AltME groups: search

Qi Ben NanoNote http://sharism.cc/Now we have a need for MIPS Linux 
builds for REBOL :)

Thanks Graham, the 32-bit libraries were not loaded.  This sent me 
on a 24-hour excursion into Ubuntu Linux land where I discovered 
all kinds of things about loading 32-bit libraries into a 64-bit 
operating system that I never thought I needed to know.  After much 
exploration, several trips up and down the Ubuntu Linux technical 
Amazon, I eventually found in a dark, murky backwater the exact technical 
information I needed to complete my journey:


https://bugs.launchpad.net/ubuntu/+source/ia32-libs/+bug/431091?comments=all

On successful completion of the obligatory wget command

cd /tmp

wget http://security.ubuntu.com/ubuntu/pool/universe/i/ia32-libs/ia32-libs_2.7ubuntu6.1_amd64.deb
dpkg-deb -x ia32-libs_2.7ubuntu6.1_amd64.deb ia32-libs
sudo cp ia32-libs/usr/lib32/libstdc++.so.5.0.7 /usr/lib32/
cd /usr/lib32
sudo ln -s libstdc++.so.5.0.7 libstdc++.so.5


And the other commands above, I was delighted to find that rebcmdview 
now works on Ubuntu 10.04 on my laptop.  So, once again, thank you 
Graham for pointing me in the right direction.  I would never have 
figured this out for myself without the 32-bit library information. 
Much appreciated.

I found a second solution and have posted it on my website (so I 
don't lose it.)
http://www.francisayley.org/rebol_on_Ubuntu10.htm

Reason being that core with ssl has not been released as a free product 
... ( I do have it though as I have the sdk )

you'll need to have a bunch of X libraries installed, though

btw, didn't find a way to start REBOL view without my error:
** User Error: Bad face in screen pane!
** Near: size-text self


all I can do right now is use rebface, which doesn't do any graphics 
code on init.

even   rebcmdview -vs   fails with the above error.


note that I'm using rebol within a remote ssh with no desktop setup 
(though I have the Xlibs installed IIRC)

Jun 18 12:50:12 mail kernel: oom-killer: gfp_mask=0x201d2, order=0

Jun 18 12:50:12 mail kernel:  [<c0157322>] out_of_memory+0x142/0x170

Jun 18 12:50:12 mail kernel:  [<c0158f27>] __alloc_pages+0x2e7/0x300

Jun 18 12:50:12 mail kernel:  [<c015a68f>] __do_page_cache_readahead+0x13f/0x2d0

Jun 18 12:50:12 mail kernel:  [<c0151f46>] __delayacct_blkio_end+0x46/0x50

Jun 18 12:50:12 mail kernel:  [<c02f9eab>] __wait_on_bit_lock+0x5b/0x70
Jun 18 12:50:12 mail kernel:  [<c0153640>] sync_page+0x0/0x40
Jun 18 12:50:12 mail kernel:  [<c0153636>] __lock_page+0x56/0x60

Jun 18 12:50:12 mail kernel:  [<c01567ee>] filemap_nopage+0x2fe/0x3c0

Jun 18 12:50:12 mail kernel:  [<c0162c82>] __handle_mm_fault+0x1d2/0x1140

Jun 18 12:50:12 mail kernel:  [<c028acea>] sock_aio_read+0x6a/0x70

Jun 18 12:50:12 mail kernel:  [<c0117f9a>] do_page_fault+0x7ba/0xd64

Jun 18 12:50:12 mail kernel:  [<c01397f0>] autoremove_wake_function+0x0/0x40
Jun 18 12:50:12 mail kernel:  [<c01790e1>] sys_read+0x41/0x70

Jun 18 12:50:12 mail kernel:  [<c01177e0>] do_page_fault+0x0/0xd64
Jun 18 12:50:12 mail kernel:  [<c01058db>] error_code+0x2b/0x30
Jun 18 12:50:12 mail kernel: Mem-info:
Jun 18 12:50:12 mail kernel: DMA per-cpu:

