth1) IP configuration/DHCP returned = 0
NetworkManager: Activation (eth1) IP configuration/DHCP unsuccessful!
Ending activation...
NetworkManager: Activation (eth1) ending thread.
NetworkManager: nm_device_activation_cancel(eth1): cancelled.
NetworkManager: nm_state_modification_monitor(): beginning activation
for device 'eth0'
NetworkManager: nm_device_activation_worker (eth0) started...
NetworkManager: dhcp_interface_init: MAC address = 00:0f:1f:fe:9d:2f
NetworkManager: ClassID = "Linux 2.6.10-1.766_FC3 i686"
ClientID = "1.0.15.1F.FE.9D.2F.00.00"
NetworkManager: Broadcasting DHCP_DISCOVER
NetworkManager: DHCP: Starting request loop
NetworkManager: DHCP: Sending request packet...
NetworkManager: DHCP: Sent request packet.
NetworkManager: DHCP: Waiting for reply...
NetworkManager: DHCP waiting for data, overall end_time = {1109323109s,
138338us}
NetworkManager: DHCP waiting for data of minimum size 28, remaining
timeout = {5s, 96347us}
NetworkManager: DHCP: Got some data to check for reply packet.
NetworkManager: DHCP: actual data length was 339
NetworkManager: debug_dump_dhcp_options: 7 options received:
NetworkManager: i=1 (subnetMask) len=4 option = 255.255.255.0
NetworkManager: i=3 (routersOnSubnet) len=4 option = xxxxxxxx
NetworkManager: i=6 (dns) len=8 option = xxxxxxxx
NetworkManager: i=6 (dns) len=8 option = xxxxxx
NetworkManager: i=15 (domainName) len=31 option = "xxxxxxxxxxxxxx"
NetworkManager: i=51 (dhcpMessageType) len=4 option = 43200
NetworkManager: i=53 (dhcpParamRequest) len=1 option = 2
NetworkManager: i=54 (dhcpMsg) len=4 option = xxxxxx
NetworkManager: dhcp_msg->yiaddr = xxxxxx
NetworkManager: dhcp_msg->siaddr = xxxxxxx
NetworkManager: dhcp_msg->giaddr = 0.0.0.0
NetworkManager: dhcp_msg->sname = ""
NetworkManager: Server Hardware Address = 00.10.5A.DE.FF.B1
NetworkManager: broadcastAddr option is missing in DHCP server response.
Assuming xxxxxx.255
NetworkManager: dhcpIPaddrLeaseTime = 43200 in DHCP server response.
NetworkManager: dhcpT1value is missing in DHCP server response. Assuming
21600 sec
NetworkManager: dhcpT2value is missing in DHCP server response. Assuming
37800 sec
NetworkManager: DHCP_OFFER received from (xxxxxxxx)
NetworkManager: Broadcasting DHCP_REQUEST for xxxxxxxx
NetworkManager: DHCP: Starting request loop
NetworkManager: DHCP: Sending request packet...
NetworkManager: DHCP: Sent request packet.
NetworkManager: DHCP: Waiting for reply...
NetworkManager: DHCP waiting for data, overall end_time = {1109323110s,
-600706us}
NetworkManager: DHCP waiting for data of minimum size 28, remaining
timeout = {5s, 132041us}
NetworkManager: DHCP: Got some data to check for reply packet.
