Linux-Advocacy Digest #587, Volume #32 Thu, 1 Mar 01 19:13:04 EST
Contents:
Re: The Windows guy. (Steve Mading)
Re: M$ Worker to Thug Ratio (Chris Ahlstrom)
Re: Information wants to be free, Revisited ("B.B.")
Re: The Windows guy. ("Edward Rosten")
Windoze Domination/Damnation (Chris Ahlstrom)
So, here's something to chew on... ("Masha Ku'Inanna")
If I delete using rm? ("Interconnect")
Re: KDE or DOJ ? ("Gary Hallock")
Re: SSH vulnerabilities - still waiting [ was Interesting article ] (Markus Friedl)
Re: The Windows guy. (Steve Mading)
Re: Something Seemingly Simple. (Steve Mading)
Re: MS websites: a tale of total and humiliating failure! (.)
----------------------------------------------------------------------------
From: Steve Mading <[EMAIL PROTECTED]>
Subject: Re: The Windows guy.
Date: 1 Mar 2001 22:53:36 GMT
The Ghost In The Machine <[EMAIL PROTECTED]> wrote:
: In comp.os.linux.advocacy, Edward Rosten
: <[EMAIL PROTECTED]>
: wrote
: on Thu, 01 Mar 2001 01:09:04 +0000
: <97k7gb$h1s$[EMAIL PROTECTED]>:
:>>>> A single-tasking OS is fundamentally incapable of fulfilling this
:>>>> definition properly.
:>>>
:>>>
:>>> A simpler definition is:
:>>>
:>>> a mechanism which allows the output of one process to be put in to the
:>>> input of another process in the order that it (the data) was outputted.
:>>>
:>>> DOS pipes still do not satisfy this definition.
:>>
:>> Pedant point: Yes, they do. While the data is long stale by the time
:>
:>Not quite.
:>
:>There are examples of programs I can run where the output of the first
:>program will never reach the input of the second under DOS pipes, but
:>will always work under a multitasking syetem (under my definition of
:>pipes). Since DOS pipes do not fit the definition, they are not pipes.
:>
:>
:>prog_that_will_never_finish | some_other_prog
: Oops, my bad. :-) You're right; if the first program never
: finishes, the second program will see nothing.
Technically what matters is that it never finishes the file,
not that the program itself never finishes. The following
(obviously contrived) example allowed tail to see the output,
even though the feeder program never finished:
At the prompt:> gcc spew.c -o spew
At the prompt:> ./spew | tail -10
The code for "spew.c" is as follows:
#include <unistd.h>
#include <string.h>
int main( int c, char **v )
{
int x;
char buf[80];
for( x = 0 ; x < 10000 ; x++)
{
/* NOTE: avoiding stdio functions.
* I wanted to see the effect without the
* buffering that stdio does.
*/
sprintf( buf, "iteration %d\n", x );
write( STDOUT_FILENO, buf, strlen(buf) );
}
close( STDOUT_FILENO ); /* Remove this, then tail gets no data */
while( 1 ) { /*loop forever*/ }
return 1;
}
------------------------------
From: Chris Ahlstrom <[EMAIL PROTECTED]>
Subject: Re: M$ Worker to Thug Ratio
Date: Thu, 01 Mar 2001 23:00:39 GMT
Bill Antigates wrote:
>
> If you were to break down M$'s headcount & payroll, how much of it actually
> gets to the software "engineer"? And how much of it is spent defending the
> empire through lies, deception, misleading magazines, corny "right to
> innovate" ads, and etc?
I like your cooooool modified message header!
X-browser: Microsoft Innerbutt Exploiter 7.1Beta
X-Newsreader: Microsoft OutaLuck Express 5.00.30.24.99.whatever.dude
X-MimeOLE: Code Written and Composed By Bill
"Bling-bling" Gates, Chief Software Architect
& RapArtist-PeaceOut!
X-Microsoft_UserPermit: PC subject to random inspection. Exp:July2001
X-Accept-Language: ebonique, anglais, francaise
I'm posting it here for the Outlook users to see.
Chris
--
[X] Check here to always trust content from Chris
[ ] Check here if you're a dazed follower of Bill Gates
------------------------------
From: "B.B." <[EMAIL PROTECTED]>
Crossposted-To:
comp.sys.mac.advocacy,comp.sys.next.advocacy,comp.os.ms-windows.advocacy
Subject: Re: Information wants to be free, Revisited
Date: Thu, 01 Mar 2001 17:06:10 -0600
In article <[EMAIL PROTECTED]>,
Aaron Kulkis <[EMAIL PROTECTED]> wrote:
@> Either way, that really has nothing to do with wether or not you're
@> wasting huge quantities of bandwidth. To address that: You're not--it's
@> only text. Just one of the millions of binaries posted a day takes up
@> more than all the .sigs from all your articles for the last few weeks.