Jun 18 12:50:12 mail kernel: cpu 0 hot: high 186, batch 31 used:107

Jun 18 12:50:12 mail kernel: cpu 0 cold: high 62, batch 15 used:14
Jun 18 12:50:12 mail kernel: DMA32 per-cpu: empty
Jun 18 12:50:12 mail kernel: Normal per-cpu: empty
Jun 18 12:50:12 mail kernel: HighMem per-cpu: empty

Jun 18 12:50:12 mail kernel: Free pages:        2508kB (0kB HighMem)

Jun 18 12:50:12 mail kernel: Active:91716 inactive:257 dirty:0 writeback:0 
unstable:0 free:627 slab:2796 mapped:9 pagetables:434

Jun 18 12:50:12 mail kernel: DMA free:2508kB min:2532kB low:3164kB 
high:3796kB active:366864kB inactive:1028kB present:401408kB pages_scanned:439557 
all_unreclaimable? yes
Jun 18 12:50:12 mail kernel: lowmem_reserve[]: 0 0 0 0

Jun 18 12:50:12 mail kernel: DMA32 free:0kB min:0kB low:0kB high:0kB 
active:0kB inactive:0kB present:0kB pages_scanned:0 all_unreclaimable? 
no
Jun 18 12:50:12 mail kernel: lowmem_reserve[]: 0 0 0 0

Jun 18 12:50:12 mail kernel: Normal free:0kB min:0kB low:0kB high:0kB 
active:0kB inactive:0kB present:0kB pages_scanned:0 all_unreclaimable? 
no
Jun 18 12:50:12 mail kernel: lowmem_reserve[]: 0 0 0 0

Jun 18 12:50:12 mail kernel: HighMem free:0kB min:128kB low:128kB 
high:128kB active:0kB inactive:0kB present:0kB pages_scanned:0 all_unreclaimable? 
no
Jun 18 12:50:12 mail kernel: lowmem_reserve[]: 0 0 0 0

Jun 18 12:50:12 mail kernel: DMA: 9*4kB 5*8kB 0*16kB 2*32kB 1*64kB 
0*128kB 1*256kB 0*512kB 0*1024kB 1*2048kB 0*4096kB = 2508kB
Jun 18 12:50:12 mail kernel: DMA32: empty
Jun 18 12:50:12 mail kernel: Normal: empty
Jun 18 12:50:12 mail kernel: HighMem: empty

Jun 18 12:50:12 mail kernel: Swap cache: add 0, delete 0, find 0/0, 
race 0+0
Jun 18 12:50:12 mail kernel: Free swap  = 0kB
Jun 18 12:50:12 mail kernel: Total swap = 0kB
Jun 18 12:50:12 mail kernel: Free swap:            0kB
Jun 18 12:50:12 mail kernel: 100352 pages of RAM
Jun 18 12:50:12 mail kernel: 0 pages of HIGHMEM
Jun 18 12:50:12 mail kernel: 1992 reserved pages
Jun 18 12:50:12 mail kernel: 83 pages shared
Jun 18 12:50:12 mail kernel: 0 pages swap cached
Jun 18 12:50:12 mail kernel: 0 pages dirty
Jun 18 12:50:12 mail kernel: 0 pages writeback
Jun 18 12:50:12 mail kernel: 9 pages mapped
Jun 18 12:50:12 mail kernel: 2796 pages slab
Jun 18 12:50:12 mail kernel: 434 pages pagetables

check out that you have enough swap space for the processes you are 
running - this will prevent the kernel killing processes, although, 
if you are swapping a lot the system may get so slow that you get 
a worse situation than what you have now. much better would be to 
have enough RAM to fit everything.

Thanks. I have cross-checked. This is a system running in a xen VM. 
I have raised the memory, and updated the app that caused the problem. 
Not it works.