NetworkManager: DHCP: actual data length was 339
NetworkManager: debug_dump_dhcp_options: 10 options received:
NetworkManager: i=1 (subnetMask) len=4 option = xxxxxx.0
NetworkManager: i=3 (routersOnSubnet) len=4 option = xxxxxxxx
NetworkManager: i=6 (dns) len=8 option = xxxxxxxx
NetworkManager: i=6 (dns) len=8 option = xxxxxxx
NetworkManager: i=15 (domainName) len=31 option = "xxxxxxxx"
NetworkManager: i=28 (broadcastAddr) len=4 option = xxxxxx.255
NetworkManager: i=51 (dhcpMessageType) len=4 option = 43200
NetworkManager: i=53 (dhcpParamRequest) len=1 option = 5
NetworkManager: i=54 (dhcpMsg) len=4 option = xxxxxxxxx
NetworkManager: i=58 (dhcpClassIdentifier) len=4 option = 0
NetworkManager: i=59 (dhcpClientIdentifier) len=4 option = 0
NetworkManager: dhcp_msg->yiaddr = xxxxxxxx
NetworkManager: dhcp_msg->siaddr = xxxxxxxx
NetworkManager: dhcp_msg->giaddr = 0.0.0.0
NetworkManager: dhcp_msg->sname = ""
NetworkManager: Server Hardware Address = 00.10.5A.DE.FF.B1
NetworkManager: dhcpIPaddrLeaseTime = 43200 in DHCP server response.
NetworkManager: dhcpT1value is missing in DHCP server response. Assuming
21600 sec
NetworkManager: dhcpT2value is missing in DHCP server response. Assuming
37800 sec
NetworkManager: DHCP_ACK received from (xxxxxxxx)
NetworkManager: Your IP address = xxxxxxxx
NetworkManager: : Adding nameserver: xxxxxxx
NetworkManager: : Adding nameserver: xxxxxxx
NetworkManager: : Adding domain search: xxxxxxx
NetworkManager: Activation (eth0) IP configuration/DHCP returned = 1
NetworkManager: Activation (eth0) IP configuration/DHCP successful!
NetworkManager: nm_state_modification_monitor() activated device eth0
NetworkManager: nm_device_activation_worker(eth0): device activated
Result: the resolv.conf file is properly filled.
--
Arnaud Abélard
Administrateur Systèmes et Réseaux
Faculté de Sciences et Techniques
Université de Nantes
19 years, 1 month
kde-3.4 (3.3.92) in fc4
by Rex Dieter
I know it's getting close, but I think it's important to try to get
kde-3.4 (3.3.92beta2 at the moment) into fc4. If not, fedora users will
have to wait for fc5 before seeing this (please correct me if I'm wrong).
-- Rex
19 years, 1 month
rawhide report: 20050224 changes
by Build System
Removed package abiword
Removed package Numeric
Removed package octave
Removed package bzflag
Removed package gnumeric
Removed package koffice
Removed package ggv
Removed package gpdf
Removed package kinput2
Removed package xemacs
Removed package xfce4-panel
Removed package xfce4-iconbox
Removed package xfce-utils
Removed package xfce4-systray
Removed package xfce-mcs-plugins
Removed package xfce-mcs-manager
Removed package libxfce4mcs
Removed package libxfcegui4
Removed package libxfce4util
Removed package Maelstrom
Removed package exim
Removed package lapack
Removed package cfengine
Removed package kdetoys
Removed package FreeWnn
Removed package nabi
Removed package system-switch-im
Removed package gv
Removed package skkinput
Removed package miniChinput
Removed package xcin
Removed package dbskkd-cdb
Updated Packages:
ddskk-12.2.0-5
--------------
* Wed Feb 23 2005 Elliot Lee <sopwith(a)redhat.com> 12.2.0-5
- Remove xemacs
flim-1.14.7-2
-------------
* Wed Feb 23 2005 Elliot Lee <sopwith(a)redhat.com> 1.14.7-2
- Remove xemacs subpackage
gaim-1:1.1.3-4
--------------
* Tue Feb 22 2005 Warren Togami <wtogami(a)redhat.com> 1:1.1.3-4
- Test fixes for #149190 and #149304
gcc-3.4.3-20
------------
* Tue Feb 22 2005 Jakub Jelinek <jakub(a)redhat.com> 3.4.3-20
- update from gcc-3_4-branch
- PRs c++/14479, c++/19487, c++/19739, c++/19755, c++/19762, c++/19787,
c++/20028, libstdc++/19829, libstdc++/19946, libstdc++/19955,