@> However, your .sig is extremely annoying.
@
@Check out the bandwidth consumed by ONE jpeg and get back to me.
What are you trying to say?
--
B.B. --I am not a goat! [EMAIL PROTECTED] @airmail.net
------------------------------
From: "Edward Rosten" <[EMAIL PROTECTED]>
Subject: Re: The Windows guy.
Date: Thu, 01 Mar 2001 23:08:30 +0000
> :>There are examples of programs I can run where the output of the first
> :>program will never reach the input of the second under DOS pipes, but
> :>will always work under a multitasking syetem (under my definition of
> :>pipes). Since DOS pipes do not fit the definition, they are not pipes.
> :>
> :>
> :>prog_that_will_never_finish | some_other_prog
>
> : Oops, my bad. :-) You're right; if the first program never
> : finishes, the second program will see nothing.
>
> Technically what matters is that it never finishes the file, not that
> the program itself never finishes. The following
> (obviously contrived) example allowed tail to see the output,
> even though the feeder program never finished:
Not counting that your program uses unistd.h, it would still not work
under DOS, because it has to finish before the next process starts.
-Ed
--
| u98ejr
| @
Share, and enjoy. | eng.ox
| .ac.uk
------------------------------
From: Chris Ahlstrom <[EMAIL PROTECTED]>
Subject: Windoze Domination/Damnation
Date: Thu, 01 Mar 2001 23:11:51 GMT
I basically dislike Microsoft, and think their Windows operating
systems range from shit (Win 98) to passable (Win 2000 on heavy
hardware). The following link makes me sad that the idiocy
continues:
http://www.infoworld.com/articles/hn/xml/01/02/28/010228hnvise.xml?p=br&s=3
Partial quote:
HE DOMINANCE OF Microsoft's Windows operating system, the centerpiece
of the U.S. Department of Justice's antitrust case against the company, remains
undisputed, according to research released Wednesday by IDC.
Worldwide shipments of Windows 95, Windows 98, Windows NT Workstation,
Windows 2000 Professional, and Windows Millennium Edition comprised 92
percent of all client operating systems shipped last year, up from 89 percent the
year before, said Al Gillen, research manager at IDC's operating environments
program.
Windows 98 shipments were up 36 percent over the prior year, while Windows 95
shipments fell off dramatically, Gillen said. Overall, Microsoft Windows 9x and
Windows Millennium Edition operating system shipments were up by 8 percent.
Linux remains a bit player on the desktop with less than 2 percent market share,
although that is a 25 percent jump from 1999, IDC's research shows. Linux also
continues to garner backing from IT industry leaders, including IBM,
Hewlett-Packard, and Dell Computer, all of which are shipping workstations and
low-end servers with the Linux operating environment.
End of quote.
I'm not so worried about Linux "market share", since it says that it's
PAID percentage is small. I'm more worried why Win 98, a piece of crap
if there ever was one, has gone up in shipments.
Perhaps there are misleading numbers here. However, it looks like Microsoft
is now coated with Teflon. Even their crap sells.
Chris
--
[X] Check here to always trust content from Chris
[ ] Check here if you're a dazed follower of Bill Gates
------------------------------
From: "Masha Ku'Inanna" <[EMAIL PROTECTED]>
Subject: So, here's something to chew on...
Date: Thu, 1 Mar 2001 18:23:51 -0500
Reply-To: "Masha Ku'Inanna" <[EMAIL PROTECTED]>
So, under Windows 2000 Pro, i was merrily surfing along with 256M RAM, and
ICQ, AIM and Outlook Expres, and Livejournal's client running in the
background..
Went to start Musicmatch Jukebox, and things started to act odd..so, i
closed MM. Things were still acting odd. CTRL-ALT-DELETE, brought up the
"taskmanager" and found t he PID for Windows Explorer was using 95-100% CPU
time.
Time to kill it.
Clicked "Kill Process."
"Action denied."
What the fuck?..
Clicked it again.
"Action denied."
Grr..
Logged out of my non-admin user account, and tried as Administrator.
"Action denied."
Grr!!!..
Logged back into the non-admin account. Still had the same processes in the
background. Still frozen at 95-100% CPU time.
Now, I assume that Windows will refuse to let one disable things and turn
off things, to maintain system stability. But when you have a runaway
process like that, you've got to try to kill it, and restart it again,
right? I mean, under the UNIX world, I can always hit a keystroke, log into
another console, and ps all i like to find the runaway process, and kill it
by process id, log back out, and carry on my merry day, right?
I mean, rebooting is a sign of surrender, isn't it? When you've exhausted
everything you can think of, you reboot.
