World: r3wp

anyway, i prefer just working for the shell most of the times.

Is this going to break anything for getting a valid status code out 
of call?

cap: func [calls /local sherr status] [
    sherr: copy ""
    append calls { ; echo -n $? >&2}
    status: call/wait/info/error calls sherr
    status/exit-code: to integer! find/last/tail sherr newline
    status
]

Of course this is stripped down  (the whole /console /output etc, 
combination thing to handle), but it seems to be working for getting 
at exit status.  Dirty pool? cheap trick? or functional?

Gabriele; any hints or comments?

Update;  set -o pipefail  needs to be set in the forked shell to 
get proper pipeline status.
...
insert head calls {set -o pipefail ; }
append calls { ; echo -n $? >&2}
...

{ ; echo -ne "\n$?" >&2}  works better...when you test with commands 
that don't fail or fail without stderr  :)  gotta love being not-quite-right 
all the time.

it is an acceptable workaround in most cases i guess.

Henrik;  Remember a while back you got a  Killed  message?  I just 
got one, and here is a snippet from the kernel log.
       1 dev kernel: oom-killer: gfp_mask=0x200d2, order=0
...addresses and other cracker info removed...

       1 dev kernel: Out of Memory: Kill process 6448 (bash) score 95256 
       and children.
       1 dev kernel: Out of memory: Killed process 6469 (orebol).

I'm guessing you had the same scenario.

btiffin, ok. I can hardly remember it now :-)

http://osnews.com/comment.php?news_id=18304&offset=0&threshold=&msg=8#257159


I love the "I wonder..." thread. It says a lot about Linux development.

Yeah. For the record, we in the Syllable project don't think this 
means much. It's a small addition to the Linux kernel that can't 
be used for much

http://apcmag.com/6735/interview_con_kolivas<--- interesting interview 
with a guy who recently quit linux kernel development. Amiga is mentioned.

Yes, an interesting read. I didn't know, things were that bad.

as a Linux newbie, I can attest that Linux for the user is bloated.To 
get more people to use Linux,they need to put the stuff that Joe 
Sixpack wants

Linux has been there for so many years, so I don't see that happen 
in the future. Some of the code might be emigrated to other OSs with 
other designs, but then it isn't Linux anymore. As I see it, Linux 
as a desktop OS is a dead end.

I think the main thing that a lot of linux developers won't see, 
is that in the long run an "evolution" approach to development (code 
something and see if it'll float or sink), doesn't work as well as 
the "intelligent design" approach. I could see that in a discussion 
recently on OSNews.

I'm seeing Linux kernel development as this big mountain of clay 
and 100s of people are throwing pieces of clay on it to make it bigger. 
:-)

if the clay sticks, it's good code.

Funny, I was thinking of the "evolution" approach also, when I read 
the article. I agree, the "evolution" approach is a bad idea, when 
dealing with computing code. "Keep it simple!" is what they often 
forget.

the thing is, one thing is to do it generally, because you don't 
have the energy for design, but another thing is to actually defend 
it as a good development method. that is why I think also that Linux 
as a desktop won't move forward. Perhaps now the best parts of Linux 
is actually what came out in the beginning, with the original GNU 
tools and not new tools being developed today.

I'm of a different opinion.  GNU/Linux will make it to the desktop. 
 It IS my desktop.  Vista is buried under the GRUB and that's were 
it stays.  Buried.  Many many people are working on the Desktop issues 
right now....like right now...3:15pm eastern standard time.  Con's 
experience may just be a matter of bad-timing.  The first brave front 
runners get cut down, then the masses have a way of getting to the 
goal.  His efforts will not go unnoticed, but he may feel nothing 
but pain and misery for it.  Human nature.

Sounds like Con wrote excellent clay and was thrown out by the cathedral-designers. 
Not the other way around. Seems linux-kernel lost its bazaar-abilities.

BTW a similar scheduler was written by a hig priest and was quickly 
included. quite unfair. and the article is about kernel, not desktop, 
as con mentions in comments

Too many cooks in the kitchen

and the problems interest-problems stem from interest of the developers, 
which are enterprise-driven today.

not according to that article.

rebols have to worry about the same disenfranchisement imho.  Somedays 
it seems the lure of wxPython or Tcl/Tk, or D, or LUA, or even frick'n 
Windows ... would be an easier long term road to get and keep happy 
customers.  That feeling passes with the next little nugget of knowledge 
(and a little stubborness).  REBOL WILL get RIF, REBin, LNS, Rebcode, 
formatted strings and all the cool stuff that seems just around the 
corner.  AND it will all run on Debian GNU/Linux and I'll be able 
to dump Wine for once and for all, and then slowly convince people 
to move away from the darkside.  :)


Can you tell I'm a little s and t of cleaning off people's PC's just 
to watch them click up IE and Outlook and MSN AGAIN.

