gmane.os.plan9.general

Re: public accounts for interested folks at 9srv.net

Anthony Sorace — 2012-05-25T15:26:47

You did not misunderstand. I'm currently traveling. I return late
monday, and will deal with this (and the other similar requests I
got) within a day or two thereafter.

On May 24, 2012, at 4:31 , John Francis Lee wrote:

Re: sed line length limit

erik quanstrom — 2012-05-25T12:11:23

please see /n/sources/contrib/quanstro/sed.c.  i think i've taken
care of all these cases.  there is at least diagnostic output.
i was somewhat concerned about breaking things, so i didn't
let toolong() sysfatal, but that's an easy mod.

- erik

Re: test -w on directory

Charles Forsyth — 2012-05-25T08:48:23

i'd suppose it was probably going to reduce the buffer size on a
wirelessly-connected ipaq

sed line length limit

dexen deVries — 2012-05-25T07:49:56

hi list,


it seems p9p sed has line lenght limit at about 1024 bytes.
however, instead of crashing and burning, or returning some error, it just 
happily spews truncated output.


i've just been bitten by it :D

Re: exec crashing

erik quanstrom — 2012-05-25T05:14:42

no, there's not.  however, rio is a thread-library program and you can't just
exec() out without confusing other processes you're sharing memory with.

rio already has a hook where it will provide an Exit menu option if ".out"
appears somewhere in argv[0].  so, you could have an rc script named rio,
and rename rio to rio.out.  the rc script would just need to read something
like

#!/bin/rc
while()
rio.out $*

- erik

Re: test -w on directory

erik quanstrom — 2012-05-25T04:41:04

in both cases, the test returning the wrong result won't easily
be noticed.  lp just mounts somthing else on /tmp on failure,
and juke restarts/remounts whatever.

i'm not sure anyone has noticed.

by the way, i notice juke has this gem:

kb=4096

say what!?  fittingly, it's unused.

- erik

test -w on directory

2012-05-25T04:07:04

yesterday. anth_x discovered that test -w on directory always
returns false even tho the directory is writable.

test works by calling access(AWRITE) on the file which tries
a open(OWRITE) on the directory. this always fails because
you can't open directory for writing.

some rc scripts seem to assume that test -w should work on
directories. examples are /rc/bin/lp which tests for /tmp
being writable. another is /rc/bin/juke that tests on
/mnt/juke.

i dont know if there are other programs that use access()
to test on directory writability. (i guess not)

its hard for access() to figure out effective directory
writability except by trying to create a file in the
dir (very bad idea). (is access() allowed to return
false positives but no false negatives?)

its not clear to me from the manpage that access() is
expected to work on directories.

access bug? is test at fault here? did it work once?
are these rc scripts that use test -w on directories bogus?

suggestions? :)

--
cinap

exec crashing

Burton Samograd — 2012-05-25T03:01:49

I'm trying to add a restart feature to my modified rio by calling exec on a
saved argv[0] and copy of argv that has been null terminated, but it crashes
every time.  Is there something  about exec on plan9 that I should be aware
of that's different than unix?

--
Burton Samograd

formatting the manual from plan9 ports?

Aharon Robbins — 2012-05-24T19:03:25

Is there a simple recipe to use p9p troff to format the manual pages
from the plan 9 dist?  I.e.:

Install Plan 9 Ports
wget plan9.iso from bell labs
mount said iso /some/where
9 rc
cd /some/where/sys/man
# magic stuff here

(Or tar off the dist to regular files where things are writable.)

Thanks!

Arnold

Re: plan 9 in the cloud - amazon ec2

Richard Miller — 2012-05-24T18:51:01

By simple I meant you don't have to ask permission, just do it.

Andrey's right, technically it's a bit more challenging.  But not much.

Re: plan 9 in the cloud - amazon ec2

andrey mirtchovski — 2012-05-24T18:36:32

can't resist to bastardize the guide. "this is obviously some strange
usage of the word simple that I wasn't previously aware of"

Re: plan 9 in the cloud - amazon ec2

Richard Miller — 2012-05-24T18:24:26

No, it's simple now - the AMI uses pv-grub (paravirtualised grub) as
the "kernel", which then loads the real kernel from an ext2 partition
on your own virtual disk which you control.

Re: plan 9 in the cloud - amazon ec2

Skip Tavakkolian — 2012-05-24T18:11:33

Thank you!  very nice.

Are there any hoops to jump through to import a VM? When I last
checked, it didn't seem possible to load a non-sanctioned OS image.