middle-end/19697, preprocessor/19077, target/19019, target/19715
- rename PowerPC IBM long double helper routines _xlq* to __gcc_*,
but keep _xlq*(a)GCC_3.4 aliases around (#148841, PR target/19019)
- change __cxa_demangle to match cxx-abi change
http://www.codesourcery.com/archives/cxx-abi-dev/msg01877.html
(Jason Merrill, #133406)
- only obsolete < NVR gcc-gnat and libgnat if %{build_ada} is 0
(#139537)
- revert the alloca vs. VLA patch, it has too many false positives (#147758)
- make sure libjava GC memory is executable for libffi trampolines sake
(#149348, PR libgcj/19823)
* Thu Feb 10 2005 Jakub Jelinek <jakub(a)redhat.com> 3.4.3-19
- use crtendS.o instead of crtend.o on ppc -pie
- use execv instead of execl in libgcc_post_upgrade to avoid bringing
in malloc and friends into the statically linked binary (which increases
its size ~ 10 times)
* Thu Feb 10 2005 Jakub Jelinek <jakub(a)redhat.com> 3.4.3-18
- update from gcc-3_4-branch
- PRs c++/18370, c++/19366, c++/19499, c++/19733, libstdc++/19642,
middle-end/19775, target/15384, target/16201, target/17771,
target/19293, target/19329, target/19393, target/19803
- fix c++filt/__cxa_demangle segfault on invalidly mangled names
generated by G++ 3.4 (#145781, PR c++/16240)
- make sure libgcj.so is not PT_GNU_STACK RWE
- disallow dlopening libgnat-3*.so, as it must be PT_GNU_STACK RWE
due to its extensive use of trampolines
- fix PRs c++/18838 and c++/19367 (Mark Mitchell, backported by
Alexandro Oliva)
- fix ICE in fold_convert (Andrew Pinski, #146385, PR c++/19666)
hal-cups-utils-0.5.2-9
----------------------
* Tue Feb 22 2005 John (J5) Palmieri <johnp(a)redhat.com> - 0.5.2-9
- For some reason -8 was never built into rawhide. Bumping the
release to 9 and rebuilding for smooth upgrade from fc3
* Fri Oct 22 2004 John (J5) Palmieri <johnp(a)redhat.com> - 0.5.2-8
- use_usb_if_null patch makes sure that the make and model of the
printer is never null even if hal does not populate the
printer.vendor and info.product fields. (Bug #136666)
- (printer_update.hal, printer_remove.hal): Use $UDI instead of
$HAL_PROP_INFO_UDI which in some instances may not be populated
mew-4.1-2
---------
* Wed Feb 23 2005 Elliot Lee <sopwith(a)redhat.com> 4.1-2
- Remove xemacs
redhat-rpm-config-8.0.33-2
--------------------------
* Wed Feb 23 2005 Arjan van de Ven <arjanv(a)redhat.com> 8.0.33-2
- add -Wall on Ulrichs request
rpmdb-fedora-1:4-0.20050224
---------------------------
wl-2.10.1-5
-----------
* Wed Feb 23 2005 Elliot Lee <sopwith(a)redhat.com> 2.10.1-5
- Remove xemacs
19 years, 1 month
Re: fedora-devel-list Digest, Vol 12, Issue 86
by Martin Sevior
On Thu, 24 Feb 2005, Peter Robinson wrote:
> So with these removed from Core I assume they are going to be in Extras.
> Correct? In Extras will they continue to be maintained as before or will
> they need someone to step up and maintain them? The reason I ask is it
> would be a pity to see abiword and gnumeric disappear as I find them
> considerably faster than OOo and in the case of Abiword it supports a
> number of word files that don't work to well with OOo.I also use grip.
> While I'm happy for them to be in extras rather than core I would still
> like them to be around.
> The plan is to add these packages to Extras as demand requires. Volunteers
> are always wanted to help maintain packages. Are you interested? Please
> see http://fedoraproject.org/wiki/Extras_2fCvsAccess
How sad. We constantly push the envolope, fix bugs, provide upgrades and
provide rpm's for stable and development version of our program for
FC3, FC2, FC1 and redhat 9 on our website.