So, in so many words, I have an operating system that tells me "No, asshole,
you cannot do that because *I* know that if you disable that I will lock up,
crash, or grow unstable." Even though I realize the implications, and also
realized that it had already grown unstable, and found out what was burning
up CPU time.
Even though it was already unstable? I could not go in and try to remove
what was causing the problem, because my OS said it would not allow me to do
it?
At least with UNIX, there never is a question to whether or not your
computer will flat out refuse to do something that you tell it, as root. It
can question, but it will not refuse, to the best of my knowledge. It does
not assume to know more about what you need to do than you know. If you tell
it to do something boneheaded, by gosh, it will do exactly that.
I have never had to wrestle with a UNIX system because of simple problems. I
have never had to wonder if my computer will deny me access to something
because it felt it should not do such a thing. My only worry was because I
do not know UNIX enough, that I will type the wrong command and toast root.
Or /usr/sbin. or /usr anything. But that is because the fault would lie
squarely on my shoulders if something stupid were to happen.
So, I rebooted.
Uptime for Win2k Pro -- three hours.
=======================================================
Adrian Feliciano
[EMAIL PROTECTED]
***
Do What thou wilt shall be the whole of the law
Love is the law, love under will
-Aleister Crowley
Harm None
-Wiccan Law
As above, so below
-Hermetic Philosophy
=======================================================
------------------------------
From: "Interconnect" <[EMAIL PROTECTED]>
Subject: If I delete using rm?
Date: Fri, 2 Mar 2001 10:26:52 +1100
If I accidentally delete a subdirectory and files is there any way of
recovering these in Linux. That is without resorting to the tape backups?
Thanks for any hints or tips.
------------------------------
From: "Gary Hallock" <[EMAIL PROTECTED]>
Subject: Re: KDE or DOJ ?
Date: Thu, 01 Mar 2001 18:32:20 +0500
In article <c_mn6.918$[EMAIL PROTECTED]>, "Erik Funkenbusch"
<[EMAIL PROTECTED]> wrote:
> I used KDE 2.0 quite a bit. I can't imagine that 2.1 has changed it
> that radically in only the last 2 or 3 months since 2.0 was released.
Just as I thought. If you can't imagine that 2.1 has changed radically
then you lack imagination.
Gary
------------------------------
From: [EMAIL PROTECTED] (Markus Friedl)
Crossposted-To: comp.os.ms-windows.nt.advocacy,comp.security.ssh
Subject: Re: SSH vulnerabilities - still waiting [ was Interesting article ]
Date: 1 Mar 2001 23:41:45 GMT
In <lNsn6.1818$[EMAIL PROTECTED]> "Chad Myers"
<[EMAIL PROTECTED]> writes:
>I have also cited SSH.com representatives themselves who claim SSH1 is
>"fundamentally flawed". He even went on to criticize OpenSSH for using
>this protocol as it endangers many people who use OpenSSH.
you have no idea what you are talking about. the SSH1 flaws listed on
the ssh.com site are fixed in openssh.
-markus
------------------------------
From: Steve Mading <[EMAIL PROTECTED]>
Subject: Re: The Windows guy.
Date: 1 Mar 2001 23:41:37 GMT
Erik Funkenbusch <[EMAIL PROTECTED]> wrote:
: "Steve Mading" <[EMAIL PROTECTED]> wrote in message
: news:97m303$b16$[EMAIL PROTECTED]...
:> Erik Funkenbusch <[EMAIL PROTECTED]> wrote:
:> : "Steve Mading" <[EMAIL PROTECTED]> wrote in message
:> : news:97kca1$enc$[EMAIL PROTECTED]...
:> :> Erik Funkenbusch <[EMAIL PROTECTED]> wrote:
:> :> : "Steve Mading" <[EMAIL PROTECTED]> wrote in message
:> :> : news:97jp4h$ice$[EMAIL PROTECTED]...
:> :>
:> :> :> NO. Running code is NOT a program. Running code is a process.
:> :> :> The word "program" refers to the image in its static form, either
:> :> :> as an executable file (and the associated execution library files),
:> :> :> or as a loaded bunch of code in RAM. It doesn't become a "process"
:> :> :> until it is running. Here's an analogy: Program is to screenplay
:> :> :> as process is to movie.
:> :>
:> :> : Well then, I fail to understand your refusal to clasify running DOS
: code
:> : as
:> :> : a process then.
:> :>
:> :> I fail to see why you think that's what I've been saying, King
: Strawman.
:> :> Running DOS code IS a process, I even SAID that DOS is a single
:> :> process that never dies, right in this very thread. The pertinent
:> :> point, that you keep missing, is that it is ONE, count them, ONE
: process.
:> :> Inter-process requires that there be actual processes (plural) to
:> :> talk to each other. One process talking to itself using a temp
:> :> file is not "interprocess" by any stretch of the imagination.