-Skip

On Thu, May 24, 2012 at 12:47 AM, Richard Miller <9fans< at >hamnavoe.com> wrote:

public accounts for interested folks at 9srv.net

John Francis Lee — 2012-05-24T08:31:08

From: Anthony Sorace <a< at >9sr...>
Subject: Governance question???
Date: Tue, 15 May 2012 12:08:44 -0400

I wrote a couple of emails to a< at >9srv.net but never got a reply ... did I 
misunderstand what seemed to me to be an invitation, or did my email or 
the reply get treated as spam, or have I failed the silent screening?

 > 9srv.net (which I run) provides public accounts for interested folks. 
 > See the wiki for instructions on making a request. It's not as
 > well-developed as tip9ug was, and we've got about a dozen users (same 
 > caveat as above), but it's growing. I would like to do something like 
 > tip9ug's faces page, but haven't gotten there yet.
 > Anthony

plan 9 in the cloud - amazon ec2

Richard Miller — 2012-05-24T07:47:42

I've made a few tweaks to the paravirtualised plan 9 kernel in
/n/sources/xen/xen3 to make it compatible with the xen environment
used in amazon's "elastic compute cloud".  There's an example public
AMI (in zone eu-west-1) named ami-03c9f377 which can be used to
instantiate a cpu server with fossil on a 1GB "elastic block storage"
volume.  This fits within the free usage tier which you get for the
first 12 months after creating an amazon web services account.

When first launched, your server will start with authid = ec2, authdom
and sysname equal to the ec2 instance id, and a randomly generated
password which will be echoed to the system console, where it can be
retrieved - only by the instance's owner - using the Get System Log
function of the aws web interface.  (Be patient; there is often a
delay of some minutes after launching an instance before the console
log becomes available.)

Alternatively, if the "user data" field, which you can create as part
of launching an ec2 instance, contains a shell script (anything
following a line beginning with #!/bin/rc), this will be run from
/bin/cpurc.local at boot time.  You can use this to inject your own
initial authentication details into factotum, by putting something
like this in the user data field:
  #!/bin/rc
  auth/factotum -g 'proto=p9sk1 user=ec2 dom=my.auth.dom !password=XXXX'

The random password or factotum script will allow you to connect to
your server for the first time using cpu, drawterm or ssh1.  (For the
first two, don't forget to open port 17010 or 17013 in the "security
group" firewall.) You'll then want to use auth/wrkey to put new
credentials securely into nvram, and reboot or run 'auth/readnvram
the old temporary password from user data, but only when the instance
is stopped.

For the curious, I've set up a plan 9 instance at ec2.hamnavoe.com,
using authentication domain outside.plan9.bell-labs.com so anyone
with a sources account can cpu into it.  I'll leave it running for
a few days (until my monthly free usage tier quota runs out).

Re: Showing an image on an o/live window

2012-05-23T02:18:45

By the way, have you tried to write, say, canvas panel?

Kenji

Re: Showing an image on an o/live window

Nemo — 2012-05-22T12:49:10

native there means in the host os of
the pc and the namespace you had there. 


On May 22, 2012, at 2:29 PM, kokamoto< at >hera.eonet.ne.jp wrote:

Re: Showing an image on an o/live window

2012-05-22T12:34:08

Oups!

$OCTOPUS/usr/octopus/port/x/xex.b

sorry

Kenji

Re: Showing an image on an o/live window

2012-05-22T12:29:07

This may also be from the different version of inferno, too.
The $OCTOPUS/usr/octopus/port/lib/os.b file shows the
line of
system(xctx, cmd);
and the cmd= "os -d <dir> rc -c <cmd>" here.

According to my inferno man page, "os" command of the
inferno sh has -m mountpoint option, which we dont't use.
If that option is not used, the defdault is at /cmd directory
which would be linked to _local_ bin...

Too much complicated to me for now.

Kenji

Re: Showing an image on an o/live window

2012-05-22T03:22:01

I'm still confusing.

The PC file tree is exported in the inferno environment, styx, so
the exported PC's (this case Ubuntu) root is /usr/octopus, but not
/ of the ubuntu machine.

I issued the ';page' command from 9front terminal, where ';' indicates
the native OS's command.   What is 'native' here?

Actually, when I issue '!' command on the same terminal,
olive try to find the comand under the _local_ /usr/octopus/dis
directory.   I this this is reasonable, but the ';' case is not.
How do you think?

Kenji

malloc

steve — 2012-05-21T20:31:04

i posted before i got to the end of the thad,

i apologise for the noise.

-Steve