Reply-To munging summary (was: Bug in mailing lists; unfriendly to non-subscribers)
by Felipe Contreras
Hi,
I haven't seen a summary of this big thread, so I'm going to try to
write one. I'll focus on Reply-To munging here (orthogonal to
non-subscribers).
Before starting it's worth to keep in mind that munging is a *default*
that is possible to manually change, and some clients have the option
to ignore it. However, the people who do that are the _minority_; most
people just hit reply, and many use typical clients. In order to fully
understand the benefits of one method over the other, we shall only
consider the behavior of the mailing list as a whole, not what some
particular individual might or might not do.
For example, when analyzing personal level indicators, an advantage of
not munging; it's not useful to concentrate on the fact that *some*
people might ignore the Reply-To header, and therefore the indicator
would appear for that particular mail. The fact remains that for most
of the posts in the mailing list, the feature will not work. IOW: what
is important is whether or not the user can rely on the feature.
== Advantages ==
1) Personal level indicators.
Gmail has a feature that places a neat marker (">") near the subject
of the mail so that you can see directly that the mail was addressed
to you. This is specially a nice feature in high traffic mailing
lists. Here's a screenshot:
http://people.freedesktop.org/~felipec/l-o-personal-marker.png
mutt can be configured to have similar highlights:
color index brightwhite black "~p" # to me
2) Recipient filtering
It's possible to search say 'to:linux-kernel@vger.kernel.org to:me';
with munging it's not possible because the To header is automatically
changed to the mailing list.
Similarly, it's possible to have a filter rule for all mail from
linux-kernel to move to a certain folder, but have an exception for
mail addressed to the user, which might go directly to the inbox.
This way, both with personal level indicators and recipient filtering,
it's much easier to quickly be alerted of incoming mail addressed to
oneself.
3) Consistency
When there's Reply-To munging, people might choose to ignore it (if
their client supports) that. So in a typical munged mailing list you
have both munged and non-munged posts.
This is not a strong argument, but it's useful to keep in mind while
reading the non-issues below.
There are other advantages, not not relevant in the current Fedora context.
== Disadvantages ==
1) Inadvertently forgetting to "reply to all"
When Reply-To munging is not used, the user is presented (in typical
clients) with the option to either "reply" or "reply to all".
Sometimes the user will select the wrong one, which is an annoyance.
The severity of this problem is impossible to calculate as it depends
on background of the members of each mailing list, which might be very
varied, and in flux.
=== Non issues ===
Many arguments were presented as disadvantages of munging which are
not valid IMO and I presented reasons for that (as I'm doing here).
1) Ignorance of "reply to all"
While there might be some rare cases where the user doesn't know about
"reply to all" (i.e. mail newbies, grandmas, etc.), the typical mail
user is well aware of "reply to all" as it's the only way to keep a
conversation between more than two persons. It's probably one of the
first things to learn.
Personally I have never seen cases of this, and nobody has provided
evidence. If there's any cases of this, they are probably very
marginal.
2) Two copies of the same mail
When a mail is not munged, the original author would be in the To
field, and the list on Cc. If the author is subscribed on the mailing
list (the only case we care about here), then the author would receive
a direct copy and another copy from the list.
However, this is not the case as mailing list servers are aware of
these cases and don't send the list copy. For example, mailman has an
option "Avoid duplicate copies of messages" which does precisely that,
and is enabled by default.
This was agreed to not be an issue.
3) Unwanted direct addressing
There was an argument that some people didn't want to be directly
addressed and wanted mails addressed to the mailing list only.
There was no mention of why people would want that, or any evidence of
such cases (if there's any).
Moreover, Reply-To munging doesn't prevent them from being directly
Cc'ed (that header filed is not munged), they can manually be added to
the To field, or somebody might ignore the Reply-To header.
At best this is a policy issue, not related to Reply-To munging.
4) FWD noise
If the user mistakenly misses the "reply to all" option, the mail is
sent privately, then the user might notice and choose to fwd the mail
back to the mailing list.
This is a double mistake by the user, the proper way to correct the
original mistake is to send the mail again properly, and as to discard
the private copy.
Just like top-posting; this is solved with policy.
5) Mail escaping
This is defined as mail that is sent privately by mistake and is never
noticed, hence escaping the mailing list.
No evidence was provided of this considerably affecting any mailing
list, so at best the damages are marginal.