:>
:> : Then tell me, how is it that TSR's can run concurrently with DOS
:> : applications? Are you going to claim that the TSR and the DOS
: application
:> : are the same process? Clearly, they're not.
:>
:> TSR's are not processes any more than the UNIX kernel is a "process".
: What is it then? You claim that a process is the instance of running code.
: The UNIX kernel doesn't run in the context of the user processes, so
The UNIX kernel behaves more like a Windows DLL than a Windows EXE.
It doesn't "run" (except at boot time when it's setting itself up,
much like a TSR). The kernel is a big library that other processes
can call into to implement the more low-level parts of their code, and
those parts of the code that require permission. The "mode switch" is
NOT a "context switch". It's a switch to turn off all the traps that
were put on memory and hardware so the process can do privileged things,
but the same process that ran the user-mode code also runs the
kernel-level code after the switch in modes is made.
: therefore, if it's running...
It's not.
: it must be in a process of some type (based on
: your loose definitions of a process).
The parts of a UNIX system that actually need to be doing
"thinking" while the ordinary processes run are themselves
also processes. They have names like "swapper", "init",
"kflush", and so on. These behave just like any other
process, but they often spend a large portion of their
time in kernel routines. In fact, some kernel routines
are written exclusively for the use of these processes.
It's these processes that do the OS's work. They make
heavy USE OF the kernel, but are not the kernel themselves.
This approach allows the kernel routines that have to be
"always running" to actually be running as seperate
processes. The MS-Windows way is to make the kernel be
one large executable, with threads inside it that do
these tasks. UNIX implementations typically do not do
that, instead opting for seperate processes as described
above. (However, theoretically an implementation COULD
do it the Windows way, and still be a "unix", but none of
the ones I KNOW about do. The reasoning is this: We need
to let any process make calls into the kernel. Seeing as
how we have to do that anyway, why not make the OS's
maintenence routines do the same thing all the other
processes are doing. Sure, mode switches are slow, but
if we put most of the logic in the kernel routine, then
there is little mode switching. These proceses will be
in kernel mode most of the time, and we'll write the
kernel routines specificly for these processes's main
loops. In this way, the kernel has no process of its own,
and no context record of its own. It's just a pile of code
that other processes are using.)
------------------------------
From: Steve Mading <[EMAIL PROTECTED]>
Crossposted-To: comp.lang.c
Subject: Re: Something Seemingly Simple.
Date: 1 Mar 2001 23:57:10 GMT
In comp.os.linux.advocacy Aaron Kulkis <[EMAIL PROTECTED]> wrote:
: This ALL comes under the section called "implementation dependant",
: and therefore, is not *really* defined, merely ***suggested***.
: Suggested != proscribed (i.e. defined, absolutely, no if's, and's, or but's)
See, whenever you use an implementation dependant feature,
like naming your own routine "printf", the standard DOES have
something defined to say about it. It says, "Your compiler
writer is allowed to make this fail and if he does, it's not
his fault - it's yours for using the feature we told you we
were not going to require to work."
"It's okay by us (the standards committee) if some compiler
makes your program that uses this feature break." That's very
different from the kind of "implementation dependant" that arises
when the standards committee says nothing at all about it (perhaps
because they didn't notice the situation might occur.)
Often when a feature is explicitly mentioned to be
"implementation dependant", it's not because they
want you to be able to do it, but it's because the
technical details of how to force you NOT to do it
would be large and ugly to implement. This might be
one such example. To force you to not name clash with
libc, when you aren't even linking with libc is still
a really good idea, with one exception: if you are
trying to write your own re-implementation of libc, or
perhaps you are the guy who's writing the first libc that
others will have to use. The problem is that it is hard
to enforce the rule while at the same time allowing for
this important exception case. That might be why they
chose to say "implementation dependant".
------------------------------
From: . <[EMAIL PROTECTED]>
Crossposted-To:
alt.destroy.microsoft,comp.os.ms-windows.advocacy,comp.os.ms-windows.nt.advocacy
Subject: Re: MS websites: a tale of total and humiliating failure!
Date: Fri, 2 Mar 2001 13:10:47 +1300
> No, they don't have any reason to change it.
> It is working? Stick with it.
This is not, apparently, part of the MS philosophy. If it's working,
it's time to incorporate it into Windows.
> You don't waste money where you don't need to.
You might not, and I might not... but MS? They have something to prove.
It's widely considered that NT and 2k aren't anywhere near the
reliability of decent UNIX systems. Whether that's true or not, MS now
has to try and prove that NT/2k really can cut it, and that's why they
would try and shift things to their own platform.
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list by posting to comp.os.linux.advocacy.
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Advocacy Digest
******************************