Hi, need an advice. I am setting up very simple CGI, and I use ClearOS, 
and CZ installation. But one of CGI scripts seems to be in UTF-8 
or so, and I think error I am getting has nothing in common with 
Apache or its config. When I press enter on the file in Midnight 
Commander, I get following error (the text is my english translation, 
no exact wording):


./test.cgi: line 1: #!/usr/local/rebol-sdk-cmd/bin/rebol: not 
a file nor a directory


It seems like file is containing an unicode BOM marker at the very 
beginning, so even shebang line can't be interpreted? How can I solve 
it, apart from converting file into some CZ compatible charset?

Have you checked if the file does start wth a BOM? My str-enc-utils.r 
at rebol.org includes a BOM? function. 


 Apache will process a file with a correct BOM - https://issues.apache.org/bugzilla/show_bug.cgi?id=16687

http://sankuru.biz/en/blog/8-joomla-configuration-issues/46-crushing-the-head-of-the-bom-marker-monster.html

- has a 'grep' line that checks if a file contains the BOM, and a 
bash script that can remove it from all files in a directory.

Untested, as I've not seen this myself before.

During boot of Kubuntu 7.10 linux, I noticed a message that flashed 
by, something like
... corrupt .. not cleanly unmounted(?)... 

I checked all the logfiles listed by syslogd-listfiles -a and didn't 
find "corrupt" or "clean" in any of them.
These are the files I checked:
$ lsa `syslogd-listfiles  -a`
-rw-r----- 1 syslog adm  10197 2010-07-09 17:04 /var/log/auth.log

-rw-r----- 1 syslog adm 190194 2010-07-09 16:02 /var/log/daemon.log
-rw-r----- 1 syslog adm 119543 2010-07-09 15:56 /var/log/debug
-rw-r----- 1 syslog adm 210453 2010-07-09 15:56 /var/log/kern.log
-rw-r----- 1 syslog adm 191106 2010-07-09 17:02 /var/log/messages
-rw-r----- 1 syslog adm   8051 2010-07-09 17:02 /var/log/syslog
-rw-r----- 1 syslog adm   3580 2010-07-09 15:56 /var/log/user.log

I'd like to know the way to capture those boot messages.
Any ideas?

(though it's strange it's not in /var/log/messages. maybe it was 
just a fsck message? for some reason the fs was not unmounted correctly 
last time so fsck told you.)

when a file system is not unmounted cleanly, fsck says "corrupt or 
not cleanly unmounted", so, the "corrupt" does not necessarily mean 
bad. in particular, if you didn't notice anything weird, you're most 
likely fine.

Ubuntu is set up do automatically do a fsck check every 30 mounts, 
although it would seem odd if that message pops up on a scheduled 
check..

A question for Linux users: I decided to try to build REBOL on an 
Ubuntu Live boot. I could build Core but not View... because the 
X11 includes are missing... or so it seems.  Tried to run Synaptic, 
but won't load. Any ideas?

You were doing pretty well if you aren't a Linux user :)

A live boot cd is probably out of date ...

you may be able to create your own live cd that does what you need 
but... i think that's more work than installing, or using a virtual 
machine.

Hello everybody, I have a problem with Linux:  DRAW   choose a font 
that give problem displaying:

example: ['text "Hello word!"]
view layout [ box 100x100 effect [draw example ]]

on linux it gives a blank window.
I added:

example: ['text "Hello word!"]
if (pick system/version 4) = 4 [
	fnt1: make face/font [ 
		name: "/usr/share/fonts/truetype/freefont/FreeSans.ttf"
		size: 12
		]
	os: "linux"
	]
if os = "linux" [ insert example reduce ['font fnt1]	]
view layout [ box 100x100 effect [draw example ]]

This solution is good, but using REBGUI  it doesn't work anymore.
What can I do?

