🚀 go-pugleaf

In article <4itbl2$320@news.erie.net>, pce@moose.erie.net (Paul
Christenson) wrote:

> Nevin ":-]" Liber (nevin@CS.Arizona.EDU) wrote:
> : In article <4in381$ri9@steel.interlog.com>, daveacg@interlog.com (Dave
> : Glue) wrote:

> : > You mean in 1991, the Mac OS could dynamically allocate memory to apps
> : > without the user "specifying" how much memory each apps needs by
> : > him/herself?

> : On the flip side, Unix users have to set process limits so that things
> : like Netscape don't bring the whole machine to a snail's pace.

> I have *NEVER* had to set a "process limit" on ANY *nix system that I've
> used, from my (primarily) CLI Linux system here at home, to the SGIs at
> school.  I've run Netscape concurrently with several shells and programs
> such as Matlab.

What version of Netscape?  Go to a JavaScript page and let it sit there
for a while.  People have reported not only noticeable slowdowns of their
machines, but even crashes of the supposedly uncrashable Unix (it has
memory protection; how could an app possibly crash it?  How could the
Titanic sink?).

> The only time I would need to set priorities is when I'm
> running a timing-sensitive process such as RealAudio.

The claim was these types of problems went away with PMT.  Guess not.

> Many of the MEs at school regularly run Netscape while waiting for a long
> ANSYS job to run.  Most programs (like Netscape) spend most of their time 
> waiting for input, be it user input or data from the net.

Are you absolutely sure that someone hasn't imposed a per-process limit on
the amount of VM a process is allowed to allocate?  There is no difference
between an admin setting it and you setting it, in my eyes.  A limit is a
limit.
-- 
        Nevin ":-)" Liber       nevin@CS.Arizona.EDU    (520) 293-2799

In <nevin-2203960538520001@mac-nevin.cs.arizona.edu> nevin@CS.Arizona.EDU (Nevin ":-]" Liber) writes:

>What version of Netscape?  Go to a JavaScript page and let it sit there
>for a while.  People have reported not only noticeable slowdowns of their
>machines, but even crashes of the supposedly uncrashable Unix (it has
>memory protection; how could an app possibly crash it?  How could the
>Titanic sink?).

Give me an URL which will cause Netscape 2.0 to crash my Unix box or cut
the crap.

Dan
--
Dan Pop
CERN, CN Division
Email: danpop@mail.cern.ch 
Mail:  CERN - PPE, Bat. 31 R-004, CH-1211 Geneve 23, Switzerland

: > : On the flip side, Unix users have to set process limits so that things
: > : like Netscape don't bring the whole machine to a snail's pace.

: > I have *NEVER* had to set a "process limit" on ANY *nix system that I've
: > used, from my (primarily) CLI Linux system here at home, to the SGIs at
: > school.  I've run Netscape concurrently with several shells and programs
: > such as Matlab.

: What version of Netscape?  Go to a JavaScript page and let it sit there
: for a while.  People have reported not only noticeable slowdowns of their
: machines, but even crashes of the supposedly uncrashable Unix (it has
: memory protection; how could an app possibly crash it?  How could the
: Titanic sink?).

Netscape 2.0.  The closest I have ever had to a "crash" was a core dump; it
simply shut down the task and went on its merry way.

The only time that happened with Netscape 2 is when I didn't install it
correctly.  Netscape, being a precompiled app, requires certain files to
be in certain locations.

: > The only time I would need to set priorities is when I'm
: > running a timing-sensitive process such as RealAudio.

: The claim was these types of problems went away with PMT.  Guess not.

I said "would need to", not "had to".  Big difference.  RealAudio takes 
about 50% of processor power.  If I valued a steady audio stream above 
everything else I'm doing, then I change the process priority.  Otherwise,
I let the OS do its job.

: > Many of the MEs at school regularly run Netscape while waiting for a long
: > ANSYS job to run.  Most programs (like Netscape) spend most of their time 
: > waiting for input, be it user input or data from the net.

: Are you absolutely sure that someone hasn't imposed a per-process limit on
: the amount of VM a process is allowed to allocate?  There is no difference
: between an admin setting it and you setting it, in my eyes.  A limit is a
: limit.

No limits that I can detect.  I can drive the machine into swap with one 
process if I care to.  I can also have 100% processor utilization with
one process.  The allocations are dynamic.

On 24 Mar 96 21:30:51 GMT, danpop@mail.cern.ch (Dan Pop) wrote:

>In <nevin-2203960538520001@mac-nevin.cs.arizona.edu> nevin@CS.Arizona.EDU (Nevin ":-]" Liber) writes:
>
>>What version of Netscape?  Go to a JavaScript page and let it sit there
>>for a while.  People have reported not only noticeable slowdowns of their
>>machines, but even crashes of the supposedly uncrashable Unix (it has
>>memory protection; how could an app possibly crash it?  How could the
>>Titanic sink?).
>
>Give me an URL which will cause Netscape 2.0 to crash my Unix box or cut
>the crap.
>
>Dan

	I can't remember the URL but NS 2.0 has crashed twice on our
Unix system (irix I think) It crashed so that I had no keyboard
control or anything. However, not a *fatal* crash: I could rlogin from
another machine and kill it. It still sucks that it can do that
though.