Maybe we should just stick with source and Debian.
I've been seriously thinking of moving to Ubuntu anyway.
Well I guess that's just about it for FC as a community distro.
Martin Sevior
Core AbiWord Developer.
Hint type "word Processor" into google and see what you get. It isn't either
MS Word or OOo
19 years, 1 month
Re: I want more CDs in the distribution
by Jamie Zawinski
[ oops, sent from wrong email addr last time ]
Daryll Strauss wrote:
>
> What if we break the distribution in to more CD's each with fewer
> packages more tightly organized around usage. For a made up example:
> CD1 & CD2 - Base OS (practically everyone uses)
> CD3 - Core Server packages
> CD4 - Core Dekstop packages
> CD5 - Additional Server packages
> CD6 - Additional Desktop packages
> CD7 - Really bizarre packages (anything else we include for some reason)
Eh, I think that means I'd always have to download at least 6 CDs,
because chances are gnome, httpd, and xemacs will be on different discs,
so it seems (selfishly) like no gain.
What I generally do is install the absolutely minimal fedora I can, and
then install the rest after the machine has booted. This is good
because it gets me past the part I might actually have to fuck around
with (partitioning, debugging boot loader problems) right away, and then
most importantly, *gets me a network* while things are installing, so
I'm not sitting there bored out of my mind while it grinds away:
Anaconda doesn't let me ssh out to check my mail.
Also it means I don't end up installing 80% of the packages twice (the
version on the CD, and then the upgrade from yum.)
I know that the installer currently does that two-stage thing where
firstboot (or whatever) asks you to install additional packages after
the machine is on the net with a desktop. I'd like to see stage 1 get a
lot smaller, and almost everything go into stage 2.
--
Jamie Zawinski jwz(a)jwz.org http://www.jwz.org/
jwz(a)dnalounge.com http://www.dnalounge.com/
http://jwz.livejournal.com/
19 years, 1 month
rawhide report: 20050228 changes
by Build System
New package fonts-chinese
Chinese TrueType Font -- Simplified and Traditional Chinese Ming and Kai Face.
Updated Packages:
byacc-1.9-29
------------
* Sun Feb 27 2005 Florian La Roche <laroche(a)redhat.com>
- Copyright: -> License:
cdecl-2.5-32
------------
* Sun Feb 27 2005 Florian La Roche <laroche(a)redhat.com>
- Copyright: -> License
cpufreq-utils-1:0.2-1.1.11
--------------------------
emacs-21.3-25
-------------
* Mon Feb 28 2005 Jens Petersen <petersen(a)redhat.com> - 21.3-25
- add tramp-2.1.3 to site-lisp (David Woodhouse, 149703)
- move removal of info dir to after its installation
- add tramp-init.el to put tramp into load-path
firefox-0:1.0.1-2
-----------------
* Sun Feb 27 2005 Christopher Aillon <caillon(a)redhat.com> 0:1.0.1-2
- Add upstream fix to reduce round trips to xserver during remote control
- Add upstream fix to call g_set_application_name
genromfs-0.5.1-2
----------------
* Sun Feb 27 2005 Florian La Roche <laroche(a)redhat.com>
- Copyright: -> License
jed-0.99.16-7
-------------
* Sun Feb 27 2005 Florian La Roche <laroche(a)redhat.com>
- Copyright: -> License
m4-1.4.2-2
----------
* Sun Feb 27 2005 Florian La Roche <laroche(a)redhat.com>
- rebuild
memtest86+-1.51-1
-----------------
* Sat Feb 19 2005 Warren Togami <wtogami(a)redhat.com> - 1.51-1
- 1.51
rpmdb-fedora-1:4-0.20050228
---------------------------
sox-12.17.6-2
-------------
* Sun Feb 27 2005 Florian La Roche <laroche(a)redhat.com>
- Copyright: -> License:
tftp-0.40-3
-----------
* Sun Feb 27 2005 Florian La Roche <laroche(a)redhat.com>
- Copyright: -> License
19 years, 1 month