Dear ManuM, my problem is with graphic. I have to display a graph 
with lines, rectangules and texts, so I use DRAW. Have you ever used 
DRAW on Linux?

I am assuming this is a linux thing. I am learning really basic stuff 
as a Rebol newbie. When I try running the expamle from http://www.rebol.com/docs/core23/rebolcore-3.html

print read ftp://ftp.rebol.com/pub

I get the following error:

>> print read ftp://ftp.rebol.com/pub
** User Error: Server error: tcp 530 Login authentication failed
** Near: print read ftp://ftp.rebol.com/pub

I was expecting this to work. What am I doing wrong?

they were hit by a hacker who defaced their site ...

or, you can do this

>> trace/net on
>> print read ftp://ftp.rebol.com/
URL Parse: none none ftp.rebol.com none none none
Net-log: ["Opening" "tcp" "for" "FTP"]
connecting to: ftp.rebol.com
Net-log: [none ["220" "230"]]

Net-log: {220---------- Welcome to Pure-FTPd [privsep] [TLS] ----------}
Net-log: "220-You are user number 3 of 150 allowed."
Net-log: "220-Local time is now 20:15. Server port: 21."
Net-log: "220-This is a private system - No anonymous login"
Net-log: {220-IPv6 connections are also welcome on this server.}

Net-log: {220 You will be disconnected after 30 minutes of inactivity.}
Net-log: [["USER" port/user] "331"]
Net-log: "331 User anonymous OK. Password required"
Net-log: [["PASS" port/pass] "230"]
** User Error: Server error: tcp 530 Login authentication failed
** Near: print read ftp://ftp.rebol.com/

Do you have a file FreeSans.ttf at that location?
Look for it with linux shell command:
	locate FreeSans.ttf

I'm not using View for a very long time, but I guess you must first 
make the font object!

MaxV has released a Debian package ... http://synapse-ehr.com/forums/showthread.php?130-Ubuntu-Debian-package&p=764#post764

You haven't specified a file to write to

Thanks Oldes, I'll give that a try.

Maybe there is some better way.. I don't remember it as I'm not using 
such a script.

but the way you have it above you haven't specificed a target file

sorry ... you need a path element in your object

You can tell I haven't done this for a long while either

So I need to add a path in the block statement?

Still no luck. I have looked at over 100+ google pages. Does anyone 
have a working example of a tcp connection using the non-standard 
method? So far I have not found one. Plenty of 'examples', but no 
actual working code. None of the 'examples' work for me. Everyone 
seems to be using the standard short method. Graham: Neither Path 
or Target worked. I checked the connection on my website and cpanel 
said I was logged in, so I have a connection, but I cannot get the 
correct syntax for 'write', so it goes back to the scheme host path 
port user pass path etc block. I'll buy someone a pint (I'm a Brit) 
or a cup of coffe for a working example of that code!

Well, two good friends of mine are moving to NZ at the end of the 
year. I'll ask them to deliver it or buy a fresh one for you when 
they arrive.

my snippet is from a client app of mine and definitively works for 
me

my example allows you to insert just a few bytes at a time and update 
a progress bar, if you replace the 'INSERT line by a loop which only 
copies 2000 bytes at a time until the copy returns none

You have to be brave if you're a soldier in the rebolution!

I'm a Brit living in WA state USA. My sister lives in the midlands. 
I'll visit Sunanda next time I visit the UK.

Looks like I owe two coffees. Give me a delivery address(es) and 
I'll oblige. Graham, you will havo to wait until December/January, 
but that's a Moccachino on its way.

At home we just use those single use Gregg's powdered Mocas that 
you pour into a cup and tip hot water into ... :)

If i get it right, it will prove invaluable to my customers and me, 
as I will have more to offer them. If I don't I can just leave the 
PHP in place. I've been called a practical idealist by some people 
(and much worse by others). Besides, I'll be able to promote Rebol. 
Always a good thing.