Volker:  You must also remember that everyone is human - I think 
that's what most people forget.  How he presented his code to the 
developers could easily have been the reason they didn't include 
it.  I haven't seen the kernel mailing list myself, so I can't say 
for sure, but - he said it's a very unfriendly place, and so likely 
positioned himself accordingly, defensively.  He said himself that 
defended his work as much as possible - without budging.  That's 
the kind of person I hate working with, and the kind of person I 
actively ignore.

And, I agree with btiffin - the Linux desktop is getting there, slowly 
but surely.  Especially Ubuntu, which my *mom* wants me to install 
when she gets her laptop, after having checked out mine a couple 
of times.  I only use Windows for video games that don't run well 
under WINE.

enterprise versus joe sixpack is a big problem. I can understand 
the enterprise model to get into the Fortune 500 but then they have 
to have the joe model for the workers/users. For the Linix office 
worker, a browser/oo/few games/email/ is about all they need at work. 
At home add more games/Gimp/irc/torrent client/media apps and most 
will be happy. It's true those distros are out there but for joe 
sixpack, they are hard to find and most joes only know the big Linux 
names

To be honest, I'm kinda of pro/con with open source in the fortune 
500 server room.  If they kick back some of the savings into funding 
some developers....yeehah.  If they use OSS simply to increase the 
bottom line and only train their elites in it's use ... then boohiss. 
 It is the small, in their own basement  developers and the rare 
business tycoon that supports them that this issue teeters on.  Can 
freedom reach a critical mass, or are we doomed?  Today I'd bet on 
doomed, but will rail against it the whole way down.


And while I'm here, preaching to the choir, I'll cheerlead for those 
brave souls that will buck the trends at their own risk and expense. 
 Go Doc Go.   Give me an A, give me an l, a t, an M and and e.  Ashley, 
Sunanda and all the rest...GO REBOL.  All rebols and gnubies reading 
this, pat yourself on the back for the remarkable achievement of 
seeing the light through the glare.  And other stuff.

I think we found a REBOL evangelist!

disestablishmentarianism - it's the only way to go.  :)

btiffin -- Did you ever hear back from Leo Laporte regarding trying 
out REBOL?  I read about your question to him on the Lab a while 
back.

Nope.  No word.

no offense.. but Laporte doesn't seem to me to be the kind of guy 
who follow up on things like REBOL.

You're right.  It's not really in the realm of his target audience.

he has a big audience though

Yep. I like watching the show and try to lean people that run small 
businesses toward watching.

I read that Con Kolivas interview, and I don't want to sound repetitive, 
but the answer is right there in front of everybody. Or right here, 
in this case

I've been playing with Kubuntu on and off for a month or so. Sometimes, 
after updates to the system, something breaks.

I would like to know what is the best way of backing up and restoring 
the system.
I think the best way might be to copy partitions back and forth.
I have a few LiveCDs like the System Rescue CD.
What do you guys do ?

Backing up whole partitions can be good if you have the space

When you do a new installation, it's good to put your /home directory 
on a different partition. That way, you can backup just your data, 
or wipe the root partition and re-install the system withoug loosing 
your data and settings

Use Debian  :)  My Dev system has been auto-updating since Potato 
(2.2), I rarely use Unstable packages, but Dev lives in Testing and 
I've not had any problems.  Spike and Chester stay with Stable sources.list 
and I trust the Debian development model (not released on any time 
schedules, but only releasesd when the DD's have finished bashing 
each other around).  I hope they fight the urge to go with release 
dates, even Ian Murdoch is pushing for release shedules :(  but he's 
at Sun now, so I hope that pressure subsides.


To be more to the point.  I keep all my work (and user accounts) 
rsynced on three different machines, backed up to CD and USB memory 
stick on a semi-irregular basis.  I plan for system restore from 
the boot disks, the list from apt-show-versions and a documented 
list of manual steps for things like the Linksys, /etc/hosts, /etc/ssh/sshd_config 
et al and Cheyenne config if anything does eventually go sideways. 
 It might be a (much) longer restore cycle, but I trust that more 
than backup/restore mirroring.  I've avoided the VM model, but that 
seems to be getting some good press now-a-days.

kubuntu and updating without problems here, but, i've seen that happen 
in the past. i always managed to fix stuff though.

Ubuntu has been a big disappointment to me just because I thought 
I would finally be able to upgrade

My opinion, release dates are going to hurt them in the long run...great 
for getting new adopters...bad for long term guaranteed stability

Much like R3...I'd prefer disappointment in "missed" dates than any 
guaranteed release schedule.  Patience is a virtue...release code 
when it's ready. not before, and never under the urging of some project 
manager (the common case).

Well, we have had a lot of success with Syllable by releasing early 
and often

Managed expectations :)  You haven't stamped any "shrinked wrapped, 
use this to run your business" versions yet, have you?  I'd urge 
your team to lean toward the Debian model over the Ubuntu schedule 
model.  But test releases...yep as often as possible, but even then, 
not because a manager said "It goes out on Friday".

We're just discussing someone who wants to automate his company with 
Syllable :-)

We want people to be able to rely on us. That includes not just stability, 
but also schedules