		Jaliya

Jaliya Jayawardena wrote:
> 
> On 24 Mar 96 21:30:51 GMT, danpop@mail.cern.ch (Dan Pop) wrote:
> 
> >In <nevin-2203960538520001@mac-nevin.cs.arizona.edu> nevin@CS.Arizona.EDU (Nevin ":-]" Liber) writes:
> >
> >>What version of Netscape?  Go to a JavaScript page and let it sit there
> >>for a while.  People have reported not only noticeable slowdowns of their
> >>machines, but even crashes of the supposedly uncrashable Unix (it has
> >>memory protection; how could an app possibly crash it?  How could the
> >>Titanic sink?).
> >
> >Give me an URL which will cause Netscape 2.0 to crash my Unix box or cut
> >the crap.
> >
> >Dan
> 
>         I can't remember the URL but NS 2.0 has crashed twice on our
> Unix system (irix I think) It crashed so that I had no keyboard
> control or anything. However, not a *fatal* crash: I could rlogin from
> another machine and kill it. It still sucks that it can do that
> though.
> 
>                 Jaliya

That's not crashing.  If you can switch to another console (or virtual console)
and kill it, it has *not* crashed you're system.  All it did was trap the input
device (keyboard) so the OS couldn't tell that you wanted to kill it.  It's a
bit like the difference between hijacking a boat (if that's how you would say
it) and sinking it.

That's why Unix is said to be uncrashable.  In, say, DOS, stupid, careless 
errors on th part of the programmer can bring the computer to its knees, and
you helpless to recover, because DOS trusts any running program with exclusive
control over the computer.  *That* is crashing a computer, my friend (and a
stupid move on the part of the authors of DOS)

In article <4j7him$vst@news.erie.net>, pce@moose.erie.net (Paul
Christenson) wrote:

> I said "would need to", not "had to".  Big difference.  RealAudio takes 
> about 50% of processor power.

What kind of machine are you running?  It certainly doesn't take that much
on my lowly Quadra 800.
-- 
        Nevin ":-)" Liber       nevin@CS.Arizona.EDU    (520) 293-2799

G'Day Dan Pop, you wrote:
: In <nevin-2203960538520001@mac-nevin.cs.arizona.edu> nevin@CS.Arizona.EDU (Nevin ":-]" Liber) writes:

: >What version of Netscape?  Go to a JavaScript page and let it sit there
: >for a while.  People have reported not only noticeable slowdowns of their
: >machines, but even crashes of the supposedly uncrashable Unix (it has
: >memory protection; how could an app possibly crash it?  How could the
: >Titanic sink?).

: Give me an URL which will cause Netscape 2.0 to crash my Unix box or cut
: the crap.

Okay. Place a large picture on you background.
(The one I used to have was the Enterprise at 1024x768x24)
Pick a graphic intensive sight, and then watch Netscape go bye-byes.

The terminals we use are running the SunOS with 4Mb of Graphics Memory, and
the main server has over 180Mb RAM AFAIK.

Netscape is the ONLY programme I have seen do this too.

: Dan
: --
: Dan Pop
: CERN, CN Division
: Email: danpop@mail.cern.ch 
: Mail:  CERN - PPE, Bat. 31 R-004, CH-1211 Geneve 23, Switzerland

In <4kafvi$n4t@pellew.ntu.edu.au> jamesm@it.ntu.edu.au (James McArthur) writes:

>G'Day Dan Pop, you wrote:
>: In <nevin-2203960538520001@mac-nevin.cs.arizona.edu> nevin@CS.Arizona.EDU (Nevin ":-]" Liber) writes:
>
>: >What version of Netscape?  Go to a JavaScript page and let it sit there
>: >for a while.  People have reported not only noticeable slowdowns of their
>: >machines, but even crashes of the supposedly uncrashable Unix (it has
>: >memory protection; how could an app possibly crash it?  How could the
>: >Titanic sink?).
>
>: Give me an URL which will cause Netscape 2.0 to crash my Unix box or cut
>: the crap.
>
>Okay. Place a large picture on you background.
>(The one I used to have was the Enterprise at 1024x768x24)
>Pick a graphic intensive sight, and then watch Netscape go bye-byes.

Did you actually bother to read what I was asking?  If you did, you'd have
noticed that I wanted to see Solaris go bye-bye because of Netscape, not
Netscape itself.

Dan
--
Dan Pop
CERN, CN Division
Email: danpop@mail.cern.ch 
Mail:  CERN - PPE, Bat. 31 R-004, CH-1211 Geneve 23, Switzerland

Dan Pop (danpop@mail.cern.ch) wrote:
: Did you actually bother to read what I was asking?  If you did, you'd have
: noticed that I wanted to see Solaris go bye-bye because of Netscape, not
: Netscape itself.

A tremendous number of people don't understand that "crash proof" systems 
like UNIX don't actually stop APPS from crashing, but rather they stop 
apps from crashing OTHER apps and the OS. When a UNIX app crashes, you 
just look at the core dump and find out what went wrong, recompile, and 
you're back up without a reboot.