I've read about Cheyenne. I probably will when I can code competently 
in Rebol and I have my own server, which are cheap these days. So 
I am on a steep Rebol learning curve right now. I am a fast learner 
when I'm motivated. But Cheyenne will have to wait several months.

I will certainly take a look at Cheyenne before I start doing complex 
stuff, and before I spend time designing this project on paper, which 
I have yet to do. ATM I'm playing around to discover what can be 
done with Rebol.

I think a better solution is to fix the protocols.

still this is a fix right?

this is a fix
 - no, because the %40 is converted to @

It might work if you do use %2540 (double encoding)...

well, it works for me... doing a trace/net .. I saw the username 
being passed correctly to the ftp server

>> read ftp://gchiu%2540compkarori.co.nz:[password-:-ftp-:-rebol-:-com]

URL Parse: gchiu%40compkarori.co.nz password ftp.rebol.com none none 
none
Net-log: ["Opening" "tcp" "for" "FTP"]
connecting to: ftp.rebol.com
Net-log: [none ["220" "230"]]

Net-log: {220---------- Welcome to Pure-FTPd [privsep] [TLS] ----------}
Net-log: "220-You are user number 3 of 150 allowed."
Net-log: "220-Local time is now 03:37. Server port: 21."
Net-log: "220-This is a private system - No anonymous login"
Net-log: {220-IPv6 connections are also welcome on this server.}

Net-log: {220 You will be disconnected after 30 minutes of inactivity.}
Net-log: [["USER" dehex port/user] "331"]
Net-log: {331 User [gchiu-:-compkarori-:-co-:-nz] OK. Password required}
Net-log: [["PASS" port/pass] "230"]
** User Error: Server error: tcp 530 Login authentication failed

** Near: read ftp://gchiu%40compkarori.co.nz:[password-:-ftp-:-rebol-:-com]

Someone, add a ticket to integrate Gabriele's work to R3 !

Adapt, not integrate. We can't integrate it in a compatible way, 
but we can adapt the code to make the existing DECODE-URL work properly.

How can I reference parameters in a linux shell without xargs when 
piping commands?

ls *.o | ar rs mylib.a $1

Where $1 is the output of ls *.o

Maybe try back-tick quoting:

ar rs mylib.a `ls *.o`

The quoted part is evaluated first by the shell.

Alternative to back-ticks that I find more readable (and also allows 
nesting if needed):

ar rs mylib.a $(ls *.o)

Or a multiline version using variables:

OUT=$(ls *.o)
ar rs mylib.a $OUT

I'll do the renaming, but it's a binary. What dependendies I should 
add?

libc6, libstdc++6, libx11-6, libxcb1, libxaw7, libfreetype6

is a more minimal set that will require the rest via transitive dependencies

Very nice. The dependencies for the amd64 Debian package are wrong, 
though. As REBOL still is a 32-bit binary, you'll need to depend 
on the 32-bit compatibility libs.

So to correct this, a depends on "ia32-libs" should be sufficient 
for the amd64 package.

But don't take my word for it, do a ldd on the binary and see for 
yourself :)

I just tried out "tiny linux core" (11 megs). Boots in seconds, installs 
in a minute (frugal), and runs rebol view out of the box....
Includes simple desktop...pretty cool

Great for kiosk/appliance devices running rebol apps

I thought one or more of you gurus would have some advice. I'm trying 
to back up a website via ftp to move all the files to another server. 
Some of the files have owner "apache" and won't let me move them. 
There are a lot of those files in various directories. Not being 
a linux guy, is there an easy way to chown a whole directory with 
subdirectories of files?

chown -R src dest 


and if you can use scp or rcp (depending on your setup) to copy your 
files over, it should be less hassle than ftp which often has many 
issues with stranger filenames and permissions.


on one server, for example, a file was created with ftp, which is 
a legal (albeit twisted) path, but it cannot be access or deleted 
from that same ftp access, because the path gets mangled when it 
goes through the url path parser of that ftp server.


