Rubyx on OS News

[trans. (T. Onoma)]

Rubyx splotlighted on OS News.

http://www.osnews.com/story.php?news_id=8530

Andrew Walrond is very good coder. In fact I've looked at the Rubyx script in
detail and it's about the most compact code I have ever seen --quite amazing,
albeit I don't necessarily recommend coding in that style generally,
nonetheless it is an impressive piece of work.

T.

(I have one major caveat about Rubyx though, I don't like how all the
"packages" are stuffed into that one script.)

 

[Mauricio Fernández]

Rubyx splotlighted on OS News.

http://www.osnews.com/story.php?news_id=8530

Andrew Walrond is very good coder. In fact I've looked at the Rubyx script in
detail and it's about the most compact code I have ever seen --quite amazing,
albeit I don't necessarily recommend coding in that style generally,
nonetheless it is an impressive piece of work.

I am very interested in the rubyx script, but I couldn't find any simple
way to get just that file (without having to download a full ISO) last
time I tried.

I would be most grateful if somebody could give me a copy ;-)

 

[Andrew Walrond]

(I have one major caveat about Rubyx though, I don't like how all the
"packages" are stuffed into that one script.)

No longer the case

Andrew Walrond

 

[Andrew Walrond]

I am very interested in the rubyx script, but I couldn't find any simple
way to get just that file (without having to download a full ISO) last
time I tried.

I would be most grateful if somebody could give me a copy ;-)

I've put a copy here

ftp://ftp.walrond.org/rubyx/

Andrew

 

[trans. (T. Onoma)]

30, trans. (T. Onoma) wrote:
| > (I have one major caveat about Rubyx though, I don't like how all the
| > "packages" are stuffed into that one script.)
|
| No longer the case

Ooo.. that's promising.

I tired to download using ww. It took ~24 hours to download cd iso, but when
it finished it gave:

0 remaining
ww.execute() caught an unexpected exception: Loader exited due to error:
Extracted file digest is incorrect: rubyx-82-i386-cd.iso

So I do not have much hope that it will work. I don't know the details but it
sure would be nice if it could go back and fix itself, (rsync?) or something.

Have you considered trying linuxiso.org? Maybe they will mirror a copy.

Also, I found article on chrooting services that you might be interested in:

http://www.linuxfocus.org/English/January2002/article225.shtml

I know Rubyx does something like this for building software, I wondered if it
did so for running services too --might be nice.

Thanks,
T.

 

[Bill Atkins]

What is the advantage of serving Rubyx through White Water rather
than, say, BitTorrent, which is much more common?

I think Rubyx might be interesting, but the obscure download method is
going to cause some people to just move on.

Bill

 

[trans. (T. Onoma)]

On Wednesday 13 October 2004 11:41 am, Bill Atkins wrote:
| What is the advantage of serving Rubyx through White Water rather
| than, say, BitTorrent, which is much more common?
|
| I think Rubyx might be interesting, but the obscure download method is
| going to cause some people to just move on.
|
| Bill

Rubyx used to use BitTorrent actually. I'm not sure why Andrew wrote his own.
It doesn't bother me --in fact I kind of like ww since it is just one
program, I was able to get it working in minutes. BitTorrent, on the other
hand, well, I'm still not sure how to do. That might seem dumb on my part,
but you know, I only have so much time to figure things out. When I hit
bt<tab> in my shell I get:

btcompletedir btlaunchmanycurses.bittornado
btcompletedir.bittornado btmakemetafile
btcompletedirgui btmakemetafile.bittornado
btcompletedirgui.bittornado btmaketorrentgui
btcopyannounce btreannounce
btdownloadcurses btreannounce.bittornado
btdownloadcurses.bittornado btrename
btdownloadgui btrename.bittornado
btdownloadgui.bittornado bts
btdownloadheadless btsethttpseeds
btdownloadheadless.bittornado btshowmetainfo
bterm btshowmetainfo.bittornado
btlaunchmany bttrack
btlaunchmany.bittornado bttrack.bittornado
btlaunchmanycurses

Spare me.

I wish ww was written in Ruby though!

T.

 

[Mark Hubbart]

On Wednesday 13 October 2004 11:41 am, Bill Atkins wrote:
| What is the advantage of serving Rubyx through White Water rather
| than, say, BitTorrent, which is much more common?
|
| I think Rubyx might be interesting, but the obscure download method is
| going to cause some people to just move on.
|
| Bill

Rubyx used to use BitTorrent actually. I'm not sure why Andrew wrote his own.
It doesn't bother me --in fact I kind of like ww since it is just one
program, I was able to get it working in minutes.

well, I guess you're lucky. I can't get whitewater to compile on my
system (powerpc/darwin) so I couldn't download the iso. That ended my
experiment in Rubyx :/ Perhaps I can install it just using the rubyx
script mentioned above? Perhaps I'll see...

cheers,
Mark

 

[Andrew Walrond]

well, I guess you're lucky. I can't get whitewater to compile on my
system (powerpc/darwin) so I couldn't download the iso. That ended my
experiment in Rubyx :/ Perhaps I can install it just using the rubyx
script mentioned above? Perhaps I'll see...

I'm happy to work with you to get this fixed; I'm already working with another
user on PPC. (I only have x86 and x86_64 other-endian test machines).

What compiler are you using, and what errors does it report? Note that gcc
3.3.1 or better is a requirement (see INSTALL file). Older gcc's have lots of
fundamental c++ bugs which break ww.

You might move this discussion to the ww or rubyx mailing lists, or to my
personal mail since the traffic here is already enormous

Andrew

 

[Andrew Walrond]

I tired to download using ww. It took ~24 hours to download cd iso, but
when it finished it gave:

0 remaining
ww.execute() caught an unexpected exception: Loader exited due to error:
Extracted file digest is incorrect: rubyx-82-i386-cd.iso

I'm afraid the iso hasn't many proxy mirrors yet, and the demand on my home
adsl has been huge, hence long download times. The situation is better now
(more mirrors)

What version of ww are you using? If you used a permanent cache (so you can
resume if a problem occurs like above) then just run it again (with
--verify-cache) and all will be well in a few moments. Use the latest version
if you aren't already.

So I do not have much hope that it will work. I don't know the details but
it sure would be nice if it could go back and fix itself, (rsync?) or
something.

I can and does; but older versions had a bug where corrupt blocks didn't get
deleted on huge downloads, but that is fixed now. Even so, a quick rerun with
--verify-cache sorts the problem.

Have you considered trying linuxiso.org? Maybe they will mirror a copy.

No; perhaps I should. I haven't really been encouraging use of the ISO's; The
preferred method is to build your own custom distro from source using just
the rubyx script (See www.rubyx.org).

Also, I found article on chrooting services that you might be interested
in:

http://www.linuxfocus.org/English/January2002/article225.shtml

I know Rubyx does something like this for building software, I wondered if
it did so for running services too --might be nice.

Yea; I need some technology in the linux kernel which will make rubyx really
elegant (no symlinks) and chrooted services efficient and elegant. Namely,
extended union mounting, or mutiple level transparent mounts. It's a pet
project I will do one day, unless somebody gets there first...

Andrew Walrond