== Personal opinion ==
My feeling is that Reply-To munging is being widely used because it's
being widely used; IOW: inertia. It seems like a good idea at first,
until you realize the loss of functionality. Most people don't realize
the features they are missing, and instead only concentrate on the
convenience of not having to choose between two options.
== Conclusion ==
Once you weed out all the invalid arguments, the analysis is very simple:
1) Not munging allows some nice features
2) The only thing munging does is prevent inadvertent private replies
So it's a matter of putting in a balance what is more desirable: the
features, or preventing the mistakes. IMO the result is obvious, since
the people that would benefit from the features would do it in a very
big way, and the amount of people that would be affected by mistakes
is not known.
Also, people that are used to Reply-To munging can get used to
clicking "reply to all"; it's a matter of habit, but no functionality
is lost. OTOH the people that enjoy the benefits of not munging can't
really do anything to make the features work; it's not possible,
period. So IMO munging advocates are being egotistic by not giving
away a tiny bit of their convenience in order to enable features other
people love and rely on for efficient mail handling.
Cheers.
--
Felipe Contreras
13 years, 11 months
kickstart problems with f13
by Chris Rouch
I have a heavily customised installation (lots of locally built rpms)
that i'm trying to install from kickstart and nfs. If I use a netinst
cdrom and pass the nfs location of the kickstart file without using
dhcp then this all works well. But If I use pxe/dhcp/tftp instead, it
gives me the "retrieving installation repo" message and then complains
about missing packages and installation groups.
If I switch the VT2 I can see that on the working installation I have
the remote nfs directory mounted on /mnt/source. However on the PXE
installation, this directory is mounted on /mnt/isodir and /dev/loop1
is mounted on /mnt/source. This latter contains a completely
different set of packages. This is completely repeatable - if I
comment out the dhcpd.conf entry and restart dhcpd everything works.
If I leave it in (even when booting from the cdrom) then it always
fails
I'm seeing this from a vmware client, but I also see the PXE error
from a real laptop.
How do I stop this happening?
I don't have a any repo= lines in the ks.cfg or on the kernel command
line. The ks.cfg has the following to specify the installation media:
nfs --server 192.168.6.7 --dir /data/distros/f13/cdrom
Any help will be appreciated.
Regards,
Chris
13 years, 11 months
VBOX problem
by RAMAKISHOREBABU KOPPULA
I am using VBOX with F13 to run Windows XP. Today I got following error:
Start tag expected, '<' not found.
Location: '/home/rkbabu/.VirtualBox/Machines/WinXP/WinXP.xml', line 1 (0),
column 1.
/home/vbox/vbox-3.2.0/src/VBox/Main/MachineImpl.cpp[664] (nsresult
Machine::registeredInit()).
Result Code:
NS_ERROR_FAILURE (0x80004005)
Component:
VirtualBox
Interface:
IVirtualBox {3f36e024-7fed-4f20-a02c-9158a82b44e6}
and now it is not working.
Please someone help me to solve this problem.
Regards.
Kishore
13 years, 11 months
Q about swap size
by Robert G. (Doc) Savage
I've been running Fedora on my old Thinkpad A22p (P-III/M w/512MB) since
FC4 or before, upgrading from one version to the next using the same
partition sizes. For F13 I increased the size of /boot, but kept swap at
1GB:
# fdisk -l /dev/sda
Disk /dev/sda: 120.0 GB, 120034123776 bytes
255 heads, 63 sectors/track, 14593 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0x000304c6
Device Boot Start End Blocks Id System
/dev/sda1 * 1 64 512000 83 Linux
Partition 1 does not end on cylinder boundary.
/dev/sda2 64 14463 115658752 83 Linux
/dev/sda3 14463 14594 1048576 82 Linux swap / Solaris
The last couple of releases have run agonizingly slowly with very high
load averages (between 2.x and 10.x).
I just read
http://docs.fedoraproject.org/en-US/Fedora/13/html/Installation_Guide/s2-...
which recommends a minimum of 2GB swap size for memory sizes up to 4GB.
I've been using 2x RAM for swap size ever since the 2.6 kernel was
introduced. At the moment (running 'yumex') 'top' indicates:
top - 10:43:15 up 2 days, 6 min, 5 users, load average: 8.62, 6.84, 6.26
Tasks: 244 total, 7 running, 237 sleeping, 0 stopped, 0 zombie
Cpu(s): 60.7%us, 23.5%sy, 9.2%ni, 0.0%id, 0.0%wa, 5.4%hi, 1.3%si, 0.0%st
Mem: 509800k total, 502624k used, 7176k free, 16280k buffers
Swap: 1048568k total, 317372k used, 731196k free, 101944k cached
This indicates only about 30% of my 1GB swap is being used. Would
increasing that to 2GB have any affect on the load average performance?