also unless both source and destination ftp servers are the same, 
you can have other nasties.


its always best (and much faster) to gzip your whole directory and 
copy over one file, and then unpack it on the other server (you also 
get a free backup ;-)

Thanks Maxim. I use Putty to access the server though I use it so 
infrequently that each time is like a new experience. I appreciate 
your help.

but remember that if you xfer files to another unix and the user/group 
ids don't exist, you will have to do a chown on the other system 
(unless everything runs as root then, hey who cares right ?  ;-)

Oh, I mean if later the php app cares whether or not a file is owned 
by a different user.

yes, obviously, a chmod is also very important


chown -R   for data should work without problems and its a good thing 
to do.

Before mucking with a recursive chown, I would recursively list your 
files and save into a file, eg:
$ ls -lR > listedfiles
(that first option is a lowercase L ).

Then later if you are worried about having blasted some particular 
permissions, you just refer to listedfiles.

Thanks Anton. That's a wise move as well.

For all Linux gurus:
- I have a xen disk image file
- I want to mount it to read-back some files from it

How to do this? I have tried some stuff but...

- It seems that the images might have some problems

Is there a way to rescue a disk image?

It really looks, that I have lost the data in the image... :-( I 
can't mount nor re-create a partition table.

Is the image containing a single partition or a partitioned virtual 
disk?

If it is a single partition, loop-mounting it as Izkata already mentioned 
should suffice.

If it is a full disk, you'll need to mount the proper partition within 
the disk. There are various approaches possible, depending on your 
actual scenario. You could loop-mount with an offset, if you now 
your disk layout fairly well. Alternatively, use losetup to set up 
a loop device bound to the disk image. Then use kpartx to have the 
kernel read the partition table from this new device and map the 
partitions to their own devices. Finally mount the device corresponding 
to the partition you want to access.

Depends a bit on the details. Feel free to contact me privately.

It's just one partition. The problem is that I can't mount it, as 
the file-structure is damaged. I tried to extract meaningful file 
parts but no luck. I already spent to much time on it. We throw it 
away as we have the data locally and now need to find a way how to 
restore our AltMe content.

I think he wants a publicly available Linux build of R3 with the 
GUI parts included :)

IIRC there exists a command different from FIND that will print all 
absolut file locations for a given filename. Any idea?

Anyone using Dovecot IMAP mailserver here? I need to know if I can 
delete files on the filesystem from the mailbox dirs without bombing 
the mailserver. I have a couple of duplicate files I need to get 
rid of.

Ok, Thunderbird and a plug-in is your friend.

Just redid my netbook with Ubuntu Netbook 10.10, and it works much 
better than in did with WinXP. Had to partition manually as the tweak 
to make the installer SSD friendly is still a proposal. R2 and R3 
work, though (as someone used to the Windows versions) I am having 
trouble with the console not opening a terminal when prompted to. 
I guess the button in the R2 View desktop doesn't work. Time to figure 
out the application shortcut creation method for this distro.

Regarding the console, fire up REBOL directly from a terminal and 
you'll be fine.

Creating a "Application in Terminal" type launcher should do the 
trick.

If the Unity launcher (at least I think that's how the launchbar 
in the netbook thing is called) still uses .desktop files, it's probably 
only a matter of finding the associated .desktop file and modifying 
it to use Terminal=true

Does anyone do anonymous pipes with REBOL? I've found that when doing 
an "insert system/ports/output my-data", that it won't actuallly 
send "my-data" until you halt or quit. Is there a way to force it 
to send so the script can continue to do other things and send more 
data later?

MaxV, It is possible to use call, but then the scripit would be executing 
like CGI. I wanted to pipe two REBOL scripts from the command line 
and have the first script periodically send data to the second. Of 
course I could use a socket, but that's a little more complex.

Hello, I encounter a problem with the draw dialect under linux (rebol/view 
2.7.8). I have tried the following code from the official documentation: 
view layout [
    box black 100x100 effect [
        draw [
            pen red
            line 30x30 50x20 70x70 40x50

            pen blue
            box 20x20 80x80

            fill-pen 0.100.0
            box 20x60 40x80

            pen white
            text 8x25 "Example"

            fill-pen gold
            flood 2x2
        ]
    ]
]

might be a font problem? Better wait for some real experts, though 
...

Better to choose a font which works on Linux... last time I used 
Rebol/View on linux, the default font look was very bad.

But newbies need a working rebol on Linux., then they will choose 
their font. So, what is the default font on Rebol?

? system/view/vid/vid-styles/text/font
helvetica
so, is it enogh  having helvetica font?
No, I made a link to FreeSans.ttf this way:
ln -s FreeSans.ttf helvetica.ttf
but it doesn't work.
I think that the path is wrong. The standard linux path  now is:
/usr/share/fonts/truetype/freefont/

so if Rebol path would be /usr/share/fonts/truetype/freefont/FreeSans.ttf 
, we have resolved all problems!

Is there a way to know Rebol font path?

rebol uses system fonts, so its the other way round, is there a way 
to know system font path.  on linux  no one has done any in-depth 
assessment if there is a single way to find out what are the installed 
fonts for *any* linux.

To clarify the 'font issue' a bit:

In REBOL2 there are two different 'font engines':
engine1:  is used when you set face/text field
engine2:  is used when you use face/effect: [draw [text "..."]]


On Windows both engines use the same OS functions for the font handling. 
Therefore no differencies are seen at the script level. 


On Linux the Engine1 us using some X-windows api and the Engine2 
is using FreeType2 library. 

The X-windows font api doesn't support antialiasing and the font 
output in general is not so nice (depending on the distro).

The FreeType2 lib has better font quality output and also supports 
antialiasing but you have to know the full-path to the font file.


Currently noone suggested any efficient method how to get the 'font 
path' in unified way (considering the 'mess' between different Linux 
distros). 

When I tried to investigate a bit someone pointed me to the FontConfig 
lib that should take care of this problem on Linux. But:

1. I don'r know if the fontconfig is really widely supported and 
what other 'framewrorks' are possible on the rest of distros that 
don't have fontconfig...

2. I don't know how easy/hard is to get the info from fontconfig, 
havent studied it

3. as I said I'm not using Linux for desktop stuff so it is nto my 
high-priority item.


Another solution is to intorduce something like global font-path 
in the Rebol system object which would be used for the font lookup. 
This variable could be set individually by user or author of the 
rebol program. We could also provide some font-name/type->font-file 
matching table but that would take some research as well and won't 
work for all setups. But at least something.


So If anyone here from the 'Linux power users' group knows a good 
solution or even propose a prototype of the solution I believe Carl 
will be glad for your help.

I noticed a difference between Linux and Windows:
Widndows: ?  system/view/vid/vid-styles/text/font    == "arial"

Linux: ?  system/view/vid/vid-styles/text/font    == "helvetica" 
 
Is it normal?

MaxV: It's all about writin a proposal 'how to solve the font issues 
on Linux' then it can be discussed in some more serious way and in 
the end it can be implemented in future releases of R2. So there 
is nothing that blocks you to do the research and propose a solution.

i fear you'll be hard-pressed to come up with a very general solution

I want to use PK authentification with SSH from my client account 
but be able to use two or more different logins. So, one PK for Username-A, 
one for Username-B etc.


How do I set this up? How do I have to name my local private keys 
in that SSH will find them? How and where to store the public keys 
on the server?

I don't think you should care about that, just use:
username-a: "my_string"
username-b: "my_other_string"
call (reform [ "ssh" username-a   mycommands etc] )

However ssh stores files in 
~/.ssh/authorized_keys