--Doc Savage
Fairview Heights, IL
13 years, 11 months
F12 -> F13. Preupgrade Error
by Dave Cross
I'm trying to use preupgrade to upgrade my Dell Studio XPS from Fedora
12 to Fedora 13.
It got through the bit where it downloads all of the required data
from the internet and prompts me to reboot the PC.
I reboot the PC and select "upgrade to Goddard" from the grub menu.
After a while, I then get the message:
"Unable to find any device of the type needed for this installation
type. Would you like to manually select your driver or use a driver
disk."
There's an option to choose a driver from a list. But the list as huge
and seems to contain entries for all sorts of drivers - network cards,
storage devices, probably others - I don't know what I should be
choosing.
Can anyone suggest what might be going wrong here? And what I can do to fix it.
Thanks,
Dave...
13 years, 11 months
problem playing vcd/dvd
by anees a a
not able to view or copy from original vcd and dvd of films
following errors shown for default movie player
could not read from resource
no uri handler implemented for vcd
and for vlc following error shown
file reading failed
pls suggest solutions
13 years, 11 months
HP Deskjet F4140 , will not setup in FC13
by Jim
FC13-i686/KDE
Can't install a HP Deskjet F4140 in FC13.
I have installed;
1. All Cups rpm's
2. Gutenprint-cups
3. All the hplip rpms
4. hpijs
And I still can't get the HP Deskjet F4140 to show up in Selection Window.
I have even run HP-Setup and it can't install the F4140 .
HP-Setup can see the F4140 is on USB port but can't install it.
I went to the HP website and it says the printer is Supported in the
hplip drivers
13 years, 11 months
[F13] KeyboardSignal not working (Upstart/init)
by Andreas M. Kirchwitz
Hello Fedora Users!
Similar to Control-Alt-Delete triggering a "reboot", I would like
to configure Control-Alt-Insert to run "poweroff" when I'm on a
regular Linux console (no GDM or other graphics stuff running).
Up to Fedora 8, that was easy. Just put this to /etc/inittab:
ci::kbrequest:/sbin/shutdown -t3 -h -P now
And run this is /etc/rc.d/rc.local:
# 82 + 110 are the two Insert keys on regular PC keyboards
cat << EOF
control alt keycode 82 = KeyboardSignal
control alt keycode 110 = KeyboardSignal
EOF
) | loadkeys
In Fedora 10 to 12 (upstart), I had to put this in /etc/event.d/kbdrequest:
start on kbdrequest
exec /sbin/shutdown -h -P now "Control-Alt-Insert pressed"
But what about Fedora 13?
I tried to create /etc/init/control-alt-insert.conf like this:
start on keyboard-signal
exec /sbin/shutdown -h -P now "Control-Alt-Insert pressed"
But I would not ask here if it worked. ;-)
The keyboard map seems to be okay. If I configure Control-Alt-Insert
to output a simple string like "Hello World", that works perfectly fine
on all Linux text consoles.
It's Upstart's "init" process that doesn't seem to recognize the
KeyboardSignal event in Fedora 13. Instead of "start on keyboard-signal"
I've also tried "start on kbdrequest" and "start on kbrequest". No luck.
I tried pressing Control-Alt-Insert on all six console ttys. No luck.
What's wrong here?
Help appreciated ... Andreas
13 years, 11 months
Converting DVD copy to AVI
by Patrick O'Callaghan
I have a DVD structure (VIDEO_TS etc. etc.) on my hard disk and would
like to convert the main feature to an AVI file, i.e. basically what
rippers do but without the physical DVD. I know I can use ffmpeg or
transcode to convert individual VOB files, but this seems error prone
(e.g. how do I know I got all of the main feature?). Does anyone have a
script for this sort of thing? Unfortunately K3B is no good as it
insists on there being a physical DVD to rip from.
poc
13 years, 11 months
Thynderbird sounds -
by Bob Goodwin
I can tag Thunderbird received messages with color but what I
really need is to tag them with a sound. I would like to play a
unique sound as the output of an e-mail filter.
I can hear the incoming mail tone from other parts of the house
but I would like to know when the message goes into one of the
filter directories I've created, else if I am waiting for a
response I have to keep coming back to the computer to see if it
signals the awaited message.
Is there a way I can do that?
Bob
--
13 years, 11 months